name: Ron Sercely establishment: Convex Computer Corporation email: sercely@convex.com telephone: 214.497.4667 date sent: 29/Nov/95 date benchmarked (DD/MMM/YY): 29/Nov/95 benchmark suite and revision: Parkbench message passing library and revision: PVM 3.3.10.2 (to be released Dec '95) benchmark: poly1 problem size (if applicable): number of nodes: 4 file number if one of a group: manufacturer: Convex model number: Exemplar 1200 cpu: HP-PA/RISC 7200 cpu speed (MHz): 120Mhz fpu: integrated primary cache: 250kbytes secondary cache: none other cache: none memory: 512M disk and connection: SCSI interconnect type: shared memory switching: none other hardware: none operating system and version: SPP-UX 3.2 (to be released Dec ' 95) compilers, libraries and versions: Convex FORTRAN, vs 9.3 /usr/convex/cc/cc: 6.3 compiler switches: -O2 (full optimization) additional tuning parameters: other software: additional comments: PARKBENCH Version 1.0 *** UNOFFICIAL release test package *** This run was with unmodified code PVM_ARCH = CSPP PVMBUFSIZE=1000000 ================================================= === === === GENESIS / PARKBENCH Parallel Benchmarks === === === === POLY1 === === === === Program: In-cache (r-hat f-half) === === Version: Standard Fortran 77 === === Author: Roger Hockney === === === ================================================= POLYNOMIAL BENCHMARK FOR FHALF ------------------------------ (DOUBLE PRECISION) This benchmark tests severity of memory bottlenecks by varying the amount of arithmetic per memory reference which is called the computational intensity of the loop. The performance for long loop (vector) lengths, RINF, is represented as : RINF = RHAT/(1 + FHALF/F) where RHAT = peak Mflop/s rate of arithmetic pipeline approached as F goes to infinity and F = computational intensity = ratio floating operations/memory references FHALF = F required to obtain RINF=RHAT/2 The loop executed is polynomial evaluation by Horners rule, F varies as the order increases from 1 to 10 Roger Hockney, August 1989 ------------------------------------------------------------ POLYNOMIALS : A(I)=S0+B(I)*(S1+B(I)*(S2+B(I)*S3)) ICASE= 1 NCASE= 1 NMAX= 10000 NTIMES= 1000000 NPROCS= 1 COMPUTATIONAL INTENSITY = flop per mem ref 1.0 ------------------------------------------ Floating operations per iteration = 2 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 2.704E-07 .000000 .000 .000E+00 .00 7.397759 2.704E-07 0 5 4.691E-07 40.250221 4.441 8.882E+00 .00 21.317017 2.207E-07 0 10 7.369E-07 38.527096 4.147 8.294E+00 5.49 27.139957 2.153E-07 0 20 9.646E-07 55.276791 7.866 1.573E+01 58.29 41.466385 2.846E-07 0 50 1.745E-06 69.484779 11.885 2.377E+01 40.82 57.312935 3.421E-07 0 100 3.247E-06 68.875984 11.663 2.333E+01 20.07 61.591721 3.387E-07 0 200 5.733E-06 73.657074 14.128 2.826E+01 16.00 69.769012 3.836E-07 0 300 8.300E-06 75.274666 15.242 3.048E+01 11.76 72.287910 4.050E-07 0 400 1.152E-05 73.086479 13.280 2.656E+01 12.30 69.429474 3.634E-07 0 600 1.635E-05 74.423866 15.001 3.000E+01 9.80 73.391968 4.031E-07 0 800 2.168E-05 74.902496 15.796 3.159E+01 7.35 73.801155 4.218E-07 0 1000 2.702E-05 75.058678 16.119 3.224E+01 5.70 74.013435 