From wade@cs.utk.edu Tue Feb 9 10:47:32 1993 Received: from THUD.CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA26386; Tue, 9 Feb 93 10:47:32 -0500 Received: from LOCALHOST.cs.utk.edu by thud.cs.utk.edu with SMTP (5.61++/2.7c-UTK) id AA26430; Tue, 9 Feb 93 10:47:26 -0500 Message-Id: <9302091547.AA26430@thud.cs.utk.edu> To: pbwg-comm-archive@surfer.EPM.ORNL.GOV Cc: wade@cs.utk.edu Subject: test Date: Tue, 09 Feb 93 10:47:25 EST From: Reed Wade From owner-pbwg-comm@CS.UTK.EDU Sun Feb 14 16:56:44 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA14958; Sun, 14 Feb 93 16:56:44 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA24403; Sun, 14 Feb 93 16:56:04 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Sun, 14 Feb 1993 16:56:03 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from DASHER.CS.UTK.EDU by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA24397; Sun, 14 Feb 93 16:56:02 -0500 From: Jack Dongarra Received: by dasher.cs.utk.edu (5.61++/2.7c-UTK) id AA01941; Sun, 14 Feb 93 16:56:01 -0500 Date: Sun, 14 Feb 93 16:56:01 -0500 Message-Id: <9302142156.AA01941@dasher.cs.utk.edu> To: pbwg-comm@cs.utk.edu Subject: Parallel Benchmark Working Group Meeting For planning purposes, I would like to know how many people will be attending the Parallel Benchmark Working Group (PBWG) meeting on March 1th and 2nd, 1993 in Knoxville, Tennessee. Let me know if your will or will not be attending the meeting. Best wishes, Jack From @ecs.soton.ac.uk,@diana.ecs.soton.ac.uk:C.D.Collier@ecs.southampton.ac.uk Mon Jan 11 12:30:42 1993 Return-Path: <@ecs.soton.ac.uk,@diana.ecs.soton.ac.uk:C.D.Collier@ecs.southampton.ac.uk> Received: from sun2.nsfnet-relay.ac.uk by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA04153; Mon, 11 Jan 93 12:30:09 -0500 Via: uk.ac.southampton.ecs; Mon, 11 Jan 1993 17:27:19 +0000 Via: diana.ecs.soton.ac.uk; Mon, 11 Jan 93 17:20:06 GMT From: Christine Collier Message-Id: <6780.9301111725@diana.ecs.soton.ac.uk> Subject: Meeting Minneapolis 18th November, 1992 To: dbailey@nas.nasa.com, iyb@lanl.gov, sbveit@ksv.com, carterl@watson.ibm.com, thec@newton.national-physical-lab.co.uk, dongarra@cs.utk.edu, dem@cxa.dl.ac.uk, j.flemming@cray.com, gcfe@npac.syr.edu, danielf@kgnvma.vnet.ibm.com, paulg@meiko.com, gent@genias.de, cmg@cray.com, harp@revmes.mod.uk, siamak@fai.com, hempel@gmd.de, ajgh@ecs.soton.ac.uk, rwh@uk.pac.soton.ecs, j1mart@kgnvmz.vnet.ibm.com, hcooke@parsys.co.uk, messina@caltech.edu, Bminto@cray.com, dennisp@think.com, schneid@csrd.uiuc.edu, simon@nas.nasa.gov, actstea@ml.ruu.cc, frannie@Parsytech.de Date: Mon, 11 Jan 1993 17:25:46 +0000 (GMT) X-Mailer: ELM [version 2.4 PL8] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 7692 Status: RO Parallel Benchmark Working Group Summary of 1st Meeting held in Minneapolis Convention Centre Wednesday November 18th, 1992 1. Introduction and Welcome The meeting was opened by Professor Roger Hockney who welcomed all the participants and asked Tony Hey to say a few words about the background to the meeting. Tony Hey outlined the history of attempts in Europe to establish credible and useful benchmarks for the evaluation of Distributed Memory MIMD systems. At the time the European Genesis work began in 1988 there were no suitable DM message-passing benchmarks - although the Caltech group also undertook a study at about the same time. Four years later the scene has moved on, and besides the Genesis benchmarks for message-passing Fortran programs, HPF/Fortran 90 benchmarks are now appearing. The Perfect Club are also now looking at benchmarks for DM systems and the NASA Ames 'Pencil and Paper' benchmarks are being taken seriously by the vendors. From a situation where there were very few suitable benchmark codes we are now approaching a situation where there is an over-abundance of benchmarks and inevitable duplication of effort. This is undesirable both for the vendors and for the end users of systems. With the adoption of HPF and the current proposal for a standard message-passing interface (MPI), discussed yesterday at the message-passing standards workshop, there seems to be a real window of opportunity to gather together US and European benchmarkers to agree on a useful subset. Roger Hockney then asked those present to introduce themselves and make some remark about their interest in this activity. A list of attendees is attached to this summary. There was widespread support for such an activity and a summary of contributions as follows: Horst Simon of NASA Ames suggested that we follow the HPF model and form working groups in identified areas. He also voiced reservations about any attempt to create a fully comprehensive benchmark suite. Denis Parkinson of TMC supported the activity but pointed out that it was important that this did not become "yet another" set of benchmarks for vendors to implement. He also stressed the need for HPF versions as well as message-passing versions, and raised the question of scalability of benchmarks. Aard van der Steen raised the question of Japanese participation in such an activity. Siamak Hassanzadeh of Fujitsu supported the proposal and suggested the inclusion of seismic benchmarking codes. Rolf Hempel of the GMD in Sankt Augustin stressed his organization's support for the activity and briefly described the RAPS benchmarking initiative in Europe. This is a consortium of users and software houses, led by ECMWF and including ESI and AVL, supported by Convex, Cray, Fujitsu, IBM, Intel and Meiko RAPS is an acronym for Real Applications on Parallel Systems. The Genesis benchmark codes are included as a public domain subset of RAPS but the major codes would not be publicly available. Bill Minto of Cray UK stressed that real applications were important and that it was also necessary to address as wide a spread of applications as possible. Geoffrey Fox of Syracuse suggested that it was only realistic at this stage at attempt to co-ordinate benchmarking activities. His activities at Syracuse had lately been concerned with constructing benchmarks for validating HPF and Fortran 90 compilers. Trevor Chambers of NPL, UK gave his support to the activity and announced that a major new European benchmarking project PEPS with NPL were involvement has just started. Gordon Harp of DRA, Malvern UK said that the DRA (Defence Research Agency) were involved in a collaboration with agencies in the US, Canada and Australia looking at benchmarks for defence applications. He was concerned that there was a need for real applications and issues such as scalability and power consumption were important from the DRA perspective. Francis Wray of Parsytec and Paul Garrett of Meiko both voiced concern that this activity should not generate more rather than less work in procurement activities. David Schneider from CRSD Illinois, representing the Perfect Club, welcomed an attempt to eliminate redundant effort. He stressed that benchmarks had multiple uses - such as education and compiler evaluation - as well as specific application knowledge. He thought that there may be budget problems if a very organised activity was envisaged. Benchmarks spanning a range of architectures and the need for public domain codes were also stressed. Joanne Martin of the HPSSL IBM Kingston welcomed the initiative and stressed how important it was for vendors not to be confronted with many sets of "standard" benchmarks. 2. Objectives There were no objections to the draft objectives for the group. These objectives are: 1. To establish a comprehensive set of parallel benchmarks that is generally accepted by both users and vendors of parallel system. 2. To provide a focus for parallel benchmark activities and avoid unnecessary duplication of effort and proliferation of benchmarks. 3. To set standards for benchmarking methodology and result-reporting together with a control database/repository for both the benchmarks and the results. 3. Mode of Working 3.1 It was agreed that on HPF-like forum style of working should be adopted with a view to convergence to an agreed set of benchmarks and procedures within 12 months. 3.2 There was not seen to be a need for meetings every six weeks but in order to generate momentum for the project it was thought that two meetings a year were too few. 3.3 Jack Dongarra volunteered to set up a database for benchmarks and results at ORNL. NPL were willing to maintain a European copy of this database. 3.4 It was agreed that all present should send existing benchmarks to Jack Dongarra at ORNL. 3.5 Jack Dongarra and Aard van der Steen agreed to examine the available benchmarks submitted and attempt to classify them appropriately. 3.6 Three other working groups were identified with named individuals taking responsibility for each group. These were as follows: Methodology: Bailey, Hockney and Schneider Kernel Benchmarks: Dongarra, Hockney, van der Steen, Wray Compiler Benchmarks: Fox, Grassl 3.7 A number of application areas were discussed as possible working groups e.g. CFD, Seismic, QCD etc., but it was thought premature to activate such groups at this time. 4. Future Activity 4.1 Jack Dongarra agreed to set up a mail refector at ORNL for the Parallel Benchmark Working Group (PBWG) and to organize the relevant subgroups along the lines of the HPF forum. 4.2 Jack Dongarra also agreed to host the next meeting of the PBWG. Subsequent discussions after the formal close of the meeting led to the dates of March 1st/2nd being selected. 4.3 At the March meeting, each subgroup will produce a discussion document and a benchmark database classification will be proposed. Further discussion on procedures for the PBWG was also deferred until then. From owner-pbwg-comm@CS.UTK.EDU Sun Feb 21 12:08:07 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA10794; Sun, 21 Feb 93 12:08:07 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA19152; Sun, 21 Feb 93 12:07:34 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Sun, 21 Feb 1993 12:07:32 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from THUD.CS.UTK.EDU by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA19145; Sun, 21 Feb 93 12:07:28 -0500 From: Jack Dongarra Received: by thud.cs.utk.edu (5.61++/2.7c-UTK) id AA05996; Sun, 21 Feb 93 12:07:24 -0500 Date: Sun, 21 Feb 93 12:07:24 -0500 Message-Id: <9302211707.AA05996@thud.cs.utk.edu> To: pbwg-comm@cs.utk.edu Subject: change of room for the PBWG meeting We have had a change in the meeting room for the Parallel Benchmark Working Group. The new meeting room is in the University Center Room 201. The postscript file below contains a map that may help. Look forward to seeing you next week. Regards, Jack %!PS-Adobe-2.0 EPSF-1.2 %%DocumentFonts: Helvetica-Bold Courier Courier-Bold Times-Bold %%Pages: 1 %%BoundingBox: 39 -113 604 767 %%EndComments /arrowHeight 10 def /arrowWidth 5 def /IdrawDict 54 dict def IdrawDict begin /reencodeISO { dup dup findfont dup length dict begin { 1 index /FID ne { def }{ pop pop } ifelse } forall /Encoding ISOLatin1Encoding def currentdict end definefont } def /ISOLatin1Encoding [ /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /space/exclam/quotedbl/numbersign/dollar/percent/ampersand/quoteright /parenleft/parenright/asterisk/plus/comma/minus/period/slash /zero/one/two/three/four/five/six/seven/eight/nine/colon/semicolon /less/equal/greater/question/at/A/B/C/D/E/F/G/H/I/J/K/L/M/N /O/P/Q/R/S/T/U/V/W/X/Y/Z/bracketleft/backslash/bracketright /asciicircum/underscore/quoteleft/a/b/c/d/e/f/g/h/i/j/k/l/m /n/o/p/q/r/s/t/u/v/w/x/y/z/braceleft/bar/braceright/asciitilde /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/dotlessi/grave/acute/circumflex/tilde/macron/breve /dotaccent/dieresis/.notdef/ring/cedilla/.notdef/hungarumlaut /ogonek/caron/space/exclamdown/cent/sterling/currency/yen/brokenbar /section/dieresis/copyright/ordfeminine/guillemotleft/logicalnot /hyphen/registered/macron/degree/plusminus/twosuperior/threesuperior /acute/mu/paragraph/periodcentered/cedilla/onesuperior/ordmasculine /guillemotright/onequarter/onehalf/threequarters/questiondown /Agrave/Aacute/Acircumflex/Atilde/Adieresis/Aring/AE/Ccedilla /Egrave/Eacute/Ecircumflex/Edieresis/Igrave/Iacute/Icircumflex /Idieresis/Eth/Ntilde/Ograve/Oacute/Ocircumflex/Otilde/Odieresis /multiply/Oslash/Ugrave/Uacute/Ucircumflex/Udieresis/Yacute /Thorn/germandbls/agrave/aacute/acircumflex/atilde/adieresis /aring/ae/ccedilla/egrave/eacute/ecircumflex/edieresis/igrave /iacute/icircumflex/idieresis/eth/ntilde/ograve/oacute/ocircumflex /otilde/odieresis/divide/oslash/ugrave/uacute/ucircumflex/udieresis /yacute/thorn/ydieresis ] def /Helvetica-Bold reencodeISO def /Courier reencodeISO def /Courier-Bold reencodeISO def /Times-Bold reencodeISO def /none null def /numGraphicParameters 17 def /stringLimit 65535 def /Begin { save numGraphicParameters dict begin } def /End { end restore } def /SetB { dup type /nulltype eq { pop false /brushRightArrow idef false /brushLeftArrow idef true /brushNone idef } { /brushDashOffset idef /brushDashArray idef 0 ne /brushRightArrow idef 0 ne /brushLeftArrow idef /brushWidth idef false /brushNone idef } ifelse } def /SetCFg { /fgblue idef /fggreen idef /fgred idef } def /SetCBg { /bgblue idef /bggreen idef /bgred idef } def /SetF { /printSize idef /printFont idef } def /SetP { dup type /nulltype eq { pop true /patternNone idef } { dup -1 eq { /patternGrayLevel idef /patternString idef } { /patternGrayLevel idef } ifelse false /patternNone idef } ifelse } def /BSpl { 0 begin storexyn newpath n 1 gt { 0 0 0 0 0 0 1 1 true subspline n 2 gt { 0 0 0 0 1 1 2 2 false subspline 1 1 n 3 sub { /i exch def i 1 sub dup i dup i 1 add dup i 2 add dup false subspline } for n 3 sub dup n 2 sub dup n 1 sub dup 2 copy false subspline } if n 2 sub dup n 1 sub dup 2 copy 2 copy false subspline patternNone not brushLeftArrow not brushRightArrow not and and { ifill } if brushNone not { istroke } if 0 0 1 1 leftarrow n 2 sub dup n 1 sub dup rightarrow } if end } dup 0 4 dict put def /Circ { newpath 0 360 arc patternNone not { ifill } if brushNone not { istroke } if } def /CBSpl { 0 begin dup 2 gt { storexyn newpath n 1 sub dup 0 0 1 1 2 2 true subspline 1 1 n 3 sub { /i exch def i 1 sub dup i dup i 1 add dup i 2 add dup false subspline } for n 3 sub dup n 2 sub dup n 1 sub dup 0 0 false subspline n 2 sub dup n 1 sub dup 0 0 1 1 false subspline patternNone not { ifill } if brushNone not { istroke } if } { Poly } ifelse end } dup 0 4 dict put def /Elli { 0 begin newpath 4 2 roll translate scale 0 0 1 0 360 arc patternNone not { ifill } if brushNone not { istroke } if end } dup 0 1 dict put def /Line { 0 begin 2 storexyn newpath x 0 get y 0 get moveto x 1 get y 1 get lineto brushNone not { istroke } if 0 0 1 1 leftarrow 0 0 1 1 rightarrow end } dup 0 4 dict put def /MLine { 0 begin storexyn newpath n 1 gt { x 0 get y 0 get moveto 1 1 n 1 sub { /i exch def x i get y i get lineto } for patternNone not brushLeftArrow not brushRightArrow not and and { ifill } if brushNone not { istroke } if 0 0 1 1 leftarrow n 2 sub dup n 1 sub dup rightarrow } if end } dup 0 4 dict put def /Poly { 3 1 roll newpath moveto -1 add { lineto } repeat closepath patternNone not { ifill } if brushNone not { istroke } if } def /Rect { 0 begin /t exch def /r exch def /b exch def /l exch def newpath l b moveto l t lineto r t lineto r b lineto closepath patternNone not { ifill } if brushNone not { istroke } if end } dup 0 4 dict put def /Text { ishow } def /idef { dup where { pop pop pop } { exch def } ifelse } def /ifill { 0 begin gsave patternGrayLevel -1 ne { fgred bgred fgred sub patternGrayLevel mul add fggreen bggreen fggreen sub patternGrayLevel mul add fgblue bgblue fgblue sub patternGrayLevel mul add setrgbcolor eofill } { eoclip originalCTM setmatrix pathbbox /t exch def /r exch def /b exch def /l exch def /w r l sub ceiling cvi def /h t b sub ceiling cvi def /imageByteWidth w 8 div ceiling cvi def /imageHeight h def bgred bggreen bgblue setrgbcolor eofill fgred fggreen fgblue setrgbcolor w 0 gt h 0 gt and { l b translate w h scale w h true [w 0 0 h neg 0 h] { patternproc } imagemask } if } ifelse grestore end } dup 0 8 dict put def /istroke { gsave brushDashOffset -1 eq { [] 0 setdash 1 setgray } { brushDashArray brushDashOffset setdash fgred fggreen fgblue setrgbcolor } ifelse brushWidth setlinewidth originalCTM setmatrix stroke grestore } def /ishow { 0 begin gsave fgred fggreen fgblue setrgbcolor /fontDict printFont printSize scalefont dup setfont def /descender fontDict begin 0 [FontBBox] 1 get FontMatrix end transform exch pop def /vertoffset 1 printSize sub descender sub def { 0 vertoffset moveto show /vertoffset vertoffset printSize sub def } forall grestore end } dup 0 3 dict put def /patternproc { 0 begin /patternByteLength patternString length def /patternHeight patternByteLength 8 mul sqrt cvi def /patternWidth patternHeight def /patternByteWidth patternWidth 8 idiv def /imageByteMaxLength imageByteWidth imageHeight mul stringLimit patternByteWidth sub min def /imageMaxHeight imageByteMaxLength imageByteWidth idiv patternHeight idiv patternHeight mul patternHeight max def /imageHeight imageHeight imageMaxHeight sub store /imageString imageByteWidth imageMaxHeight mul patternByteWidth add string def 0 1 imageMaxHeight 1 sub { /y exch def /patternRow y patternByteWidth mul patternByteLength mod def /patternRowString patternString patternRow patternByteWidth getinterval def /imageRow y imageByteWidth mul def 0 patternByteWidth imageByteWidth 1 sub { /x exch def imageString imageRow x add patternRowString putinterval } for } for imageString end } dup 0 12 dict put def /min { dup 3 2 roll dup 4 3 roll lt { exch } if pop } def /max { dup 3 2 roll dup 4 3 roll gt { exch } if pop } def /midpoint { 0 begin /y1 exch def /x1 exch def /y0 exch def /x0 exch def x0 x1 add 2 div y0 y1 add 2 div end } dup 0 4 dict put def /thirdpoint { 0 begin /y1 exch def /x1 exch def /y0 exch def /x0 exch def x0 2 mul x1 add 3 div y0 2 mul y1 add 3 div end } dup 0 4 dict put def /subspline { 0 begin /movetoNeeded exch def y exch get /y3 exch def x exch get /x3 exch def y exch get /y2 exch def x exch get /x2 exch def y exch get /y1 exch def x exch get /x1 exch def y exch get /y0 exch def x exch get /x0 exch def x1 y1 x2 y2 thirdpoint /p1y exch def /p1x exch def x2 y2 x1 y1 thirdpoint /p2y exch def /p2x exch def x1 y1 x0 y0 thirdpoint p1x p1y midpoint /p0y exch def /p0x exch def x2 y2 x3 y3 thirdpoint p2x p2y midpoint /p3y exch def /p3x exch def movetoNeeded { p0x p0y moveto } if p1x p1y p2x p2y p3x p3y curveto end } dup 0 17 dict put def /storexyn { /n exch def /y n array def /x n array def n 1 sub -1 0 { /i exch def y i 3 2 roll put x i 3 2 roll put } for } def %%EndProlog %%BeginIdrawPrologue /arrowhead { 0 begin transform originalCTM itransform /taily exch def /tailx exch def transform originalCTM itransform /tipy exch def /tipx exch def /dy tipy taily sub def /dx tipx tailx sub def /angle dx 0 ne dy 0 ne or { dy dx atan } { 90 } ifelse def gsave originalCTM setmatrix tipx tipy translate angle rotate newpath arrowHeight neg arrowWidth 2 div moveto 0 0 lineto arrowHeight neg arrowWidth 2 div neg lineto patternNone not { originalCTM setmatrix /padtip arrowHeight 2 exp 0.25 arrowWidth 2 exp mul add sqrt brushWidth mul arrowWidth div def /padtail brushWidth 2 div def tipx tipy translate angle rotate padtip 0 translate arrowHeight padtip add padtail add arrowHeight div dup scale arrowheadpath ifill } if brushNone not { originalCTM setmatrix tipx tipy translate angle rotate arrowheadpath istroke } if grestore end } dup 0 9 dict put def /arrowheadpath { newpath arrowHeight neg arrowWidth 2 div moveto 0 0 lineto arrowHeight neg arrowWidth 2 div neg lineto } def /leftarrow { 0 begin y exch get /taily exch def x exch get /tailx exch def y exch get /tipy exch def x exch get /tipx exch def brushLeftArrow { tipx tipy tailx taily arrowhead } if end } dup 0 4 dict put def /rightarrow { 0 begin y exch get /tipy exch def x exch get /tipx exch def y exch get /taily exch def x exch get /tailx exch def brushRightArrow { tipx tipy tailx taily arrowhead } if end } dup 0 4 dict put def %%EndIdrawPrologue %I Idraw 10 Grid 2.84217e-39 0 %%Page: 1 1 Begin %I b u %I cfg u %I cbg u %I f u %I p u %I t [ 0.799705 0 0 0.799705 0 0 ] concat /originalCTM matrix currentmatrix def Begin %I Pict %I b u %I cfg u %I cbg u %I f u %I p u %I t [ 1 0 0 1 89.1002 831.2 ] concat Begin %I Elli %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.09512e-08 0.9 -0.9 1.09512e-08 584.1 0.89999 ] concat %I 79 550 18 17 Elli End Begin %I Line %I b 65535 3 0 1 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.09512e-08 0.9 -0.9 1.09512e-08 541.8 -27 ] concat %I 110 466 110 542 Line %I 1 End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.09512e-08 0.9 -0.9 1.09512e-08 541.8 -27 ] concat %I 82 504 140 504 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-helvetica-bold-r-*-140-* Helvetica-Bold 14 SetF %I t [ 1.2168e-08 1 -1 1.2168e-08 35.5 66.1 ] concat %I [ (N) ] Text End End %I eop Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.23147 0.0157385 -0.0157385 -1.23147 409.218 -127.169 ] concat %I [ (Voluteer Boulevard) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-bold-r-*-120-* Courier-Bold 12 SetF %I t [ 3.66661e-08 3.01333 -3.01333 3.66661e-08 57.9267 164.513 ] concat %I [ (UT Campus -- Jack Dongarra's Lab) ] Text End Begin %I Rect none SetB %I b n %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 1 SetP %I t [ 1.2168e-08 1 -1 1.2168e-08 693.5 -156 ] concat %I 17 61 177 478 Rect End Begin %I Line %I b 65535 2 1 1 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 6.9812e-09 0.698798 -0.573736 8.50294e-09 453.281 -79.7194 ] concat %I 158 545 1015 545 Line %I 1 End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 5.81767e-09 0.582331 -0.478114 7.08579e-09 333.274 -81.9323 ] concat %I 257 403 4 4 Elli End Begin %I Line %I b 65535 2 0 1 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 7.02094e-09 0.702776 -0.577002 8.55135e-09 455.344 -80.5588 ] concat %I 211 544 211 17 Line %I 1 End Begin %I BSpl %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 6.9812e-09 0.698798 -0.573736 8.50294e-09 455.594 -124.448 ] concat %I 5 503 545 514 535 546 529 628 529 628 529 5 BSpl %I 1 End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 5.81767e-09 0.582331 -0.478114 7.08579e-09 333.274 76.6169 ] concat %I 257 403 4 4 Elli End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 5.81767e-09 0.582331 -0.478114 7.08579e-09 403.382 -81.1611 ] concat %I 257 403 4 4 Elli End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.15378e-08 1.1549 -0.948215 1.40528e-08 222.192 481.065 ] concat %I [ (DOWN) (TOWN) ] Text End Begin %I Poly %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.75 SetP %I t [ 2.43189e-09 0.222447 -0.19986 2.70672e-09 310.54 254.841 ] concat %I 4 301 50 850 50 850 537 301 537 4 Poly End Begin %I BSpl %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 9 251 541 251 516 251 507 257 495 263 489 413 491 472 496 485 499 485 499 9 BSpl %I 1 End Begin %I BSpl %I b 65520 2 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.18822e-09 0.43379 -0.755115 5.27834e-09 620.126 111.397 ] concat %I 6 486 498 502 500 510 504 514 514 514 529 513 542 6 BSpl %I 1 End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2378 ] concat %I 476 654 476 542 Line %I 1 End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 5.81767e-09 0.582331 -0.478114 7.08579e-09 445.797 -81.1611 ] concat %I 257 403 4 4 Elli End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 252.947 159.612 ] concat %I [ (Volunteer Boulevard) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 310.947 152.672 ] concat %I [ (Neyland Drive) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 198.965 155.756 ] concat %I [ (Cumberland Avenue) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 130.329 90.977 ] concat %I [ (Interstate 40) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 130.329 403.305 ] concat %I [ (Interstate 40) ] Text End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0 SetP %I t [ 5.81767e-09 0.582331 -0.478114 7.08579e-09 332.433 366.124 ] concat %I 257 403 4 4 Elli End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 107 542 486 541 Line %I 1 End Begin %I Line %I b 65520 2 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 487 541 664 541 Line %I 1 End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 665 541 682 540 Line %I 1 End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -51.4666 ] concat %I 681 599 681 485 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.931845 2.26773e-08 -2.26773e-08 -0.931845 205.214 462.905 ] concat %I [ (Henley Street) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.931845 2.26773e-08 -2.26773e-08 -0.931845 193.647 318.042 ] concat %I [ (17th Street) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 110.137 205.729 ] concat %I [ (17th Street Exit) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 110.137 26.0435 ] concat %I [ (Airport/Alcoa Highway Exit) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-times-bold-r-*-140-* Times-Bold 14 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 171.727 80.0262 ] concat %I [ (Cumberland Avenue Exit) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 265.915 92.3651 ] concat %I [ (Neyland Drive Exit) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 110.137 509.574 ] concat %I [ (Summit Hill Exit) ] Text End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 144 661 155 636 Line %I 1 End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 372 662 361 634 Line %I 1 End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2379 ] concat %I 752 660 736 634 Line %I 1 End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2378 ] concat %I 186 466 160 486 Line %I 1 End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 9.38371e-09 0.771182 -0.771182 9.38371e-09 628.85 -52.2378 ] concat %I 180 576 157 542 Line %I 1 End Begin %I Line %I b 65520 0 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 793 -5.99997 ] concat %I 475 50 328 492 Line %I 1 End Begin %I Line %I b 65520 0 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 793 -5.99994 ] concat %I 475 483 329 589 Line %I 1 End Begin %I Line %I b 65520 0 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 793 -6 ] concat %I 962 483 450 588 Line %I 1 End Begin %I Line %I b 65520 0 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 793 -6 ] concat %I 450 491 474 471 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 456.136 36.7289 ] concat %I [ (To Airport) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 9.63786e-09 0.792068 -0.792068 9.63786e-09 137.137 639.729 ] concat %I [ (To Ashville, Bristol) ] Text End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ -0.288762 0.966405 -0.966405 -0.288762 1070.92 66.5381 ] concat %I 743 193 51 70 Elli End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 600.094 464.292 ] concat %I 11 203 116 212 116 212 140 226 140 226 122 222 122 222 102 212 102 212 110 203 110 203 113 11 Poly End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.0258e-08 0.843029 -0.843029 1.0258e-08 619.253 516.221 ] concat %I 12 183 114 183 124 193 124 193 135 205 135 205 125 248 90 242 83 239 86 231 76 194 106 194 114 12 Poly End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.11572e-08 0.810855 -0.916932 9.86645e-09 659.418 526.458 ] concat %I 8 251 106 251 126 300 126 300 126 300 116 289 116 289 106 289 106 8 Poly End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.63639e-08 1.34483 -1.34483 1.63639e-08 1018.44 -282.452 ] concat %I 791 383 799 395 Rect End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 900.11 -9.08362 ] concat %I 12 776 339 776 352 798 352 803 356 808 350 807 348 812 342 808 336 803 341 789 343 782 333 776 333 12 Poly End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.13408e-08 0.952586 -0.932017 1.1591e-08 882.447 11.9457 ] concat %I 885 359 901 436 Rect End Begin %I Elli %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 1 SetP %I t [ -0.249802 0.836016 -0.836016 -0.249802 1016.8 155.898 ] concat %I 743 193 51 70 Elli End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.75 SetP %I t [ -0.289327 0.966236 -0.966236 -0.289327 1071.6 80.2259 ] concat %I 707 160 754 232 Rect End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 0.240969 0.462035 -0.606024 0.318423 615.648 549.235 ] concat %I 13 164 162 182 162 182 167 235 167 234 162 254 162 254 134 234 133 235 129 183 129 183 134 164 134 164 149 13 Poly End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 9.95725e-09 0.885621 -0.818318 1.07762e-08 598.215 291.204 ] concat %I 385 148 422 197 Rect End Begin %I Pict %I b u %I cfg Black 0 0 0 SetCFg %I cbg u %I f u %I p u %I t [ 1.05665 0 0 1.05665 213.224 -6.32959 ] concat Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg DkGray 0.501961 0.501961 0.501961 