Benchmarks Report for pc10 & pc0
(Linpack, Dhrystone, Whetstone)
by: Godfrey Tan
date: 02-04-9

Documents Include:

pc0 - Single 200MHz PPro
pc10 - Quad 200MHz PPro

Benchmarks Comparison Between pc0 & pc10
Linpack

 pc0 pc10 (01/30) pc10 (02/02)
Average Mflops (n=100 and runs for each category)
Single Rolled 47.96 46.91 47.05
Double Rolled 47.25 44.16 44.75
Single Unrolled 70.17 63.79 63.92
Double Unrolled 67.49 48.99 50.20

Drystone
Version 2.1
Average of two runs (number of loops=1,000,000 and 10,000,000)

pc0 pc10
Drystone/sec 379795.35 378787.9
VAX MIPS rating 216.16 215.588

Whetstone

 MWIPS Mflops1 Mflops2 Mflops3 Cosmops Expmops Fixmops Ifmops Eqmops
pc0 93.993 45.867 39.363 16.220 2.863 2.025 47.908 44.104 14.271
pc10 101.857 46.479 40.598 17.749 3.094 2.425 47.366 51.953 14.017

Details of Benchmarks on pc10
Lmbench
See attached comparison list.

Linpack
Summary:

Average Mflops (with n = 100 and 1000)
Single Rolled 46.91
Double Rolled 44.16
Single Unrolled 63.79
Double Unrolled 48.99

Linpack Benchmark in 'C/C++' n@100
Date 01-30-98
Model
CPU PPro
Clock MHz 200
Cache
H/W options
OS Solaris 2.6
Compiler gcc
Options 03
Run by GT

Rolling Rolled
Precision Single
norm. resid 1.9
resid 4.52336171e-05
machep 1.19209290e-07
x[0]-1 -1.31130219e-05
x[n-1]-1 -1.30534172e-05
matgen 1 seconds 0.00182
matgen 2 seconds 0.00182
Repetitions 312
Leading dimension 201


dgefa dgesl total Mflops
1 pass seconds 0.02000 -0.00000 0.02000
Repeat seconds 0.01491 0.00042 0.01533 44.79
Repeat seconds 0.01487 0.00046 0.01533 44.79
Repeat seconds 0.01487 0.00046 0.01533 44.79
Repeat seconds 0.01487 0.00042 0.01530 44.89
Repeat seconds 0.01487 0.00042 0.01530 44.89
Average


44.83

Linpack Benchmark in 'C/C++' n@100
Date Model CPU Clock MHz Cache H/W options OS Compiler Run by

Rolling Unrolled
Precision Single
norm. resid 1.9
resid 4.52336171e-05
machep 1.19209290e-07
x[0]-1 -1.31130219e-05
x[n-1]-1 -1.30534172e-05
matgen 1 seconds 0.00237
matgen 2 seconds 0.00237
Repetitions 394
Leading dimension 201


dgefa dgesl total Mflops
1 pass seconds 0.01000 0.00000 0.01000
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01035 0.00033 0.01068 64.31
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01032 0.00033 0.01065 64.47
Average


64.44
Leading dimension 200


Repeat seconds 0.01045 0.00030 0.01075 63.86
Repeat seconds 0.01045 0.00030 0.01075 63.86
Repeat seconds 0.01047 0.00030 0.01078 63.71
Repeat seconds 0.01047 0.00030 0.01078 63.71
Repeat seconds 0.01045 0.00033 0.01078 63.71
Average


63.77

Linpack Benchmark in 'C/C++' n@100
Date Model CPU Clock MHz Cache H/W options OS Compiler Run by

Rolling Unrolled
Precision Double
norm. resid 1.9
resid 8.39915160e-14
machep 2.22044605e-16
x[0]-1 -6.22835117e-14
x[n-1]-1 -4.16333634e-14
matgen 1 seconds 0.00224
matgen 2 seconds 0.00225
Repetitions 320
Leading dimension 201


dgefa dgesl total Mflops
1 pass seconds 0.02000 -0.00000 0.02000
Repeat seconds 0.01335 0.00031 0.01366 50.26
Repeat seconds 0.01338 0.00031 0.01369 50.14
Repeat seconds 0.01341 0.00031 0.01373 50.03
Repeat seconds 0.01335 0.00031 0.01366 50.26
Repeat seconds 0.01335 0.00031 0.01366 50.26
Average