4.295E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 2.0 ------------------------------------------ Floating operations per iteration = 4 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 3.227E-07 .000000 .000 .000E+00 .00 12.396640 3.227E-07 0 5 6.599E-07 47.450840 2.828 1.131E+01 .00 30.309498 2.384E-07 0 10 1.007E-06 52.810207 3.421 1.369E+01 15.48 39.729889 2.591E-07 0 20 1.300E-06 80.089981 7.466 2.987E+01 71.55 61.527122 3.729E-07 0 50 2.623E-06 89.493637 9.255 3.702E+01 34.54 76.260071 4.137E-07 0 100 4.632E-06 94.543953 10.537 4.215E+01 20.14 86.354965 4.458E-07 0 200 8.694E-06 96.767532 11.346 4.538E+01 10.75 92.012634 4.690E-07 0 300 1.251E-05 98.928558 12.417 4.967E+01 9.05 95.885155 5.021E-07 0 400 1.768E-05 95.084755 9.904 3.962E+01 13.90 90.489075 4.166E-07 0 600 2.465E-05 97.740547 12.438 4.975E+01 12.53 97.348534 5.090E-07 0 800 3.283E-05 98.521355 13.405 5.362E+01 9.35 97.472023 5.442E-07 0 1000 4.141E-05 98.165970 12.855 5.142E+01 7.22 96.590645 5.238E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 3.0 ------------------------------------------ Floating operations per iteration = 6 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 3.306E-07 .000000 .000 .000E+00 .00 18.150093 3.306E-07 0 5 7.957E-07 51.596302 1.843 1.106E+01 .00 37.701401 2.143E-07 0 10 1.239E-06 59.742462 2.517 1.510E+01 23.32 48.426662 2.528E-07 0 20 1.754E-06 82.389755 5.143 3.086E+01 57.92 68.396965 3.745E-07 0 50 3.642E-06 92.458015 6.720 4.032E+01 29.04 82.377602 4.361E-07 0 100 6.473E-06 99.020920 8.153 4.892E+01 19.16 92.692619 4.940E-07 0 200 1.208E-05 103.127930 9.480 5.688E+01 12.00 99.305450 5.516E-07 0 300 1.759E-05 105.032478 10.332 6.199E+01 9.07 102.325081 5.902E-07 0 400 2.496E-05 100.671005 7.733 4.640E+01 14.71 96.138321 4.609E-07 0 600 3.476E-05 103.557785 10.271 6.163E+01 13.13 103.560753 5.951E-07 0 800 4.641E-05 104.299957 11.123 6.674E+01 9.67 103.417786 6.399E-07 0 1000 5.847E-05 103.998039 10.689 6.413E+01 7.43 102.616730 6.167E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 4.0 ------------------------------------------ Floating operations per iteration = 8 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 3.613E-07 .000000 .000 .000E+00 .00 22.144981 3.613E-07 0 5 9.846E-07 51.339363 1.318 1.055E+01 .00 40.627323 2.054E-07 0 10 1.690E-06 54.281307 1.533 1.226E+01 9.10 47.334702 2.259E-07 0 20 2.454E-06 73.924438 3.683 2.947E+01 57.19 65.189171 3.986E-07 0 50 4.871E-06 90.578056 6.263 5.011E+01 37.94 82.114861 5.532E-07 0 100 9.106E-06 93.389366 6.875 5.500E+01 19.27 87.858147 5.890E-07 0 200 1.718E-05 96.184601 7.795 6.236E+01 10.85 93.152931 6.483E-07 0 300 2.657E-05 93.400246 6.509 5.207E+01 9.92 90.311157 5.575E-07 0 400 3.377E-05 95.303246 7.706 6.165E+01 10.06 94.770874 6.469E-07 0 600 5.035E-05 96.166267 8.508 6.806E+01 6.93 95.330734 7.078E-07 0 800 6.681E-05 96.572182 9.002 7.202E+01 5.17 95.797432 7.457E-07 0 1000 8.325E-05 96.789421 9.335 7.468E+01 4.06 96.099342 7.716E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 5.0 ------------------------------------------ Floating operations per iteration = 10 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 3.934E-07 .000000 .000 .000E+00 .00 25.420715 3.934E-07 0 5 1.165E-06 51.845615 1.040 1.040E+01 .00 42.922092 2.005E-07 0 10 2.143E-06 51.438797 1.012 1.012E+01 1.30 46.672348 1.968E-07 0 20 3.356E-06 64.640755 2.404 2.404E+01 42.22 59.598495 3.719E-07 0 50 7.282E-06 72.886787 3.701 3.701E+01 23.90 68.666779 5.078E-07 0 100 1.396E-05 74.243454 4.021 4.021E+01 11.83 71.613029 5.416E-07 0 200 2.707E-05 75.299675 4.417 4.417E+01 6.20 73.893463 5.866E-07 0 300 4.011E-05 75.754440 4.665 4.665E+01 4.23 74.786667 6.158E-07 0 400 5.337E-05 75.797424 4.697 4.697E+01 3.00 74.945801 6.197E-07 0 600 7.949E-05 75.988083 4.912 4.912E+01 2.05 75.483597 6.465E-07 0 800 1.067E-04 75.652077 4.410 4.410E+01 2.06 74.990013 5.830E-07 0 1000 1.324E-04 75.780251 4.653 4.653E+01 1.71 75.551262 6.141E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 6.0 ------------------------------------------ Floating operations per iteration = 12 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 4.688E-07 .000000 .000 .000E+00 .00 25.596228 4.688E-07 0 5 1.260E-06 60.660431 1.370 1.644E+01 .00 47.614922 2.710E-07 0 10 1.922E-06 74.879509 2.260 2.712E+01 35.61 62.439220 3.622E-07 0 20 3.217E-06 85.007095 3.162 3.795E+01 28.07 74.609177 4.464E-07 0 50 7.106E-06 90.427254 3.859 4.631E+01 13.97 84.440231 5.121E-07 0 100 1.365E-05 91.321045 4.030 4.837E+01 6.86 87.939049 5.296E-07 0 200 2.687E-05 91.124847 3.971 4.765E+01 3.25 89.310547 5.229E-07 0 300 3.975E-05 91.616562 4.190 5.028E+01 2.50 90.562958 5.488E-07 0 400 5.275E-05 91.833984 4.324 5.188E+01 1.94 90.995071 5.650E-07 0 600 7.886E-05 91.918953 4.402 5.283E+01 1.26 91.299576 5.747E-07 0 800 1.051E-04 91.858551 4.328 5.194E+01 .93 91.319504 5.654E-07 0 1000 1.318E-04 91.629395 3.971 4.765E+01 1.18 91.050354 5.200E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 7.0 ------------------------------------------ Floating operations per iteration = 14 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 4.690E-07 .000000 .000 .000E+00 .00 29.850546 4.690E-07 0 5 1.414E-06 59.250252 .985 1.379E+01 .00 49.499813 2.327E-07 0 10 2.196E-06 73.521530 1.807 2.530E+01 38.06 63.755432 3.441E-07 0 20 3.790E-06 82.019073 2.526 3.536E+01 26.64 73.874184 4.311E-07 0 50 8.479E-06 87.297546 3.188 4.463E+01 13.50 82.557205 5.113E-07 0 100 1.633E-05 88.440323 3.406 4.768E+01 6.80 85.730019 5.392E-07 0 200 3.211E-05 88.676971 3.479 4.870E+01 3.24 87.210114 5.492E-07 0 300 4.782E-05 88.810371 3.539 4.955E+01 2.07 87.831284 5.579E-07 0 400 6.380E-05 88.641617 3.433 4.806E+01 1.60 87.776505 5.422E-07 0 600 9.487E-05 88.918167 3.696 5.174E+01 1.37 88.537949 5.819E-07 0 800 1.263E-04 89.012192 3.815 5.341E+01 1.04 88.659721 6.000E-07 0 1000 1.578E-04 89.049835 3.875 5.425E+01 .82 88.741882 6.092E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 8.0 ------------------------------------------ Floating operations per iteration = 16 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 5.183E-07 .000000 .000 .000E+00 .00 30.867935 5.183E-07 0 5 1.553E-06 61.868366 1.004 1.607E+01 .00 51.520115 2.597E-07 0 10 2.398E-06 77.229202 1.857 2.971E+01 39.09 66.733025 3.846E-07 0 20 3.467E-06 107.092384 4.279 6.847E+01 63.89 92.292038 6.394E-07 0 50 8.079E-06 