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 8.1286e-09 0.496511 -0.668033 6.04153e-09 312.332 583.056 ] concat %I 13 204 92 204 123 226 123 226 113 248 113 248 123 264 123 264 92 248 92 248 95 226 95 226 92 225 92 13 Poly End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg DkGray 0.501961 0.501961 0.501961 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 8.1286e-09 -0.496511 -0.668033 -6.04153e-09 312.332 845.214 ] concat %I 13 204 92 204 123 226 123 226 113 248 113 248 123 264 123 264 92 248 92 248 95 226 95 226 92 225 92 13 Poly End Begin %I Rect none SetB %I b n %I cfg DkGray 0.501961 0.501961 0.501961 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 8.1286e-09 0.496511 -0.668033 6.04153e-09 311.664 582.559 ] concat %I 258 92 274 121 Rect End End %I eop Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 532.584 731.01 ] concat %I [ (Physics) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 555.924 783.814 ] concat %I [ (Geography) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 565.058 787.873 ] concat %I [ (& Geology) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 0.737179 0.804195 -0.804195 0.737179 504.155 697.286 ] concat %I [ (Biology) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 366.956 772.951 ] concat %I [ (13th Street) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 511.152 538.704 ] concat %I [ (Voluteer Boulevard) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 0.46602 0.986397 -0.986397 0.466021 521.533 631.373 ] concat %I [ (Middle Way) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 373.044 540.56 ] concat %I [ (16th Street) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 363.799 672.151 ] concat %I [ (Walters) (Life) (Sciences) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 537.034 899.133 ] concat %I [ (Daughtery) (Engineering) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.60693e-08 1.32061 -1.32061 1.60693e-08 657.627 707.825 ] concat %I [ (Neyland) (Stadium) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.0536 -0.283024 0.283024 -1.0536 624.893 639.464 ] concat %I [ (Stadium Drive) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 450.384 486.441 ] concat %I [ (Library) ] Text End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.645 -1.75617 ] concat %I 4 483 431 523 431 523 391 481 389 4 Poly End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 412.866 555.498 ] concat %I [ (University) ( Center) ] Text End Begin %I BSpl %I b 65520 2 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.438 0.0836792 ] concat %I 6 753 467 837 468 841 464 846 464 843 420 841 419 6 BSpl %I 1 End Begin %I BSpl %I b 65520 2 0 0 [12 4] 17 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.921 0.182861 ] concat %I 6 788 313 830 316 840 348 841 386 843 417 843 418 6 BSpl %I 1 End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.435 0.47699 ] concat %I 887 450 839 438 Line %I 1 End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.589 0.222778 ] concat %I 9 807 460 838 460 838 405 816 404 816 394 831 394 831 379 807 378 806 379 9 Poly End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 895.838 0.0155029 ] concat %I 796 369 781 389 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 523.481 806.106 ] concat %I [ (South) (College) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-bold-r-*-120-* Courier-Bold 12 SetF %I t [ 1.21714e-08 1.00029 -1.00029 1.21714e-08 439.801 711.069 ] concat %I [ (Ayres Hall) ] Text End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.07786e-08 0.885813 -0.885813 1.07786e-08 771.842 61.8701 ] concat %I 943 284 915 303 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 459.421 900.429 ] concat %I [ (Dabney/) (Buhler) ] Text End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.07786e-08 0.885813 -0.885813 1.07786e-08 767.413 57.441 ] concat %I 698 424 663 374 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-times-bold-r-*-140-* Times-Bold 14 SetF %I t [ 1.22729e-08 1.00862 -1.00862 1.22729e-08 464.596 743.589 ] concat %I [ (X) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09087 0.0128026 -0.0128027 -1.09087 503.693 612.151 ] concat %I [ (Stadium Drive) ] Text End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p < cc cc 33 33 cc cc 33 33 > -1 SetP %I t [ 1.07786e-08 0.885813 -0.885813 1.07786e-08 855.109 167.282 ] concat %I 465 390 498 427 Rect End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 6.73658e-10 0.0553633 -0.0553633 6.73658e-10 494.25 582.008 ] concat %I 11 509 1018 237 1018 237 1114 -35 1114 -35 1018 -307 1018 -307 746 -35 746 -35 378 509 378 509 380 11 Poly End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ 1.32746e-08 1.09095 -1.09095 1.32746e-08 485.903 693.513 ] concat %I [ (Psychology) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-bold-r-*-120-* Courier-Bold 12 SetF %I t [ 9.87788e-09 0.811792 -0.811792 9.87788e-09 526.522 575.313 ] concat %I [ (Parking) (Garage) ] Text End Begin %I Line %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p < cc cc 33 33 cc cc 33 33 > -1 SetP %I t [ 1.07786e-08 0.885813 -0.885813 1.07786e-08 774.943 159.309 ] concat %I 481 283 491 300 Line %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 368.693 657.005 ] concat %I [ (15th Street) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -1.09095 2.65492e-08 -2.65492e-08 -1.09095 374.042 893.422 ] concat %I [ (11th Street) ] Text End Begin %I BSpl %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.07785e-08 0.885813 -0.885813 1.07785e-08 786.901 202.714 ] concat %I 18 476 304 494 301 501 299 527 283 545 269 569 255 590 244 611 241 648 238 672 234 686 219 705 203 741 204 767 217 776 236 774 284 776 343 773 538 18 BSpl %I 1 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.07785e-08 0.885813 -0.885813 1.07785e-08 786.901 202.714 ] concat %I 376 429 375 538 Line %I 1 End Begin %I BSpl %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.07785e-08 0.885813 -0.885813 1.07785e-08 786.901 202.714 ] concat %I 8 376 429 375 275 369 238 352 223 345 219 323 208 303 203 303 204 8 BSpl %I 1 End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.07785e-08 0.885813 -0.885813 1.07785e-08 786.901 202.714 ] concat %I 4 301 50 850 50 850 537 301 537 4 Poly End Begin %I BSpl %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 793 -5.99995 ] concat %I 5 814 52 872 66 910 82 947 107 961 127 5 BSpl %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.520591 0.958718 -0.958718 -0.520591 712.173 866.792 ] concat %I [ (Neyland Drive) ] Text End Begin %I BSpl %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 653.5 ] concat %I 24 459 412 422 314 224 273 224 233 224 152 305 80 345 72 386 31 418 -33 418 -122 394 -316 474 -461 652 -542 854 -566 1015 -566 1152 -501 1184 -445 1176 -203 1168 15 1031 233 789 314 547 314 432 351 428 350 24 BSpl %I 8 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 653.5 ] concat %I 916 415 915 1204 Line %I 8 End Begin %I BSpl %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 620.75 ] concat %I 5 486 195 402 153 394 72 402 -73 394 -73 5 BSpl %I 8 End Begin %I BSpl %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 604.375 ] concat %I 11 133 415 132 396 110 602 231 715 316 751 330 765 387 857 351 977 344 1126 358 1041 351 1190 11 BSpl %I 8 End Begin %I MLine %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 604.375 ] concat %I 3 133 410 153 -454 613 -2226 3 MLine %I 8 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 461.75 604.375 ] concat %I 663 276 137 276 Line %I 8 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 527 522.5 ] concat %I 96 105 808 79 Line %I 8 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 527 522.5 ] concat %I 95 105 -425 89 Line %I 8 End Begin %I Line %I b 65535 3 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.52099e-09 0.125 -0.125 1.52099e-09 483.5 457 ] concat %I 99 587 3984 588 Line %I 8 End Begin %I BSpl %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 6.91136e-09 0.653301 -0.567997 7.94935e-09 450.164 -70.1652 ] concat %I 15 211 347 224 320 247 286 278 265 315 254 368 252 499 251 582 255 629 257 783 266 863 302 880 318 903 384 900 434 898 545 15 BSpl %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-bold-r-*-120-* Courier-Bold 12 SetF %I t [ 1.2168e-08 1 -1 1.2168e-08 288.5 532.5 ] concat %I [ (Jack Dongarra's office in Ayres Hall Room 107) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ -0.792068 1.92757e-08 -1.92757e-08 -0.792068 428.69 73.2605 ] concat %I [ (Airport/Alcoa Highway) ] Text End Begin %I Rect %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p < 88 44 22 11 88 44 22 11 > -1 SetP %I t [ 1.2168e-08 1 -1 1.2168e-08 610 221 ] concat %I 258 413 267 424 Rect End Begin %I Rect %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p < 88 44 22 11 88 44 22 11 > -1 SetP %I t [ 1.2168e-08 1 -1 1.2168e-08 593 221 ] concat %I 258 413 267 424 Rect End Begin %I MLine %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 773 -23 ] concat %I 3 558 602 521 602 507 598 3 MLine %I 1 End Begin %I Line %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 773 -23 ] concat %I 326 526 326 467 Line %I 1 End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.2168e-08 1 -1 1.2168e-08 773 -23 ] concat %I 564 368 588 385 Rect End Begin %I Rect %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 1.2168e-08 1 -1 1.2168e-08 773 26.9999 ] concat %I 564 368 588 385 Rect End Begin %I Poly %I b 65535 2 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg %I p 0.5 SetP %I t [ 8.96587e-09 0.736842 -0.736842 8.96587e-09 708.816 120.079 ] concat %I 4 616 414 631 414 631 431 616 431 4 Poly End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.855758 0.676646 -0.676646 -0.855758 381.905 599.524 ] concat %I [ (Law Builfinh) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.855758 0.676646 -0.676646 -0.855758 383.905 547.524 ] concat %I [ (Pan-Helenic Bldg.) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-80-* Courier 8 SetF %I t [ -0.855758 0.676646 -0.676646 -0.855758 384.905 572.524 ] concat %I [ (International House.) ] Text End Begin %I MLine %I b 65535 0 0 0 [] 0 SetB %I cfg Black 0 0 0 SetCFg %I cbg White 1 1 1 SetCBg none SetP %I p n %I t [ 1.2168e-08 1 -1 1.2168e-08 773 -23 ] concat %I 2 559 582 507 581 2 MLine %I 1 End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 186.965 539.756 ] concat %I [ (Ramada Inn) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-medium-r-*-100-* Courier 10 SetF %I t [ 1.13387e-08 0.931845 -0.931845 1.13387e-08 167.965 538.756 ] concat %I [ (Hilton) ] Text End Begin %I Text %I cfg Black 0 0 0 SetCFg %I f *-courier-bold-r-*-120-* Courier-Bold 12 SetF %I t [ 1.2168e-08 1 -1 1.2168e-08 514.5 56.5 ] concat %I [ (Directions from the airport to Ayres Hall:) () ( Alcoa Highway North to Cumberland Avenue) () ( Cumberland Avenue east to Stadium Drive) ( \(Stadium Dr. is accross from 15th St.\)) () ( Park at Parking Garage and walk up hill) ( to largest building, Ayres Hall) ] Text End End %I eop showpage %%Trailer end From owner-pbwg-comm@CS.UTK.EDU Mon Feb 22 10:36:00 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA05312; Mon, 22 Feb 93 10:36:00 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA28038; Mon, 22 Feb 93 10:34:47 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 22 Feb 1993 10:34:46 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from antares.mcs.anl.gov by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA28025; Mon, 22 Feb 93 10:34:42 -0500 Received: from jadoube.mcs.anl.gov by antares.mcs.anl.gov with SMTP id AA23129 (5.65c/IDA-1.4.4 for ); Mon, 22 Feb 1993 09:34:38 -0600 From: David Levine Received: by jadoube.mcs.anl.gov (4.1/GeneV4) id AA11892; Mon, 22 Feb 93 09:34:37 CST Date: Mon, 22 Feb 93 09:34:37 CST Message-Id: <9302221534.AA11892@jadoube.mcs.anl.gov> To: pbwg-comm@cs.utk.edu send archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Mon Feb 22 15:29:10 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA15809; Mon, 22 Feb 93 15:29:10 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15065; Mon, 22 Feb 93 15:27:47 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 22 Feb 1993 15:27:45 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from adios.brl.mil by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15057; Mon, 22 Feb 93 15:27:44 -0500 Date: Mon, 22 Feb 93 15:24:05 EST From: TomC To: pbwg-comm@CS.UTK.EDU Cc: TomC , coleman@BRL.MIL, apress@BRL.MIL Subject: Additions to pbwg maillist Message-Id: <9302221524.aa18550@ADIOS.BRL.MIL> Please include the following in the above mail list crimmins@brl.mil Tom Crimmins 410-278-6267 monte@brl.mil Monte Coleman 410-278-6261 apress@brl.mil Tony Pressley 410-278-6509 Thanks, Tom Crimmins From owner-pbwg-comm@CS.UTK.EDU Mon Feb 22 17:04:11 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA19520; Mon, 22 Feb 93 17:04:11 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA19497; Mon, 22 Feb 93 17:03:39 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 22 Feb 1993 17:03:38 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from climate.engin.umich.edu by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA19489; Mon, 22 Feb 93 17:03:36 -0500 Received: by climate.engin.umich.edu (5.67/1.35) id AA21283; Mon, 22 Feb 93 17:05:11 -0500 Date: Mon, 22 Feb 93 17:05:11 -0500 From: Hal G Marshall Message-Id: <9302222205.AA21283@climate.engin.umich.edu> To: pbwg-comm@cs.utk.edu Cc: idaho@CS.UTK.EDU Please add me to your list for MIMD benchmarks. -Hal From owner-pbwg-comm@CS.UTK.EDU Mon Mar 1 05:14:41 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA28056; Mon, 1 Mar 93 05:14:41 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15881; Mon, 1 Mar 93 05:13:28 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 1 Mar 1993 05:13:27 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from Mail.Think.COM by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15867; Mon, 1 Mar 93 05:13:24 -0500 Received: from Godot.Think.COM by mail.think.com; Mon, 1 Mar 93 05:13:22 -0500 Received: by godot.think.com (4.1/Think-1.2) id AA16751; Mon, 1 Mar 93 05:14:28 EST Message-Id: <9303011014.AA16751@godot.think.com> To: Jack Dongarra Cc: pbwg-comm@cs.utk.edu Subject: Re: Parallel Benchmark Working Group Meeting In-Reply-To: Your message of "Sun, 14 Feb 93 16:56:01 EST." <9302142156.AA01941@dasher.cs.utk.edu> Date: Mon, 01 Mar 93 05:14:28 EST From: Dennis Parkinson I apologize I will not be able to attend the meeting