50.19
Leading dimension 200


Repeat seconds 0.01331 0.00034 0.01366 50.28
Repeat seconds 0.01328 0.00034 0.01363 50.40
Repeat seconds 0.01331 0.00031 0.01363 50.40
Repeat seconds 0.01331 0.00034 0.01366 50.28
Repeat seconds 0.01328 0.00034 0.01363 50.40
Average


50.35

Linpack Benchmark in 'C/C++' n @ 1000
Date Model CPU Clock MHz Cache H/W options OS Compiler Options Run by

Rolling Rolled
Precision Single
norm. resid 1.9
resid 4.52336171e-05
machep 1.19209290e-07
x[0]-1 -1.31130219e-05
x[n-1]-1 -1.30534172e-05
matgen 1 seconds 0.00182
matgen 2 seconds 0.00182
Repetitions 311
Leading dimension 201


dgefa dgesl total Mflops
1 pass seconds 0.02000 0.00000 0.02000
Repeat seconds 0.01419 0.00045 0.01464 46.89
Repeat seconds 0.01419 0.00042 0.01461 46.99
Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Average


46.91
Leading dimension 200


Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Repeat seconds 0.01423 0.00042 0.01464 46.89
Average


46.89

Linpack Benchmark in 'C/C++' n @1000
Date Model CPU Clock MHz Cache H/W options OS Compiler Options Run by

Rolling Rolled
Precision Double
norm. resid 1.9
resid 8.39915160e-14
machep 2.22044605e-16
x[0]-1 -6.22835117e-14
x[n-1]-1 -4.16333634e-14
matgen 1 seconds 0.00211
matgen 2 seconds 0.00210
Repetitions 283
Leading dimension 201

dgefa dgesl total Mflops
1 pass seconds 0.01000 -0.00000 0.01000
Repeat seconds 0.01553 0.00042 0.01595 43.05
Repeat seconds 0.01553 0.00042 0.01595 43.05
Repeat seconds 0.01553 0.00046 0.01599 42.96
Repeat seconds 0.01553 0.00046 0.01599 42.96
Repeat seconds 0.01553 0.00042 0.01595 43.05
Average 43.01
Leading dimension 200
Repeat seconds 0.01539 0.00042 0.01582 43.42
Repeat seconds 0.01539 0.00042 0.01582 43.42
Repeat seconds 0.01536 0.00046 0.01582 43.42
Repeat seconds 0.01536 0.00046 0.01582 43.42
Repeat seconds 0.01539 0.00042 0.01582 43.42
Average


43.42

Linpack Benchmark in 'C/C++' n @ 1000
Date Model CPU Clock MHz Cache H/W options OS Compiler Options Run by
Rolling Unrolled
Precision Single
norm. resid 1.9
resid 4.52336171e-05
machep 1.19209290e-07
x[0]-1 -1.31130219e-05
x[n-1]-1 -1.30534172e-05
matgen 1 seconds 0.00237
matgen 2 seconds 0.00237
Repetitions 394
Leading dimension 201

dgefa dgesl total Mflops
1 pass seconds 0.01000 0.00000 0.01000
Repeat seconds 0.01030 0.00033 0.01063 64.62
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01032 0.00033 0.01065 64.47
Repeat seconds 0.01032 0.00030 0.01063 64.62
Average


64.53
Leading dimension 200


Repeat seconds 0.01045 0.00033 0.01078 63.71
Repeat seconds 0.01045 0.00030 0.01075 63.86
Repeat seconds 0.01045 0.00033 0.01078 63.71
Repeat seconds 0.01045 0.00030 0.01075 63.86
Repeat seconds 0.01042 0.00033 0.01075 63.86
Average


63.80

Linpack Benchmark in 'C/C++' n @ 1000
Date Model CPU Clock MHz Cache H/W options OS Compiler Options Run by
Rolling Unrolled
Precision Double
norm. resid 1.9
resid 8.39915160e-14
machep 2.22044605e-16
x[0]-1 -6.22835117e-14
x[n-1]-1 -4.16333634e-14
matgen 1 seconds 0.00167
matgen 2 seconds 0.00167
Repetitions 317
Leading dimension 201

dgefa dgesl total Mflops
1 pass seconds 0.01000 -0.00000 0.01000
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00032 0.01438 47.74
Average