107.556526 4.332 6.931E+01 24.53 99.020355 6.444E-07 0 100 1.484E-05 112.466995 5.144 8.231E+01 14.54 107.839363 7.318E-07 0 200 2.922E-05 112.325607 5.108 8.172E+01 6.84 109.499603 7.276E-07 0 300 4.312E-05 113.013031 5.363 8.581E+01 4.62 111.317390 7.593E-07 0 400 5.710E-05 113.394531 5.558 8.892E+01 3.48 112.090607 7.842E-07 0 600 8.590E-05 112.975311 5.238 8.381E+01 2.45 111.752594 7.419E-07 0 800 1.143E-04 112.794609 5.056 8.090E+01 1.83 111.952766 7.172E-07 0 1000 1.428E-04 112.690216 4.923 7.877E+01 1.45 112.039352 6.990E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 9.0 ------------------------------------------ Floating operations per iteration = 18 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 5.946E-07 .000000 .000 .000E+00 .00 30.273428 5.946E-07 0 5 1.752E-06 62.185810 1.054 1.897E+01 .00 51.358101 3.051E-07 0 10 2.752E-06 75.599922 1.805 3.250E+01 34.00 65.404259 4.299E-07 0 20 4.078E-06 101.435524 3.929 7.072E+01 56.65 88.275482 6.972E-07 0 50 9.603E-06 101.034851 3.883 6.989E+01 21.53 93.720184 6.918E-07 0 100 1.778E-05 105.112465 4.583 8.249E+01 12.73 101.236053 7.848E-07 0 200 3.512E-05 104.874557 4.519 8.134E+01 5.98 102.508698 7.756E-07 0 300 5.206E-05 105.187737 4.642 8.355E+01 3.85 103.716873 7.943E-07 0 400 6.917E-05 105.256767 4.679 8.422E+01 2.74 104.085022 8.001E-07 0 600 1.035E-04 105.198723 4.632 8.337E+01 1.75 104.358704 7.925E-07 0 800 1.378E-04 105.126358 4.554 8.197E+01 1.27 104.477745 7.797E-07 0 1000 1.720E-04 105.114838 4.538 8.169E+01 .98 104.629051 7.771E-07 0 COMPUTATIONAL INTENSITY = flop per mem ref 10.0 ------------------------------------------ Floating operations per iteration = 20 Memory references per iteration = 2 LOOP LOOP RINF N1/2 S1/2 ERROR R-AVERAGE T0 NTIM LENGTH SECS MFLOP/S VEC.LEN FLOP % MFLOP/S SECS repeat 1 5.671E-07 .000000 .000 .000E+00 .00 35.268333 5.671E-07 0 5 1.520E-06 83.952408 1.380 2.761E+01 .00 65.789398 3.289E-07 0 10 2.140E-06 115.830627 2.827 5.653E+01 55.83 93.466385 4.881E-07 0 20 4.019E-06 113.056038 2.653 5.307E+01 26.01 99.521980 4.694E-07 0 50 9.036E-06 116.962311 3.013 6.027E+01 11.27 110.669106 5.153E-07 0 100 1.719E-05 119.792557 3.412 6.824E+01 6.88 116.348465 5.697E-07 0 200 3.412E-05 119.260658 3.291 6.582E+01 3.31 117.219894 5.519E-07 0 300 5.075E-05 119.469780 3.361 6.722E+01 2.13 118.234673 5.627E-07 0 400 6.743E-05 119.577469 3.411 6.823E+01 1.54 118.637329 5.706E-07 0 600 1.031E-04 117.796005 2.157 4.313E+01 4.46 116.341347 3.662E-07 0 800 1.354E-04 118.222908 2.558 5.116E+01 3.45 118.166176 4.328E-07 0 1000 1.688E-04 118.533730 2.930 5.860E+01 2.83 118.484848 4.944E-07 0 CALCULATION OF FHALF -------------------- F RINF RHAT FHALF ERROR f/m Mflop/s Mflop/s flop/mref % 1.0 75.058678 .000000 .0000 .00 2.0 98.165970 141.828903 .8896 .00 3.0 103.998039 128.834534 .6859 1.16 4.0 96.789421 108.125870 .3078 3.36 5.0 75.780251 79.196724 -.3097 7.48 6.0 91.629395 85.346870 -.1525 7.41 7.0 89.049835 86.632896 -.1146 5.74 8.0 112.690216 101.277534 .3731 9.66 9.0 105.114838 106.247444 .5550 7.85 10.0 118.533730 116.347481 .9563 8.82 LAST VALUES : RHAT = 116.347481 Mflop/s FHALF = .9563 flop/mref TOTAL EXECUTION TIME IS 3.0915820599E+00 SECONDS.