sorry for the late notifia`cation dennis Parkinson From owner-pbwg-comm@CS.UTK.EDU Mon Mar 1 07:39:16 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA25550; Mon, 1 Mar 93 07:39:16 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA25188; Mon, 1 Mar 93 07:38:13 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 1 Mar 1993 07:38:12 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from inetg1.Arco.COM by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA25180; Mon, 1 Mar 93 07:38:11 -0500 Received: by Arco.COM (4.1/SMI-4.1) id AA15125; Mon, 1 Mar 93 06:38:08 CST Date: Mon, 1 Mar 93 06:38:08 CST From: ccm@Arco.COM (Chuck Mosher (214)754-6468) Message-Id: <9303011238.AA15125@Arco.COM> To: pbwg-comm@cs.utk.edu Subject: Addition to the List Please add me to the parallel benchmark working group mailing list. Regards, Chuck Mosher ccm@arco.com From owner-pbwg-comm@CS.UTK.EDU Mon Mar 1 10:49:36 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA29573; Mon, 1 Mar 93 10:49:36 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA05800; Mon, 1 Mar 93 10:48:55 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 1 Mar 1993 10:48:54 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from ben.uknet.ac.uk by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA05792; Mon, 1 Mar 93 10:48:45 -0500 Received: from eros.uknet.ac.uk by ben.uknet.ac.uk via UKIP with SMTP (PP) id ; Mon, 1 Mar 1993 15:48:23 +0000 Received: from parsys.co.uk by eros.uknet.ac.uk with UUCP id <9375-0@eros.uknet.ac.uk>; Mon, 1 Mar 1993 15:48:01 +0000 Message-Id: <2594.9303011242@boston.parsys.co.uk> Received: from parsys.co.uk (gmail) by boston.parsys.co.uk; Mon, 1 Mar 93 12:42:29 GMT Date: 1 Mar 1993 12:23:18 +0000 From: Heather Cooke Subject: Re: Parallel Benchmark Worki To: dongarra@cs.utk.edu, System_Manager@parsys.co.uk Cc: pbwg-comm@cs.utk.edu Mail*Link(r) SMTP RE>Parallel Benchmark Worki Received: by parsys.co.uk with SMTP;1 Mar 1993 11:49:21 +0000 Received: from uknet.UUCP by boston.parsys.co.uk; Mon, 1 Mar 93 11:45:35 GMT Received: from CS.UTK.EDU by ben.uknet.ac.uk via EUnet with SMTP (PP) id ; Mon, 1 Mar 1993 10:17:20 +0000 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15881; Mon, 1 Mar 93 05:13:28 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 1 Mar 1993 05:13:27 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from Mail.Think.COM by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA15867; Mon, 1 Mar 93 05:13:24 -0500 Received: from Godot.Think.COM by mail.think.com; Mon, 1 Mar 93 05:13:22 -0500 Received: by godot.think.com (4.1/Think-1.2) id AA16751; Mon, 1 Mar 93 05:14:28 EST Message-Id: <9303011014.AA16751@godot.think.com> To: Jack Dongarra Cc: pbwg-comm@cs.utk.edu Subject: Re: Parallel Benchmark Working Group Meeting In-Reply-To: Your message of "Sun, 14 Feb 93 16:56:01 EST." <9302142156.AA01941@dasher.cs.utk.edu> Date: Mon, 01 Mar 93 05:14:28 EST From: Dennis Parkinson I apologize I will not be able to attend the meeting sorry for the late notifia`cation dennis Parkinson From owner-pbwg-comm@CS.UTK.EDU Wed Mar 3 13:13:23 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA05199; Wed, 3 Mar 93 13:13:23 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA25195; Wed, 3 Mar 93 13:12:14 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Wed, 3 Mar 1993 13:12:12 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from orion.oac.uci.edu by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA25186; Wed, 3 Mar 93 13:12:09 -0500 Received: from balboa.eng.uci.edu by orion.oac.uci.edu with SMTP id AA14344 (5.65c/IDA-1.4.4 for ); Wed, 3 Mar 1993 10:12:00 -0800 Received: from seal.eng.uci.edu by balboa.eng.uci.edu (4.1/SMI-4.1) id AA03526; Wed, 3 Mar 93 10:12:02 PST From: mnyeu@balboa.eng.uci.edu (Maung Nyeu) Message-Id: <9303031812.AA03526@balboa.eng.uci.edu> Subject: Parallel Benhmarking Meetinh To: pbwg-comm@cs.utk.edu Date: Wed, 3 Mar 93 10:10:46 PST X-Mailer: ELM [version 2.3 PL11] Hi, I am student at UC Irvine dept. of comp science. I am interested to get the summary or the papers presented in the parallel benchmarking meeting. I would appreciate it if you send me a reply. Thanks. -- Maung Ting Nyeu email: mnyeu@balboa.eng.uci.edu From owner-pbwg-comm@CS.UTK.EDU Thu Mar 4 12:21:42 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA02456; Thu, 4 Mar 93 12:21:42 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA06000; Thu, 4 Mar 93 12:20:35 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Thu, 4 Mar 1993 12:20:33 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from timbuk.cray.com by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA05984; Thu, 4 Mar 93 12:20:31 -0500 Received: from groucho (groucho.cray.com) by cray.com (4.1/CRI-MX 2.13) id AA06976; Thu, 4 Mar 93 11:20:29 CST Received: by groucho (4.1/CRI-5.13) id AA16192; Thu, 4 Mar 93 11:20:26 CST From: kkg@ferrari.cray.com (Koushik Ghosh) Message-Id: <9303041720.AA16192@groucho> Subject: Meeting on May 24, 1993 To: pbwg-comm@cs.utk.edu Date: Thu, 4 Mar 93 11:20:25 CST X-Mailer: ELM [version 2.3 PL11] There seems to be a lot of requests for a May 24, 1993 meeting. May I suggest that if we start the meeting at 8 AM instead of 9 AM, the chances of getting everything done by 5PM might go up. Then Jack and some of us might be able to catch a flight out of Knoxville that evening. -- Koushik Ghosh kkg@cray.com Benchmarking Dept. Cray Research Inc. 1-612-683-3407 From owner-pbwg-comm@CS.UTK.EDU Thu Mar 4 13:48:21 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA05140; Thu, 4 Mar 93 13:48:21 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA11235; Thu, 4 Mar 93 13:47:13 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Thu, 4 Mar 1993 13:47:10 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from gemini.npac.syr.EDU by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA11220; Thu, 4 Mar 93 13:47:05 -0500 Received: by gemini.npac.syr.edu id AA20470 (5.65c/IDA-1.4.4 for pbwg-comm@cs.utk.edu); Thu, 4 Mar 1993 13:46:43 -0500 Date: Thu, 4 Mar 1993 13:46:43 -0500 From: Tomasz Haupt Message-Id: <199303041846.AA20470@gemini.npac.syr.edu> To: pbwg-comm@cs.utk.edu Subject: Meeting on May 24, 1993 Cc: haupt@npac.syr.edu > There seems to be a lot of requests for a May 24, 1993 meeting. Fine with me. Tom Haupt From owner-pbwg-comm@CS.UTK.EDU Fri Mar 5 13:15:58 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA04015; Fri, 5 Mar 93 13:15:58 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA28005; Fri, 5 Mar 93 13:14:48 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Fri, 5 Mar 1993 13:14:46 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from vnet.ibm.com by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA27996; Fri, 5 Mar 93 13:14:32 -0500 Message-Id: <9303051814.AA27996@CS.UTK.EDU> Received: from KGNVMZ by vnet.ibm.com (IBM VM SMTP V2R2) with BSMTP id 6262; Fri, 05 Mar 93 13:11:37 EST Date: Fri, 5 Mar 93 12:45:31 EST From: "Dr. Joanne L. Martin" To: pbwg-comm@cs.utk.edu Subject: send archive from pbwg send archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Tue Mar 9 15:08:55 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA13515; Tue, 9 Mar 93 15:08:55 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA24030; Tue, 9 Mar 93 15:01:29 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Tue, 9 Mar 1993 15:01:28 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from THUD.CS.UTK.EDU by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA24023; Tue, 9 Mar 93 15:01:27 -0500 From: Jack Dongarra Received: by thud.cs.utk.edu (5.61++/2.7c-UTK) id AA14193; Tue, 9 Mar 93 15:01:25 -0500 Date: Tue, 9 Mar 93 15:01:25 -0500 Message-Id: <9303092001.AA14193@thud.cs.utk.edu> To: pbwg-comm@cs.utk.edu Subject: PBWG posting Forwarding: Mail from '' dated: Tue, 9 Mar 93 14:30:05 -0500 Here is a note I'm planning on posting to comp.parallel. Let me know if you see any problems. I will post it on Friday. 3/12. Regards, Jack Dear Colleague, We are planning to have the Third Meeting of the Parallel Benchmark Working Group meet in Knoxville, Tennessee at the University of Tennessee on May 24th, 1993. This process formally began with a workshop held at the Supercomputer '92 meeting in November 1992. The purpose of the working group is to establish credible and useful benchmarks for the evaluation of Distributed Memory MIMD systems. The objectives for the group are: 1. To establish a comprehensive set of parallel benchmarks that is generally accepted by both users and vendors of parallel system. 2. To provide a focus for parallel benchmark activities and avoid unnecessary duplication of effort and proliferation of benchmarks. 3. To set standards for benchmarking methodology and result-reporting together with a control database/repository for both the benchmarks and the results. Mode of Working: The working group has adopted an HPF-like forum style of proceedings with a view to convergence to an agreed set of benchmarks and procedures within 10 months. If you would like to participate and attend the meeting let me know. Mailing Lists ============= The following mailing lists have been set up. pbwg-comm@cs.utk.edu Whole committee pbwg-lowlevel@cs.utk.edu Low level subcommittee pbwg-compactapp@cs.utk.edu Compact applications subcommittee pbwg-method@cs.utk.edu Methodology subcommittee pbwg-kernel@cs.utk.edu Kernel subcommittee If you are on a mailing list you will receive mail as it is posted. If you want to join a mailing list send me mail (dongarra@cs.utk.edu). All mail will be collected and can be retrieved by sending email to netlib@ornl.gov and in the mail message typing: send comm.archive from pbwg send lowlevel.archive from pbwg send compactapp.archive from pbwg send method.archive from pbwg send kernel.archive from pbwg send index from pbwg The various subcommittees will look into the following topics: Low-Level: --------- Start-up, latency, bandwidth Reduction (broadcast, sum, gather/scatter) Synchronization (e.g., SYNCH1 from Genesis) I/O Kernel: ------ Matrix operations (e.g., multiply, transpose) LU Decomposition PDE Solvers (Red/Black Relaxation) Multigrid FFT Conjugate Gradient Compact Applications: -------------------- Particle-In-Cell codes (e.g., LPM1 from Genesis) QCD Molecular Dynamics CFD ARCO Financial Applications Methodology: ------------ Guidelines for reporting performance. The meeting site will be the: University Center in room 221 University of Tennessee We have made arrangements with the Hilton Hotel in Knoxville. Hilton Hotel 501 W. Church Street Knoxville, TN Phone: 615-523-2300 When making arrangements tell the hotel you are associated with the Parallel Benchmarking Meeting. You can rent a car or get a cab from the airport to the hotel. >From the hotel to the University it is a 15 minute walk. We should plan to start at 8:30 pm May 24th and finish about 5:00 pm. The format of the meeting is: Monday 25th May 8.30 - 12.00 Full group meeting 12.00 - 1.30 Formal lunch 1.30 - 4.00 Parallel subgroup meetings 5.00 - 5.00 Full group meeting Tentative agenda for full group meeting: 1. Minutes of Minneapolis meeting 2. Reports and discussion from subgroups 3. Open discussion and agreement on further actions 4. Date and venue for next meeting Suggested subgroups - probably two in parallel Compact Applications Low-Level benchmarks and second pair: Kernels benchmarks Methodology We have setup a mail refector for correspondence, it is called pbwg-comm@cs.utk.edu. Mail to that address will be sent to the mailing list and also collected in netlib@ornl.gov. To retrieve the collected mail, send email to netlib@ornl.gov and in the mail message type: send comm.archive from pbwg If you would like to be put on the mailing list to receive the correspondence let me know. Regards, Jack Dongarra ----------- End Forwarded Message ----------- From owner-pbwg-comm@CS.UTK.EDU Tue Mar 9 17:22:57 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA15724; Tue, 9 Mar 93 17:22:57 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA01135; Tue, 9 Mar 93 17:21:35 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Tue, 9 Mar 1993 17:21:32 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from BERRY.CS.UTK.EDU by CS.UTK.EDU with SMTP (5.61++/2.8s-UTK) id AA01122; Tue, 9 Mar 93 17:21:29 -0500 From: Mike Berry Received: by berry.cs.utk.edu (5.61++/2.7c-UTK) id AA24206; Tue, 9 Mar 93 17:19:08 -0500 Date: Tue, 9 Mar 93 17:19:08 -0500 Message-Id: <9303092219.AA24206@berry.cs.utk.edu> To: pbwg-comm@cs.utk.edu Subject: PBWG Minutes (March 1-2, 1993) Minutes of the 2nd Meeting of the Parallel Benchmark Working Group (PBWG) ------------------------------------------------------------------------- Place: Room 221, University Center University of Tennessee Knoxville, TN Host: Jack Dongarra ORNL/Univ. of Tennessee Dates: March 1-2, 1993 Attendees/Affiliations: ---------------------- David Bailey, NASA Michael Berry, Univ. of Tennessee Ed Brocklehurst, National Physical Lab Jack Dongarra, Univ. of Tennessee / ORNL Koushik Ghosh, Cray Research Tom Haupt, Syracuse Univ. Tony Hey, Southampton Univ. Roger Hockney, Southampton Univ. Brian LaRose, Univ. of Tennessee David Mackay, Intel SSD Joanne Martin, IBM Robert Pennington, Pittsburgh Supercomputing Center David Walker, ORNL Pearl Wang, George Mason Univ. / US Geological Survey Agenda: March 1, 1993 --------------------- At 1:00 pm EDT, Roger Hockney gave opening remarks and welcomed all participants to the workshop. Each participant introduced him or herself by affiliation and interests. The minutes of the previous meeting at Supercomputing'92 (Minneapolis) were reviewed. Roger suggested that the group consider an alternative name for the group such as "PARABEN" or "INTERBEN". Participants were asked to think about this over dinner and cocktails that evening. Jack Dongarra mentioned that the mail reflector "pbwg-com@cs.utk.edu" is now in place and can be used as an electronic forum for the group. Currently there are about 50 persons on the mailing list. Jack also mentioned that anyone can obtain all correspondence information concerning PBWG requests and solicitations by sending the mail message "send archive from pbwg" to netlib@ornl.gov. To inquire about the contents of the archive simply send the message "send index from pbwg" to netlib@ornl.gov. It was noted that this meeting was posted in the the "comp.parallel" users group (USENET). Roger Hockney reviewed the 4 subgroups previously defined