47.65
Leading dimension 200


Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Repeat seconds 0.01407 0.00035 0.01442 47.63
Average


47.63

Dhrystone
Dhrystone Benchmark, Version 1.1 (Language: C)
Register option not selected.
Number of loops: 1000000
Runtime (sec): 2.2
Dhrystones/sec: 456621.0
VAX MIPS rating: 259.887
Dhrystone Benchmark, Version 1.1 (Language: C)
Register option not selected.
Number of loops: 10000000
Runtime (sec): 22.0
Dhrystones/sec: 454752.2
VAX MIPS rating: 258.823
Dhrystone Benchmark, Version 2.1 (Language: C)
Register option not selected.
Microseconds for one loop: 2.6
Dhrystones per second: 378787.9
VAX MIPS rating: 215.588
Dhrystone Benchmark, Version 2.1 (Language: C)
Register option not selected.
Microseconds for one loop: 2.6
Dhrystones per second: 378787.9
VAX MIPS rating: 215.588

Whetstone
Single Precision Benchmark in C/C++
Date Model CPU Clock MHz Cache H/W options OS Compiler Options Run by
Loop content Result MFLOPS MOPS Seconds
N1 floating point -1.12475025653839111 46.479
0.423
N2 floating point -1.12274754047393799 40.598
3.390
N3 if then else 1.00000000000000000
51.953 2.040
N4 fixed point 12.00000000000000000
47.366 6.810
N5 sin,cos etc. 0.49904659390449524
3.094 27.540
N6 floating point 0.99999988079071045 17.749
31.120
N7 assignments 3.00000000000000000
14.017 13.500
N8 exp,sqrt etc. 0.75110864639282227
2.425 15.710
MWIPS
101.857 100.533
Results to load to spreadsheet
MWIPS Mflops1
Mflops2 Mflops3 Cosmops Expmops Fixpmops Ifmops Eqmops
Results to load to spreadsheet
101.857 46.479
40.598 17.749 3.094 2.425 47.366 51.953 14.017

Details of Benchmarks on pc0
Lmbench
See Attached Comparison List.
Linpack
Summary:
Average Mflops (with100 and 1000)
Single Rolled 47.96
Double Rolled 47.25
Single Unrolled 70.17
Double Unrolled 67.49
Benchmark in C/C++ @ 100
Date
02-02-98		 Model
pc10		 CPU
PPro		 Clock MHz
200!		 Cache H/W options OS Compiler Options Run by
Rolling Rolled
Precision Double
norm. resid 1.9
resid 8.39915160e-14
machep 2.22044605e-16
x[0]-1 -6.22835117e-14
x[n-1]-1 -4.16333634e-14
matgen 1 seconds 0.00273
matgen 2 seconds 0.00274
Repetitions 298
Leading dimension 201

dgefa dgesl total Mflops
1 pass seconds 0.01000 -0.00000
Repeat 0.01405 0.00044 0.01448 47.41

DhrystoneVersion 1.1 (Language: C)
Register option not selected.
Number loops: 1000000
Runtime (sec): 2.2
Dhrystones/sec: 454545.5
VAX MIPS rating: 258.705 10000000 22.0 454132.6 258.470 2.1
Microseconds for one loop: 2.6
Dhrystones per second: 380228.1 216.408 379362.7 215.915

Whetstone
Loop content
Result MOPS
N1 floating point -1.12475025653839111 45.867 0.396
N2 -1.12274754047393799 39.363 3.230
N3 if then else 1.00000000000000000 44.104 2.220
N4 fixed 12.00000000000000000 47.908 6.220
N5 sincos etc. 0.49904659390449524 2.863 27.490
N6 0.99999988079071045 16.220 31.460
N7 assignments 3.00000000000000000 14.271 12.250
N8 expsqrt 0.75110864639282227 2.025 17.380
MWIPS 93.993 100.646
Results to load spreadsheet Mflops1 Mflops2 Mflops3 Cosmops Expmops Fixpmops Ifmops Eqmops