for the effort: Methodology, Classification, Compiler-based, Kernels. Roger proposed that the participants not break into subgroups due to the somewhat lower than expected turnout for the workshop. All participants agreed and a general discussion on "Methodology" followed. David Bailey was then asked to review the "Proposed Guidelines for Reporting Performance" which he distributed to the group. [These guidelines will be made available from the pbwg archive in netlib.] Roger Hockney then distributed an article entitled "A framework for benchmark performance analysis" from Supercomputer 48, IX-2 (March 1992). He recommended that the group start using correct performance symbols such as "Mflop/s" for millions of floating-point operations per second rather than simply "Mflops" which is somewhat ambiguous. As discussed in the paper, temporal performance measures such as T(N,P) and R_T=T^-1(N,P) were also suggested. Roger pointed out that the numerator term, F(N), in the performance measure R_B = F(N)/T(N,P) should be constant and clearly stated. Here, T(N,P) is the execution time, R_T is the inverse of execution time, N is the problem size, P is the number of processors, and F(N) is the total flop count. He suggested that the group steer away from measures such as speedup and efficiency. David Bailey pointed out that one must be prepared to change F(N) to match optimal implementations on different machines. Jack D. pointed out that one must be careful though not to allow performance rates which exceed peak performance. Roger H. suggested that the performance database might provide plots of the metrics as functions of N and P. Along with the distribution of the paper by Roger Hockney, Tony Hey distributed copies of 2 papers on the Genesis Benchmarks: "The Genesis Distributed-Memory Benchmarks" (Tech Report, Dept. of Electronics and Computer Science, Univ. of Southhampton, UK), and "The Genesis distributed-memory benchmarks. Part 1: methodology and general relativity benchmark with results for the SUPRENUM computer" (from Concurrency: Practice and Experience Vol. 5(1), pp. 1-22, February 1993.) Tony indicated that the Genesis benchmarks will be sent to Jack D. for inclusion in the benchmark repository in NETLIB. Ed Brocklehurst raised the concern that the European PEPS project methodology might differ from that of PBWG. He stressed the need for collaboration among the two efforts. He then gave a short lecture on the PEPS project which is a 3-year funded program ($11 million) which began in November of 1992. Participants of the PEPS project include: TS-ASM (electronics firm - France), NPL, Inecs Fistemi (Italy), Simulog (France), Sosip (Germany, France, Italy, UK), and Warwick. The benchmarks supported by PEPS are to be available to the public. Other concerns of the project involve performance monitoring and characterization. The discussion on Methodology concluded with a comment by Michael Berry that all accepted codes for the benchmark suite be self-validating. Roger H. then asked David Walker to report on the work thus far by the Classification subgroup. David W. stressed that the classification of benchmarks be based on (i.) communication characteristics, (ii.) I/O, and (iii.) load balance and data layout. Questions on the inclusion of profiles and processor utilization data with the benchmarks were raised. David W. pointed out that benchmarks may range from inherent parallel codes (e.g., Monte-Carlo) to mesh-based codes or gather/scatter-based applications. David Bailey stressed the need for classifying I/O and to see what the current measurement of I/O is in existing benchmark suites. Joanne Martin pointed out that the ARCO benchmark suite is well-received by the seismic community and should be seriously considered by PBWG. The discussion then turned to compiler-based benchmarks with Tom Haupt taking the lead. He distributed the document entitled "High Performance Fortran, Fortran-D Benchmark Suite" which is available vi anonymous ftp from minerva.npac.syr.edu. Tom H. indicated that the Syracuse compiler benchmarks can test parsers on a collection of statements from HPF, accept intrinsic functions, and for-all loops. He stressed that these benchmarks be able to handle hard-coded message-passing as well as HPF Fortran. One goal would be to find instances where HPF is difficult to use. Tom mentioned that these codes deal with parallel I/O, validate HPF Fortran, and generate meaningful statistics. The current Syracuse benchmark suite contains over 50 codes and includes many well-known benchmarks such as LAPACK, NAS, and Genesis benchmarks. Tom agreed to download this suite to machine at UT during the workshop (this was done after the meeting adjourned on March 2 and the files are available in netlib, send index from pbwg, for more details). This suite consumes about 5 megabytes and comprises about 14 directories of README files, makefiles, and source files. Brian LaRose (student of Michael B.) then gave a demonstration of the Performance Database Server (PDS) which is available in an upcoming release of XNETLIB. Beforehand, Jack D. gave a brief introduction of XNETLIB to familiarize participants with the features of the X Windows-based user interface to NETLIB. The members were pleased with the tool and made suggestions for improvements such as graphing data retrieved from the database. Brian demonstrated the various browsing, rank-ordering, and search features of the system he is designing for his Master thesis at the Univ. of Tennessee. Michael B. indicated that participants would be able to test PDS and should send anonymous ftp address/directory information to "utpds@cs.utk.edu" so a beta release of the software could be sent to them. At this point, Roger Hockney asked that Jack D. define/classify the publically-available benchmarks that might be considered for the PBWG benchmark suite. The classification into Kernels, Applications, and Compact Applications is given below. A (*) indicates availability as single-node benchmarks also. Additional benchmarks that were not explicitly classified are listed under "Miscellaneous". Kernels: ------- Linpack 100 (*) 1000 (*) Livermore Loops (*) Los Alamos (*, some of them) Tom Dunigan's (ORNL) Genesis Global Operations (e.g., sum, gather/scatter) NAS Sorting Applications: ------------ Perfect Slalom RAPS PEPS Compact Applications: -------------------- NAS Genesis Syracuse ARCO Euroben (*) Miscellaneous: ------------ flops.c IOZONE Dhrystones Whetstones Xstones Jack D. proposed that the selected codes for the PBWG suite be implemented in Fortran-77, HPF, in PVM for message-passing (this will be converted to MPI, Message Passing Interface, when the standard is complete). Roger H. stressed the need to measure startup latency (alpha, beta), and metrics such as R-sub-infinity and N-sub-one-half. Other target measures discussed included: peak rates, flop/s, communication costs, and bisection (node-node) bandwidth. David Bailey stressed that the single processor benchmarks should be separated from the parallel benchmarks too avoid confusion. At 5:00pm EDT, the workshop adjourned for a 6:30pm EDT dinner appointment at Chesapeakes's, a downtown Knoxville seafood restaurant. Agenda: March 2, 1993 --------------------- At 9:00 am EDT, Roger Hockney opened the morning session of the PBWG workshop in Room 221 of the University Center on the campus of the University of Tennessee. Prior to any technical discussion on the makeup of the PBWG benchmark suite. Participants agreed to hold a "Birds of a Feather" session at the Supercomputing'93 conference in Portland, OR in November. Joanne Martin will schedule this meeting for the group - a 4:30 pm session was agreed upon by the participants. Prior to the Supercomputing'93 meeting, Jack D. agreed to collect all desirable benchmarks and install them in NETLIB for public access. The discussion on the categorical makeup of a PBWG benchmark suite was then lead by Tony Hey. The 3 categories considered were: Low-Level, Kernels, and Compact Applications. David W. stressed that 10 would be an optimal number of codes to use. Participants generally agreed with this proposal. Potential benchmarks considered are listed by category below. Jack D. stressed that each group "justify" benchmarks by indicating exactly what machine feature(s) are to be tested and try to avoid overlap between categories. Low-Level: --------- Start-up, latency, bandwidth Reduction (broadcast, sum, gather/scatter) Synchronization (e.g., SYNCH1 from Genesis) I/O Kernel: ------ Matrix operations (e.g., multiply, transpose) LU Decomposition PDE Solvers (Red/Black Relaxation) Multigrid FFT Conjugate Gradient Compact Applications: -------------------- Particle-In-Cell codes (e.g., LPM1 from Genesis) QCD Molecular Dynamics CFD ARCO Financial Applications At this point, Roger H. asked that subgroups be formed to address the each of the above categories of benchmarks. Each subgroup will be ask to give a report at the Portland Meeting concerning the potential design/makeup of their category of PBWG benchmarks. Participants agreed that the subgroups be open to other experts from the performance community: vendors, academia, and users. A person may be in more than one subgroup and Jack D. suggested that a voting strategy be developed. A "1 vote per institution" was one candidate for consideration but nothing official was decided at the workshop. Suggestions for subgroup members not present at the workshop are included below. A (*) indicates subgroup leader. PBWG Benchmark Suite Subgroups: ------------------------------ Low-Level Kernel Compact Applications Methodology --------- ------ -------------------- ----------- PEPS/Warwick D. Walker D. Walker (*) D. Bailey (*) IBM/D. Frye J. Dongarra J. Martin D. Schneider R. Pennington T. Hey (*) E. Brocklehurst PEPS R. Hockney (*) T. Haupt K. Ghosh J. Dongarra PEPS M. Berry E. Kushner (Intel) D. Bailey D. Bailey P. Wang D. Barton The following mail reflectors have been setup: Low Level: pbwg-lowlevel@cs.utk.edu Kernel: pbwg-kernel@cs.utk.edu Compact Applications: pbwg-compactapp@cs.utk.edu Methodology: pbwg-method@cs.utk.edu Jack D. agreed to set up mail reflectors for each of the subgroups. (These mail reflectors are setup and functioning.) All participants agreed to 2 meetings that would be open to anyone interested in attending prior to the meeting at Supercomputing'93 in Portland. Drafts of the reports from each subgroup are due by the agreed-upon May 24 meeting at the Univ. of Tennessee which will begin at 8:30 am. The second meeting date will also be at the Univ. of Tennessee on August 23-24. Here's the approved schedule of PBWG meetings: Third Meeting: May 24, begins at 8:30am (Univ. of Tennessee) Fourth Meeting: August 23-24 (Univ. of Tennessee) Fifth Meeting: Week of Nov 15 (Supercomputing Portland, OR) Joanne Martin mentioned that there was a Conference on Performance Tools on April 1-4 and Tony H. suggested that PBWG might submit an announcement to that conference about PBWG's activities. A discussion on possible name changes for PBWG was considered again. Some suggestions included PES(Performance or Parallel Evaluation Suite) and PARKBENCH(Parallel Kernel Benchmarks). Since no consensus was reached, the discussion is postponed to the next meeting so members can think of other names. In preparation for the May meeting, subgroup leaders briefly mentioned some of the issues they hope to address: T. Hey : talk to vendors, provide list of kernel areas, justification D. Walker : define areas, acquire candidate codes, address scalability, address problem sizes (e.g., 4 used in Genesis), be able to specify data layouts (use vanilla layouts) R. Hockney: construct list of desirable low-level benchmarks Participants agreed that users should be able to optimize the suite by any means provided they document what they do (e.g., assembly language allowed). When possible, benchmark contributions should include the code that was used to obtain the reported results (stored in PDS). The second day of the PBWG workshop was adjourned by Roger Hockney at 11:00am EDT, and participants had informal discussions before departing the Univ. of Tennessee campus. PBWG activities will be posted to the comp.parallel newsgroup (USENET). From owner-pbwg-comm@CS.UTK.EDU Wed Mar 24 06:54:11 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA05932; Wed, 24 Mar 93 06:54:11 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA11133; Wed, 24 Mar 93 06:53:25 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Wed, 24 Mar 1993 06:53:13 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from sun2.nsfnet-relay.ac.uk by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA11125; Wed, 24 Mar 93 06:53:10 -0500 Via: uk.ac.southampton.ecs; Wed, 24 Mar 1993 11:44:52 +0000 From: R.Hockney@parallel-applications-centre.southampton.ac.uk Via: calvados.pac.soton.ac.uk (plonk); Wed, 24 Mar 93 11:37:49 GMT Date: Wed, 24 Mar 93 11:24:51 GMT Message-Id: <3278.9303241124@calvados.pac.soton.ac.uk> Apparently-To: pbwg-comm@cs.utk.edu From owner-pbwg-comm@CS.UTK.EDU Wed Mar 24 07:10:28 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA09660; Wed, 24 Mar 93 07:10:28 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA11684; Wed, 24 Mar 93 07:09:56 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Wed, 24 Mar 1993 07:09:55 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from sun2.nsfnet-relay.ac.uk by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA11676; Wed, 24 Mar 93 07:09:50 -0500 Via: uk.ac.southampton.ecs; Wed, 24 Mar 1993 12:07:41 +0000 From: R.Hockney@parallel-applications-centre.southampton.ac.uk Via: calvados.pac.soton.ac.uk (plonk); Wed, 24 Mar 93 11:44:01 GMT Date: Wed, 24 Mar 93 11:50:26 GMT Message-Id: <3341.9303241150@calvados.pac.soton.ac.uk> Apparently-To: pbwg-comm@cs.utk.edu ~r comlet1.asc From owner-pbwg-comm@CS.UTK.EDU Wed Mar 24 07:36:42 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA11698; Wed, 24 Mar 93 07:36:42 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA12605; Wed, 24 Mar 93 07:35:41 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Wed, 24 Mar 1993 07:35:40 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from sun2.nsfnet-relay.ac.uk by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA12584; Wed, 24 Mar 93 07:35:34 -0500 Via: uk.ac.southampton.ecs; Wed, 24 Mar 1993 12:32:55 +0000 Via: brewery.ecs.soton.ac.uk; Wed, 24 Mar 93 12:25:44 GMT From: Prof Roger Hockney Received: from alithea.ecs.soton.ac.uk by brewery.ecs.soton.ac.uk; Wed, 24 Mar 93 12:34:08 GMT Date: Wed, 24 Mar 93 12:34:10 GMT Message-Id: <253.9303241234@alithea.ecs.soton.ac.uk> To: pbwg-comm@cs.utk.edu Subject: PBWG Draft Report A Message from Your Chairman ---------------------------- In view of the committee's aim to agree on a draft text at the meeting on May 24, 1993, it would help if each subcommittee produces their recommendations as LATEX files to fit into a standard framework. I believe that LATEX is the most widely used text processing system in the universities, and appears to have been used for the HPF report. I have prepared the following skeleton files for the committee's report: (1) benrep1.tex - control file for main report (2) bencom1.tex - command definition file (3) benref1.bib - start of bibliography Additions to the command file and bibliography can be sent to me as files bencom2.tex, benref2.bib etc. The control file reads the following files which are to be provided by the leader of each subcommittee: (4) intro1.tex - Roger Hockney for whole committee (5) method1.tex - David Bailey for Methodology subcommittee (6) lowlev1.tex - Roger Hockney for the low-level subcommittee (7) kernel1.tex - Tony Hey for the kernel subcommittee (8) compac1.tex - David Walker for compact applications subcommittee (9) conclu1.tex - Roger Hockney for whole committee Files (1), (2), (3), (5), (7), and (8) are appended to this e-mail. Best wishes, Roger Hockney. % file : benrep1.tex % % ************************************************************** % STANDARD INTERNATIONAL BENCHMARKS FOR PARALLEL COMPUTERS % ************************************************************** % \input{bencom1.tex} % define new commands for benchmark report % ---------------------------------------------------------------------------- \documentstyle[]{report} % Specifies the document style. \textheight 8.25 true in \textwidth 5.625 true in \topmargin -0.13 true in \oddsidemargin 0.25 true in \evensidemargin 0.25 true in % The preamble begins here. \title{Standard International Benchmarks for Parallel Computers} % ---------------------------------------------------------------------------- \author{PBWG Committee \\ draft assembled by Roger Hockney (chairman)} \date{23 March 1993 - draft 1} % ---------------------------------------------------------------------------- \begin{document} % End of preamble and beginning of text. \sloppy \maketitle % Produces the title. % ---------------------------------------------------------------------------- \input{intro1.tex} % Introduction % responsibility of Roger Hockney for whole committee % ---------------------------------------------------------------------------- \input{method1.tex} % Chapter1 % responsibility of David Bailey for Methodology subcommittee % ---------------------------------------------------------------------------- \input{lowlev1.tex} % Chapter2 % responsibility of Roger Hockney for Low-level benchmarks subcommittee % ---------------------------------------------------------------------------- \input{kernel1.tex} % Chapter3 % responsibility of Tony Hey for Kernel benchmarks subcommittee % ---------------------------------------------------------------------------- \input{compac1.tex} % Chapter4 % responsibility of David Walker for Compact Applications subcommittee % ---------------------------------------------------------------------------- \input{conclu1.tex} % Conclusions % responsibility of Roger Hockney for whole committee % ---------------------------------------------------------------------------- \vspace{0.35in} {\large \bf Acknowledgments} \bibliography{benref1} \bibliographystyle{unsrt} \end{document} % End of document. % file : bencom1.tex % % ************************************************************** % LATEX COMMANDS FOR PARABEN REPORTS % ************************************************************** % \def\flop{\mathop{\rm flop}\nolimits} \def\pipe{\mathop{\rm pipe}\nolimits} \newcommand{\usec}{\mbox{\rm $\mu$s}} \newcommand{\where}{\mbox{\rm where}} \newcommand{\rmand}{\mbox{\rm and}} \newcommand{\Mflops}{\mbox{\rm Mflop/s}} \newcommand{\flops}{\mbox{\rm flop/s}} \newcommand{\flopB}{\mbox{\rm flop/B}} \newcommand{\tstepps}{\mbox{\rm tstep/s}} \newcommand{\MWps}{\mbox{\rm MW/s}} \newcommand{\Mwps}{\mbox{\rm Mw/s}} \newcommand{\spone}{\mbox{\ }} \newcommand{\sptwo}{\mbox{\ \ }} \newcommand{\spfour}{\mbox{\ \ \ \ }} \newcommand{\spsix}{\mbox{\ \ \ \ \ \ }} \newcommand{\speight}{\mbox{\ \ \ \ \ \ \ \ }} \newcommand{\spten}{\mbox{\ \ \ \ \ \ \ \ \ \ }} \newcommand{\rinf}{\mbox{$r_\infty$}} \newcommand{\Rinf}{\mbox{$R_\infty$}} \newcommand{\nhalf}{\mbox{$n_{\frac{1}{2}}$}} \newcommand{\Nhalf}{\mbox{$N_{\frac{1}{2}}$}} \newcommand{\phalf}{\mbox{$p_{\frac{1}{2}}$}} \newcommand{\Phalf}{\mbox{$P_{\frac{1}{2}}$}} \newcommand{\half}{\mbox{$\frac{1}{2}$}} \newcommand{\rnhalf}{\mbox{(\rinf,\nhalf)}} \newcommand{\RNhalf}{\mbox{(\Rinf,\Nhalf)}} \newcommand{\third}{\mbox{$\frac{1}{3}$}} \newcommand{\quart}{\mbox{$\frac{1}{4}$}} \newcommand{\eighth}{\mbox{$\frac{1}{8}$}} \newcommand{\nineth}{\mbox{$\frac{1}{9}$}} % ---------------------------------------------------------------------------- % file : benref1.bib % @book{HoJe88, author= "Roger W. Hockney and Christopher R. Jesshope", title= "Parallel Computers 2: Architecture, Programming and Algorithms", publisher= "Adam Hilger/IOP Publishing", address= "Bristol \& New York", year= "1988", note= "Distributed in the USA by the American Institute of Physics, c/o AIDC, 64 Depot Road, Colchester, VT 05445."} @book{Super, key="Super", title={Supercomputer}, publisher="ASFRA", address="Edam, Netherlands"} @book{SI75, key="Royal Society", organization="Symbols Committee of the Royal Society", title={Quantities, Units and Symbols}, publisher="The Royal Society", address="London", year=1975} @article{Berr89, author="M. Berry and D. Chen and P. Koss and D. Kuck and S. Lo and Y. Pang and L. Pointer and R. Roloff and A. Sameh and E. Clementi and S. Chin and D. Schneider and G. Fox and P. Messina and D. Walker and C. Hsiung and J. Schwarzmeier and K. Lue and S. Orszag and F. Seidl and O. Johnson and R. Goodrum and J. Martin", title="The {PERFECT} Club benchmarks: effective performance evaluation of computers", journal={Intl. J. Supercomputer Appls.}, volume=3, number=3, year=1989, pages="5-40"} @incollection{Ma88, author="F. H. McMahon", title="The {L}ivermore {F}ortran Kernels test of the numerical performance range", editor="J. L. Martin", booktitle={Performance Evaluation of Supercomputers}, publisher="Elsevier Science B.V., North-Holland", address="Amsterdam", year=1988, pages="143-186"} @article{Mess90, author="P. Messina and C. Baillie and E. Felten and P. Hipes and R. Williams and A. Alagar and A. Kamrath and R. Leary and W. Pfeiffer and J. Rogers and D. Walker", title="Benchmarking advanced architecture computers", journal={Concurrency: Practice and Experience}, volume=2, number=3, year=1990, pages="195-255"} @inproceedings{Cvet90, author="Z. Cvetanovic and E. G. Freedman and C. Nofsinger", title="Efficient decomposition and performance of parallel {PDE}, {FFT}, {M}onte-{C}arlo simulations, simplex and sparse solvers", booktitle={Proceedings Supercomputing90}, publisher="IEEE", address="New York", year=1990, pages="465-474"} @article{SUPR88, title="Proceedings 2nd International SUPRENUM Colloquium", author="U. Trottenberg", journal={Parallel Computing}, volume=7, number=3, year=1988} @article{Hey91, author="A. J. G. Hey", title="The {G}enesis Distributed-Memory Benchmarks", journal={Parallel Computing}, volume=17, year=1991, pages="1275-1283"} @book{F90, author="M. Metcalf and J. Reid", title={Fortran-90 Explained}, publisher="Oxford Science Publications/OUP", address="Oxford and New York", year=1990, chapter=6} @article{SPEC90, key="SPEC", title="{SPEC} Benchmarks Suite Release 1.0", journal={SPEC Newslett.}, volume=2, number=3, year=1990, pages="3-4", publisher="Systems Performance Evaluation Cooperative, Waterside Associates", address="Fremont, California"} @article{FGHS89, author="A. Friedli and W. Gentzsch and R. Hockney and A. van der Steen", title="A {E}uropean Supercomputer Benchmark Effort", journal={Supercomputer 34}, volume="VI", number=6, year=1989, pages="14-17"} @article{BRH90, author="L. Bomans and D. Roose and R. Hempel", title="The {A}rgonne/{GMD} Macros in {F}ortran for portable parallel programming and their implementation on the {I}ntel i{PSC}/2", journal={Parallel Computing}, volume=15, year=1990, pages="119-132"} @inproceedings{ShTu91, author="J. N. Shahid and R. S. Tuminaro", title="Iterative Methods for Nonsymmetric Systems on {MIMD} Machines", booktitle={Proc. Fifth SIAM Conf. Parallel Processing for Scientific Computing}, year=1991} @article{Bish90, author="N. T. Bishop and C. J. S. Clarke and R. A. d'Inverno", journal={Classical and Quantum Gravity}, volume=7, year=1990, pages="L23-L27"} @article{Isaac83, author="R. A. Isaacson and J. S. Welling and J.Winicour", journal={J. Math. Phys.}, volume=24, year=1983, pages="1824-1834"} @article{Stew82, author="J. M. Stewart and H. Friedrich", journal={Proc. Roy. Soc.}, volume="A384", year=1982, pages="427-454"} @article{Hoc92, author="R. W. Hockney", title="A framework for benchmark analysis", journal={Supercomputer}, volume=48, number="IX-2", year=1992, pages="9-22"} @article{Add93, author="C. Addison and J. Allwright and N. Binsted and N. Bishop and B. Carpenter and P. Dalloz and D. Gee and V. Getov and A. Hey and R. Hockney and M. Lemke and J. Merlin and M. Pinches and C. Scott and I. Wolton", title="The {G}enesis distributed-memory benchmarks. Part 1: methodology and general relativity benchmark with results for the {SUPRENUM} computer", journal={Concurrency: Practice and Experience}, volume=5, number=1, year=1993, pages="1-22"} % % ------------------------------------------------------------------------- %file: intro1.tex \chapter{Introduction}\footnotemark \footnotetext{written by Roger Hockney for whole committee} \cite{HoJe88} % % ------------------------------------------------------------------------- %file: method1.tex \chapter{Methodology} \footnote{assembled by David Bailey for Methodology subcommittee} % % ------------------------------------------------------------------------- %file: lowlev1.tex \chapter{Low-Level Benchmarks} \footnote{assembled by Roger Hockney for low-level subcommittee} % % ------------------------------------------------------------------------- %file: kernel1.tex \chapter{Kernel Benchmarks} \footnote{assembled by Tony Hey for Kernel subcommittee} % % ------------------------------------------------------------------------- %file: compac1.tex \chapter{Compact Applications} \footnote{assembled by David Walker for Compact Applications subcommittee} % % ------------------------------------------------------------------------- %file: conclu1.tex \chapter{Conclusions} \footnote{written by Roger Hockney for whole committee} % % ------------------------------------------------------------------------- % End of skeleton paper From owner-pbwg-comm@CS.UTK.EDU Mon Mar 29 12:54:46 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA19359; Mon, 29 Mar 93 12:54:46 -0500 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA04636; Mon, 29 Mar 93 12:52:54 -0500 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 29 Mar 1993 12:52:53 EST Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from gemini.npac.syr.EDU by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA04626; Mon, 29 Mar 93 12:52:42 -0500 Received: from localhost by gemini.npac.syr.edu with SMTP id AA11964 (5.65c/IDA-1.4.4 for pbwg-comm@cs.utk.edu); Mon, 29 Mar 1993 12:52:02 -0500 Message-Id: <199303291752.AA11964@gemini.npac.syr.edu> To: Prof Roger Hockney Cc: pbwg-comm@cs.utk.edu, haupt@npac.syr.edu Subject: Re: PBWG Draft Report In-Reply-To: Your message of "Wed, 24 Mar 93 12:34:10 GMT." <253.9303241234@alithea.ecs.soton.ac.uk> Date: Mon, 29 Mar 93 12:51:54 -0500 From: haupt@npac.syr.edu X-Mts: smtp I thought we agreed to add additional section to the draft on compiler benchmark. Syracuse volunteers to coordinate that effort, and to provide the text for the draft. Will you add this section to the draft, please? Tom Haupt From owner-pbwg-comm@CS.UTK.EDU Tue Apr 27 03:01:42 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA29378; Tue, 27 Apr 93 03:01:42 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA20445; Tue, 27 Apr 93 03:00:40 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Tue, 27 Apr 1993 03:00:39 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from swiba9.unibas.ch by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA20392; Tue, 27 Apr 93 03:00:32 -0400 Received: from sir.ifi.unibas.ch by swiba9.unibas.ch with SMTP (PP) id <6953-0@swiba9.unibas.ch>; Tue, 27 Apr 1993 09:00:04 +0200 Received: from charlie by sir.ifi.unibas.ch (NX5.67c/NX3.0M) id AA07530; Tue, 27 Apr 93 08:59:58 +0200 From: (Walter Kuhn) kuhn@ifi.unibas.ch Message-Id: <9304270659.AA07530@sir.ifi.unibas.ch> Received: by charlie.ifi.unibas.ch (NX5.67c/NX3.0X) id AA00297; Tue, 27 Apr 93 08:59:58 +0200 Date: Tue, 27 Apr 93 08:59:58 +0200 To: pbwg-comm@CS.UTK.edu Subject: Additions to pbwg maillist Please add me to your mail list. Walter From owner-pbwg-comm@CS.UTK.EDU Fri Apr 30 12:32:36 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA18590; Fri, 30 Apr 93 12:32:36 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA22997; Fri, 30 Apr 93 12:31:04 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Fri, 30 Apr 1993 12:30:51 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from sun2.nsfnet-relay.ac.uk by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA22949; Fri, 30 Apr 93 12:30:47 -0400 Via: uk.ac.southampton.ecs; Fri, 30 Apr 1993 16:11:19 +0100 From: R.Hockney@parallel-applications-centre.southampton.ac.uk Via: calvados.pac.soton.ac.uk (plonk); Fri, 30 Apr 93 16:03:41 BST Date: Fri, 30 Apr 93 15:10:51 GMT Message-Id: <15929.9304301510@calvados.pac.soton.ac.uk> To: pbwg-comm@cs.utk.edu Subject: Revised Report Framework Updated Report Framework and Bibliography ----------------------------------------- I append to this note updated versions of benrep1.tex and benref1.bib, there is an added chapter on Compiler benchmarks and additional references. I have sent a contribution to Chapter 2 to pbwg-method and a draft of Chapter 3 to pbwg-lowlevel. I await drafts of the other chapters from the different subcommittee leaders: Roger Hockney ------------------------------- cut here -------------------------------- % % ************************************************************** % STANDARD INTERNATIONAL BENCHMARKS FOR PARALLEL COMPUTERS % ************************************************************** % \input{bencom1.tex} % define new commands for benchmark report % ---------------------------------------------------------------------------- \documentstyle[]{report} % Specifies the document style. \textheight 8.25 true in \textwidth 5.625 true in \topmargin -0.13 true in \oddsidemargin 0.25 true in \evensidemargin 0.25 true in % The preamble begins here. \title{Standard International Benchmarks for Parallel Computers} % ---------------------------------------------------------------------------- \author{PBWG Committee \\ draft assembled by Roger Hockney (chairman)} \date{19 April 1993 - draft 2} % ---------------------------------------------------------------------------- \begin{document} % End of preamble and beginning of text. \sloppy \maketitle % Produces the title. % ---------------------------------------------------------------------------- \input{intro1.tex} % Introduction % responsibility of Roger Hockney for whole committee % ---------------------------------------------------------------------------- \input{method3.tex} % Chapter1 % responsibility of David Bailey for Methodology subcommittee % ---------------------------------------------------------------------------- \input{lowlev1.tex} % Chapter2 % responsibility of Roger Hockney for Low-level benchmarks subcommittee % ---------------------------------------------------------------------------- \input{kernel1.tex} % Chapter3 % responsibility of Tony Hey for Kernel benchmarks subcommittee % ---------------------------------------------------------------------------- \input{compac1.tex} % Chapter4 % responsibility of David Walker for Compact Applications subcommittee % ---------------------------------------------------------------------------- \input{compil1.tex} % Chapter5 % responsibility of Tom Haupt for Compiler Benchmarks subcommittee % ---------------------------------------------------------------------------- \input{conclu1.tex} % Conclusions % responsibility of Roger Hockney for whole committee % ---------------------------------------------------------------------------- \vspace{0.35in} {\large \bf Acknowledgments} \bibliography{benref1} \bibliographystyle{unsrt} \end{document} % End of document. % @book{HoJe81, author= "Roger W. Hockney and Christopher R. Jesshope", title= "Parallel Computers: Architecture, Programming and Algorithms", publisher= "Adam Hilger", address= "Bristol", year= "1981", } @book{HoJe88, author= "Roger W. Hockney and Christopher R. Jesshope", title= "Parallel Computers 2: Architecture, Programming and Algorithms", publisher= "Adam Hilger/IOP Publishing", address= "Bristol \& Philadelphia", year= "1988", edition="second", note= "Distributed in the USA by IOP Publ. Inc., Public Ledger Bldg., Suite 1035, Independence Square, Philadelphia, PA 19106."} @book{Super, key="Super", title={Supercomputer}, publisher="ASFRA", address="Edam, Netherlands"} @book{SI75, key="Royal Society", organization="Symbols Committee of the Royal Society", title={Quantities, Units and Symbols}, publisher="The Royal Society", address="London", year=1975} @article{Berr89, author="M. Berry and D. Chen and P. Koss and D. Kuck and S. Lo and Y. Pang and L. Pointer and R. Roloff and A. Sameh and E. Clementi and S. Chin and D. Schneider and G. Fox and P. Messina and D. Walker and C. Hsiung and J. Schwarzmeier and K. Lue and S. Orszag and F. Seidl and O. Johnson and R. Goodrum and J. Martin", title="The {PERFECT} Club benchmarks: effective performance evaluation of computers", journal={Intl. J. Supercomputer Appls.}, volume=3, number=3, year=1989, pages="5-40"} @incollection{Ma88, author="F. H. McMahon", title="The {L}ivermore {F}ortran {K}ernels test of the numerical performance range", editor="J. L. Martin", booktitle={Performance Evaluation of Supercomputers}, publisher="Elsevier Science B.V., North-Holland", address="Amsterdam", year=1988, pages="143-186"} @article{Mess90, author="P. Messina and C. Baillie and E. Felten and P. Hipes and R. Williams and A. Alagar and A. Kamrath and R. Leary and W. Pfeiffer and J. Rogers and D. Walker", title="Benchmarking advanced architecture computers", journal={Concurrency: Practice and Experience}, volume=2, number=3, year=1990, pages="195-255"} @inproceedings{Cvet90, author="Z. Cvetanovic and E. G. Freedman and C. Nofsinger", title="Efficient decomposition and performance of parallel {PDE}, {FFT}, {M}onte-{C}arlo simulations, simplex and sparse solvers", booktitle={Proceedings Supercomputing90}, publisher="IEEE", address="New York", year=1990, pages="465-474"} @article{SUPR88, title="Proceedings 2nd International SUPRENUM Colloquium", author="U. Trottenberg", journal={Parallel Computing}, volume=7, number=3, year=1988} @article{Hey91, author="A. J. G. Hey", title="The {G}enesis Distributed-Memory Benchmarks", journal={Parallel Computing}, volume=17, year=1991, pages="1275-1283"} @book{F90, author="M. Metcalf and J. Reid", title={Fortran-90 Explained}, publisher="Oxford Science Publications/OUP", address="Oxford and New York", year=1990, chapter=6} @article{SPEC90, key="SPEC", title="{SPEC} Benchmarks Suite Release 1.0", journal={SPEC Newslett.}, volume=2, number=3, year=1990, pages="3-4", publisher="Systems Performance Evaluation Cooperative, Waterside Associates", address="Fremont, California"} @article{FGHS89, author="A. Friedli and W. Gentzsch and R. Hockney and A. van der Steen", title="A {E}uropean Supercomputer Benchmark Effort", journal={Supercomputer 34}, volume="VI", number=6, year=1989, pages="14-17"} @article{BRH90, author="L. Bomans and D. Roose and R. Hempel", title="The {A}rgonne/{GMD} Macros in {F}ortran for portable parallel programming and their implementation on the {I}ntel i{PSC}/2", journal={Parallel Computing}, volume=15, year=1990, pages="119-132"} @inproceedings{ShTu91, author="J. N. Shahid and R. S. Tuminaro", title="Iterative Methods for Nonsymmetric Systems on {MIMD} Machines", booktitle={Proc. Fifth SIAM Conf. Parallel Processing for Scientific Computing}, year=1991} @article{Bish90, author="N. T. Bishop and C. J. S. Clarke and R. A. d'Inverno", journal={Classical and Quantum Gravity}, volume=7, year=1990, pages="L23-L27"} @article{Isaac83, author="R. A. Isaacson and J. S. Welling and J.Winicour", journal={J. Math. Phys.}, volume=24, year=1983, pages="1824-1834"} @article{Stew82, author="J. M. Stewart and H. Friedrich", journal={Proc. Roy. Soc.}, volume="A384", year=1982, pages="427-454"} @incollection{Hoc77, author="R. W. Hockney", title="Super-Computer Architecture", editor="F. Sumner", booktitle={Infotech State of the Art Conference: {F}uture {S}ystems}, publisher="Infotech", address="Maidenhead", year=1977, pages="277-305"} @article{Hoc82, author="R. W. Hockney", title="Characterization of parallel computers and algorithms", journal={Computer Physics Communications}, volume=26, year=1982, pages="285-291"} @article{Hoc83, author="R. W. Hockney", title="Characterizing Computers and Optimizing the {FACR}(l) Poisson-Solver on Parallel Unicomputers", journal={IEEE Trans. Comput.}, volume="{C}\-32", year=1983, pages="933-941"} @article{Hoc87, author="R. W. Hockney", title="Parametrization of Computer Performance", journal={Parallel Computing}, volume=5, year=1987, pages="97-103"} @article{Hoc88, author="R. W. Hockney", title="Synchronization and Communication Overheads on the {LCAP} Multiple {FPS}-164 Computer System", journal={Parallel Computing}, volume=9, year=1988, pages="279-290"} @article{HoCu89, author="R. W. Hockney and I. J. Curington", title="$f_{frac{1}{2}}$: a Parameter to Characterise Memory and Communication Bottlenecks", journal={Parallel Computing}, volume=10, year=1989, pages="277-286"} @article{Hoc91, author="R. W. Hockney", title="Performance Parameters and Benchmarking of Supercomputers", journal={Parallel Computing}, volume=17, year=1991, pages="1111-1130"} @article{Hoc92, author="R. W. Hockney", title="A framework for benchmark analysis", journal={Supercomputer}, volume=48, number="IX-2", year=1992, pages="9-22"} @article{HoCa92, author="R. W. Hockney and E. A. Carmona", title="Comparison of Communications on the {I}ntel i{PSC}/860 and {T}ouchstone {D}elta", journal={Parallel Computing}, volume=18, year=1992, pages="1067-1072"} @article{Add93, author="C. Addison and J. Allwright and N. Binsted and N. Bishop and B. Carpenter and P. Dalloz and D. Gee and V. Getov and A. Hey and R. Hockney and M. Lemke and J. Merlin and M. Pinches and C. Scott and I. Wolton", title="The {G}enesis distributed-memory benchmarks. Part 1: methodology and general relativity benchmark with results for the {SUPRENUM} computer", journal={Concurrency: Practice and Experience}, volume=5, number=1, year=1993, pages="1-22"} @techreport{StRi93, author="A. J. van der Steen and P. P. M. de Rijk", title="Guidelines for use of the {E}uro{B}en Benchmark", institution="{E}uro{B}en", year=1993, month=feb, type="Technical Report", number="{TR}\-3", address="The {E}uro{B}en Group, {U}trecht, {T}he {N}etherlands"} From owner-pbwg-comm@CS.UTK.EDU Fri Apr 30 12:49:20 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA20526; Fri, 30 Apr 93 12:49:20 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA24215; Fri, 30 Apr 93 12:48:17 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Fri, 30 Apr 1993 12:48:16 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from vnet.ibm.com by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA24207; Fri, 30 Apr 93 12:48:15 -0400 Message-Id: <9304301648.AA24207@CS.UTK.EDU> Received: from KGNVMZ by vnet.IBM.COM (IBM VM SMTP V2R2) with BSMTP id 8031; Fri, 30 Apr 93 12:48:11 EDT Date: Fri, 30 Apr 93 12:47:34 EDT From: "Dr. Joanne L. Martin ((914) 385-9572)" To: pbwg-comm@cs.utk.edu Subject: New address Please note that my new e-mail address is jmartin at vnet.ibm.com Thanks, Joanne. From owner-pbwg-comm@CS.UTK.EDU Mon May 3 11:05:02 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA17408; Mon, 3 May 93 11:05:02 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02905; Mon, 3 May 93 11:03:37 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 3 May 1993 11:03:35 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from swiba9.unibas.ch by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02897; Mon, 3 May 93 11:03:28 -0400 Received: from sir.ifi.unibas.ch by swiba9.unibas.ch with SMTP (PP) id <21115-0@swiba9.unibas.ch>; Mon, 3 May 1993 17:03:12 +0200 Received: from charlie by sir.ifi.unibas.ch (NX5.67c/NX3.0M) id AA27670; Mon, 3 May 93 17:03:07 +0200 From: (Walter Kuhn) kuhn@ifi.unibas.ch Message-Id: <9305031503.AA27670@sir.ifi.unibas.ch> Received: by charlie.ifi.unibas.ch (NX5.67c/NX3.0X) id AA00509; Mon, 3 May 93 17:03:05 +0200 Date: Mon, 3 May 93 17:03:05 +0200 Received: by NeXT.Mailer (1.87.1) Received: by NeXT Mailer (1.87.1) To: pbwg-comm@cs.utk.edu Subject: send lowlevel.archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Mon May 3 11:05:13 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA17427; Mon, 3 May 93 11:05:13 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02925; Mon, 3 May 93 11:04:06 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 3 May 1993 11:04:05 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from swiba9.unibas.ch by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02909; Mon, 3 May 93 11:03:56 -0400 Received: from sir.ifi.unibas.ch by swiba9.unibas.ch with SMTP (PP) id <21120-0@swiba9.unibas.ch>; Mon, 3 May 1993 17:03:33 +0200 Received: from charlie by sir.ifi.unibas.ch (NX5.67c/NX3.0M) id AA27674; Mon, 3 May 93 17:03:30 +0200 From: (Walter Kuhn) kuhn@ifi.unibas.ch Message-Id: <9305031503.AA27674@sir.ifi.unibas.ch> Received: by charlie.ifi.unibas.ch (NX5.67c/NX3.0X) id AA00514; Mon, 3 May 93 17:03:30 +0200 Date: Mon, 3 May 93 17:03:30 +0200 Received: by NeXT.Mailer (1.87.1) Received: by NeXT Mailer (1.87.1) To: pbwg-comm@cs.utk.edu Subject: send method.archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Mon May 3 11:05:38 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA17445; Mon, 3 May 93 11:05:38 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02939; Mon, 3 May 93 11:04:35 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 3 May 1993 11:04:33 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from swiba9.unibas.ch by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02928; Mon, 3 May 93 11:04:24 -0400 Received: from sir.ifi.unibas.ch by swiba9.unibas.ch with SMTP (PP) id <21132-0@swiba9.unibas.ch>; Mon, 3 May 1993 17:04:14 +0200 Received: from charlie by sir.ifi.unibas.ch (NX5.67c/NX3.0M) id AA27678; Mon, 3 May 93 17:04:10 +0200 From: (Walter Kuhn) kuhn@ifi.unibas.ch Message-Id: <9305031504.AA27678@sir.ifi.unibas.ch> Received: by charlie.ifi.unibas.ch (NX5.67c/NX3.0X) id AA00519; Mon, 3 May 93 17:04:10 +0200 Date: Mon, 3 May 93 17:04:10 +0200 Received: by NeXT.Mailer (1.87.1) Received: by NeXT Mailer (1.87.1) To: pbwg-comm@cs.utk.edu Subject: send compil.archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Mon May 3 11:05:58 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA17467; Mon, 3 May 93 11:05:58 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02984; Mon, 3 May 93 11:05:07 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Mon, 3 May 1993 11:05:05 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from swiba9.unibas.ch by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA02948; Mon, 3 May 93 11:04:53 -0400 Received: from sir.ifi.unibas.ch by swiba9.unibas.ch with SMTP (PP) id <21137-0@swiba9.unibas.ch>; Mon, 3 May 1993 17:04:34 +0200 Received: from charlie by sir.ifi.unibas.ch (NX5.67c/NX3.0M) id AA27683; Mon, 3 May 93 17:04:30 +0200 From: (Walter Kuhn) kuhn@ifi.unibas.ch Message-Id: <9305031504.AA27683@sir.ifi.unibas.ch> Received: by charlie.ifi.unibas.ch (NX5.67c/NX3.0X) id AA00525; Mon, 3 May 93 17:04:30 +0200 Date: Mon, 3 May 93 17:04:30 +0200 Received: by NeXT.Mailer (1.87.1) Received: by NeXT Mailer (1.87.1) To: pbwg-comm@cs.utk.edu Subject: send conclu.archive from pbwg From owner-pbwg-comm@CS.UTK.EDU Wed May 5 09:08:29 1993 Received: from CS.UTK.EDU by surfer.EPM.ORNL.GOV (5.61/1.34) id AA15181; Wed, 5 May 93 09:08:29 -0400 Received: from localhost by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA05335; Wed, 5 May 93 06:47:02 -0400 X-Resent-To: pbwg-comm@CS.UTK.EDU ; Wed, 5 May 1993 06:46:59 EDT Errors-To: owner-pbwg-comm@CS.UTK.EDU Received: from THUD.CS.UTK.EDU by CS.UTK.EDU with SMTP (5.61+IDA+UTK-930125/2.8s-UTK) id AA05327; Wed, 5 May 93 06:46:53 -0400 From: Jack Dongarra Received: by thud.cs.utk.edu (5.61+IDA+UTK-930125/2.7c-UTK) id AA08869; Wed, 5 May 93 06:46:47 -0400 Date: Wed, 5 May 93 06:46:47 -0400 Message-Id: <9305051046.AA08869@thud.cs.utk.edu> To: pbwg-comm@cs.utk.edu Subject: PBWG Meeting May 24th Dear Colleague, We are planning to have the Third Meeting of the Parallel Benchmark Working Group meet in Knoxville, Tennessee at the University of Tennessee on May 24th, 1993. This process formally began with a workshop held at the Supercomputer '92 meeting in November 1992. The purpose of the working group is to establish credible and useful benchmarks for the evaluation of Distributed Memory MIMD systems. The objectives for the group are: 1. To establish a comprehensive set of parallel benchmarks that is generally accepted by both users and vendors of parallel system. 2. To provide a focus for parallel benchmark activities and avoid unnecessary duplication of effort and proliferation of benchmarks. 3. To set standards for benchmarking methodology and result-reporting together with a control database/repository for both the benchmarks and the results. Mode of Working: The working group has adopted an HPF-like forum style of proceedings with a view to convergence to an agreed set of benchmarks and procedures within 10 months. If you would like to participate and attend the meeting let me know. Mailing Lists ============= The following mailing lists have been set up. pbwg-comm@cs.utk.edu Whole committee pbwg-lowlevel@cs.utk.edu Low level subcommittee pbwg-compactapp@cs.utk.edu Compact applications subcommittee pbwg-method@cs.utk.edu Methodology subcommittee pbwg-kernel@cs.utk.edu Kernel subcommittee If you are on a mailing list you will receive mail as it is posted. If you want to join a mailing list send me mail (dongarra@cs.utk.edu). All mail will be collected and can be retrieved by sending email to netlib@ornl.gov and in the mail message typing: send comm.archive from pbwg send lowlevel.archive from pbwg send compactapp.archive from pbwg send method.archive from pbwg send kernel.archive from pbwg send index from pbwg The various subcommittees will look into the following topics: Low-Level: --------- Start-up, latency, bandwidth Reduction (broadcast, sum, gather/scatter) Synchronization (e.g., SYNCH1 from Genesis) I/O Kernel: ------ Matrix operations (e.g., multiply, transpose) LU Decomposition PDE Solvers (Red/Black Relaxation) Multigrid FFT Conjugate Gradient Compact Applications: -------------------- Particle-In-Cell codes (e.g., LPM1 from Genesis) QCD Molecular Dynamics CFD ARCO Financial Applications Methodology: ------------ Guidelines for reporting performance. The meeting site will be the: Science Alliance Conference Room South College University of Tennessee (A postscript map in included at the end of this message, South College is the building located next to Ayres Hall.) We have made arrangements with the Hilton Hotel in Knoxville. Hilton Hotel 501 W. Church Street Knoxville, TN Phone: 615-523-2300 When making arrangements tell the hotel you are associated with the Parallel Benchmarking Meeting. The rate is $65.00/night. You can rent a car or get a cab from the airport to the hotel. From the hotel to the University it is a 15 minute walk. We should plan to start at 8:30 am May 24th and finish about 5:00 pm. The format of the meeting is: Monday 24th May 8.30 - 12.00 Full group meeting 12.00 - 1.30 Lunch 1.30 - 4.00 Parallel subgroup meetings 4.00 - 5.00 Full group meeting Tentative agenda for full group meeting: 1. Minutes of Minneapolis meeting 2. Reports and discussion from subgroups 3. Open discussion and agreement on further actions 4. Date and venue for next meeting Suggested subgroups - probably two in parallel Compact Applications Low-Level benchmarks and second pair: Kernels benchmarks Methodology We have setup a mail refector for correspondence, it is called pbwg-comm@cs.utk.edu. Mail to that address will be sent to the mailing list and also collected in netlib@ornl.gov. To retrieve the collected mail, send email to netlib@ornl.gov and in the mail message type: send comm.archive from pbwg If you would like to be put on the mailing list to receive the correspondence let me know. Regards, Jack Dongarra ---postscript map of the University of Tennessee--- %!PS-Adobe-2.0 EPSF-1.2 %%DocumentFonts: Helvetica-Bold Courier Courier-Bold Times-Bold %%Pages: 1 %%BoundingBox: 39 -113 604 767 %%EndComments /arrowHeight 10 def /arrowWidth 5 def /IdrawDict 54 dict def IdrawDict begin /reencodeISO { dup dup findfont dup length dict begin { 1 index /FID ne { def }{ pop pop } ifelse } forall /Encoding ISOLatin1Encoding def currentdict end definefont } def /ISOLatin1Encoding [ /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /space/exclam/quotedbl/numbersign/dollar/percent/ampersand/quoteright /parenleft/parenright/asterisk/plus/comma/minus/period/slash /zero/one/two/three/four/five/six/seven/eight/nine/colon/semicolon /less/equal/greater/question/at/A/B/C/D/E/F/G/H/I/J/K/L/M/N /O/P/Q/R/S/T/U/V/W/X/Y/Z/bracketleft/backslash/bracketright /asciicircum/underscore/quoteleft/a/b/c/d/e/f/g/h/i/j/k/l/m /n/o/p/q/r/s/t/u/v/w/x/y/z/braceleft/bar/braceright/asciitilde /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef /.notdef/dotlessi/grave/acute/circumflex/tilde/macron/breve /dotaccent/dieresis/.notdef/ring/cedilla/.notdef/hungarumlaut /ogonek/caron/space/exclamdown/cent/sterling/currency/yen/brokenbar /section/dieresis/copyright/ordfeminine/guillemotleft/logicalnot /hyphen/registered/macron/degree/plusminus/twosuperior/threesuperior /acute/mu/paragraph/periodcentered/cedilla/onesuperior/ordmasculine /guillemotright/onequarter/onehalf/threequarters/questiondown /Agrave/Aacute/Acircumflex/Atilde/Adieresis/Aring/AE/Ccedilla /Egrave/Eacute/Ecircumflex/Edieresis/Igrave/Iacute/Icircumflex /Idieresis/Eth/Ntilde/Ograve/Oacute/Ocircumflex/Otilde/Odieresis /multiply/Oslash/Ugrave/Uacute/Ucircumflex/Udieresis/Yacute /Thorn/germandbls/agrave/aacute/acircumflex/atilde/adieresis /aring/ae/ccedilla/egrave/eacute/ecircumflex/edieresis/igrave /iacute/icircumflex/idieresis/eth/ntilde/ograve/oacute/ocircumflex /otilde/odieresis/divide/oslash/ugrave/uacute/ucircumflex/udieresis /yacute/thorn/ydieresis ] def /Helvetica-Bold reencodeISO def /Courier reencodeISO def /Courier-Bold reencodeISO def /Times-Bold reencodeISO def /none null def /numGraphicParameters 17 def /stringLimit 65535 def /Begin { save numGraphicParameters dict begin } def /End { end restore } def /SetB { dup type /nulltype eq { pop false /brushRightArrow idef false /brushLeftArrow idef true /brushNone idef } { /brushDashOffset idef /brushDashArray idef 0 ne /brushRightArrow idef 0 ne /brushLeftArrow idef /brushWidth idef false /brushNone idef } ifelse } def /SetCFg { /fgblue idef /fggreen idef /fgred idef } def /SetCBg { /bgblue idef /bggreen idef /bgred idef } def /SetF { /printSize idef /printFont idef } def /SetP { dup type /nulltype eq { pop true /patternNone idef } { dup -1 eq { /patternGrayLevel idef /patternString idef } { /patternGrayLevel idef } ifelse false /patternNone idef } ifelse } def /BSpl { 0 begin storexyn newpath n 1 gt { 0 0 0 0 0 0 1 1 true subspline n 2 gt { 0 0 0 0 1 1 2 2 false subspline 1 1 n 3 sub { /i exch def i 1 sub dup i dup i 1 add dup i 2 add dup false subspline } for n 3 sub dup n 2 sub dup n 1 sub dup 2 copy false subspline } if n 2 sub dup n 1 sub dup 2 copy 2 copy false subspline patternNone not brushLeftArrow not brushRightArrow not and and { ifill } if brushNone not { istroke } if 0 0 1 1 leftarrow n 2 sub dup n 1 sub dup rightarrow } if end } dup 0 4 dict put def /Circ { newpath 0 360 arc patternNone not { ifill } if brushNone not { istroke } if } def /CBSpl { 0 begin dup 2 gt { storexyn newpath n 1 sub dup 0 0 1 1 2 2 true subspline 1 1 n 3 sub { /i exch def i 1 sub dup i dup i 1 add dup i 2 add dup false subspline } for n 3 sub dup n 2 sub dup n 1 sub dup 0 0 false subspline n 2 sub dup n 1 sub dup 0 0 1 1 false subspline patternNone not { ifill } if brushNone not { istroke } if } { Poly } ifelse end } dup 0 4 dict put def /Elli { 0 begin newpath 4 2 roll translate scale 0 0 1 0 360 arc patternNone not { ifill } if brushNone not { istroke } if end } dup 0 1 dict put def /Line { 0 begin 2 storexyn newpath x 0 get y 0 get moveto x 1 get y 1 get lineto brushNone not { istroke } if 0 0 1 1 leftarrow 0 0 1 1 rightarrow end } dup 0 4 dict put def /MLine { 0 begin storexyn newpath n 1 gt { x 0 get y 0 get moveto 1 1 n 1 sub { /i exch def x i get y i get lineto } for patternNone not brushLeftArrow not brushRightArrow not and and { ifill } if brushNone not { istroke } if 0 0 1 1 leftarrow n 2 sub dup n 1 sub dup rightarrow } if end } dup 0 4 dict put def /Poly { 3 1 roll newpath moveto -1 add { lineto } repeat closepath patternNone not { ifill } if brushNone not { istroke } if } def /Rect { 0 begin /t exch def /r exch def /b exch def /l exch def newpath l b moveto l t lineto r t lineto r b lineto closepath patternNone not { ifill } if brushNone not { istroke } if end } dup 0 4 dict put def /Text { ishow } def /idef { dup where { pop pop pop } { exch def } ifelse } def /ifill { 0 begin gsave patternGrayLevel -1 ne { fgred bgred fgred sub patternGrayLevel mul add fggreen bggreen fggreen sub patternGrayLevel mul add fgblue bgblue fgblue sub patternGrayLevel mul add setrgbcolor eofill } { eoclip originalCTM setmatrix pathbbox /t exch def /r exch def /b exch def /l exch def /w r l sub ceiling cvi def /h t b sub ceiling cvi def /imageByteWidth w 8 div ceiling cvi def /imageHeight h def bgred bggreen bgblue setrgbcolor eofill fgred fggreen fgblue setrgbcolor w 0 gt h 0 gt and { l b translate w h scale w h true [w 0 0 h neg 0 h] { patternproc } imagemask } if } ifelse grestore end } dup 0 8 dict put def /istroke { gsave brushDashOffset -1 eq { [] 0 setdash 1 setgray } { brushDashArray brushDashOffset setdash fgred fggreen fgblue setrgbcolor } ifelse brushWidth setlinewidth originalCTM setmatrix stroke grestore } def /ishow { 0 begin gsave fgred fggreen fgblue setrgbcolor /fontDict printFont printSize scalefont dup setfont def /descender fontDict begin 0 [FontBBox] 1 get FontMatrix end transform exch pop def /vertoffset 1 printSize sub descender sub def { 0 vertoffset moveto show /vertoffset vertoffset printSize sub def } forall grestore end } dup 0 3 dict put def /patternproc { 0 begin /patternByteLength patternString length def /patternHeight patternByteLength 8 mul sqrt cvi def /patternWidth patternHeight def /patternByteWidth patternWidth 8 idiv def /imageByteMaxLength imageByteWidth imageHeight mul stringLimit patternByteWidth sub min def /imageMaxHeight imageByteMaxLength imageByteWidth idiv patternHeight idiv patternHeight mul patternHeight max def /imageHeight imageHeight imageMaxHeight sub store /imageString imageByteWidth imageMaxHeight mul patternByteWidth add string def 0 1 imageMaxHeight 1 sub { /y exch def /patternRow y patternByteWidth mul patternByteLength mod def /patternRowString patternString patternRow patternByteWidth getinterval def /imageRow y imageByteWidth mul def 0 patternByteWidth imageByteWidth 1 sub { /x exch def imageString imageRow x add patternRowString putinterval } for } for imageString end } dup 0 12 dict put def /min { dup 3 2 roll dup 4 3 roll lt { exch } if pop } def /max { dup 3 2 roll dup 4 3 roll gt { exch } if pop } def /midpoint { 0 begin /y1 exch def /x1 exch def /y0 exch def /x0 exch def x0 x1 add 2 div y0 y1 add 2 div end } dup 0 4 dict put def /thirdpoint { 0 begin /y1 exch def /x1 exch def /y0 exch def /x0 exch def x0 2 mul x1 add 3 div y0 2 mul y1 add 3 div end } dup 0 4 dict put def /subspline { 0 begin /movetoNeeded exch def y exch get /y3 exch def x exch get /x3 exch def y exch get /y2 exch def x exch get /x2 exch def y exch get /y1 exch def x exch get /x1 exch def y exch get /y0 exch def x exch get /x0 exch def x1 y1 x2 y2 thirdpoint /p1y exch def /p1x exch def x2 y2 x1 y1 thirdpoint /p2y exch def /p2x exch def x1 y1 x0 y0 thirdpoint p1x p1y midpoint /p0y exch def /p0x exch def x2 y2 x3 y3 thirdpoint p2x p2y midpoint /p3y exch def /p3x exch def movetoNeeded { p0x p0y moveto } if p1x p1y p2x p2y p3x p3y curveto end } dup 0 17 dict put def /storexyn { /n exch def /y n array def /x n array def n 1 sub -1 0 { /i exch def y i 3 2 roll put x i 3 2 roll put } for } def %%EndProlog %%BeginIdrawPrologue /arrowhead { 0 begin transform originalCTM itransform /taily exch def /tailx exch def transform originalCTM itransform /tipy exch def /tipx exch def /dy tipy taily sub def /dx tipx tailx sub def /angle dx 0 ne dy 0 ne or { dy dx atan } { 90 } ifelse def gsave originalCTM setmatrix tipx tipy translate angle rotate newpath arrowHeight neg arrowWidth 2 div moveto 0 0 lineto arrowHeight neg arrowWidth 2 div neg lineto patternNone not { originalCTM setmatrix /padtip arrowHeight 2 exp 0.25 arrowWidth 2 exp mul add sqrt brushWidth mul arrowWidth div def /padtail brushWidth 2 div def tipx tipy translate angle rotate padtip 0 translate arrowHeight padtip add padtail add arrowHeight div dup scale arrowheadpath ifill } if brushNone not { originalCTM setmatrix tipx tipy translate angle rotate arrowheadpath istroke } if grestore end } dup 0 9 dict put def /arrowheadpath { newpath arrowHeight neg arrowWidth 2 div moveto 0 0 lineto arrowHeight neg arrowWidth 2 div neg lineto } def /leftarrow { 0 begin y exch get /taily exch def x exch get /tailx exch def y exch get /tipy exch def x exch get /tipx exch def brushLeftArrow { tipx tipy tailx taily arrowhead } if end } dup 0 4 dict put def /rightarrow { 0 begin y exch get /tipy exch def x exch get /tipx exch def y exch get /taily exch def x exch get /tailx exch def brushRightArrow { tipx tipy tailx taily arrowhead } if end } dup 0 4 dict put def %%EndIdrawPrologue %I Idraw 10 Grid 2.84217e-39 0 %%Page: 1 1 Begin %I b u %I cfg u %I cbg u %I f u %I p u %I t [ 0.799705 0 0 0.799705 0 0 ] concat /originalCTM matrix currentmatrix def Begin %I Pict %I b u %I cfg u %I cbg u %I f u %I p u %I t [ 1 0 0 1 89.1002 831.2 ] concat Begin %I Elli %I b 655