Taylor Test - Benchmark Problem
This DYNA2D analysis is a cylindrical copper projectile impacting a rigid boundary at 190 m/s. The simulation is run to 70 micro-sec in 3519 cycles (see table below graphic for run times).
1. “CPU time to 70 msec” is the processor time for the benchmark, 500
element,
2. “CPU time 500 msec” is the same problem run to 500 msec (29658 cycles).
This is a better benchmark because the CPU time with faster computers is
dominated by startup, ending and I/O.
3. TPEC is "time per zone cycle" (CPU time / # of elements / problem
cycles).
|
Computer |
MHz |
CPU Time to 70 msec |
CPU Time to 500 msec |
TPEC (msec) |
|
VAX 11-780 |
|
4594 |
|
2610 |
|
INTEL 486 |
66 |
511 |
|
290 |
|
SGI R3000/R3010 |
33 |
477 |
|
271 |
|
SUN SPARKstation
2 |
|
416 |
|
236 |
|
SGI R4000/R4010 |
100 |
208 |
|
118 |
|
INTEL P90 NOTEBOOK |
90 |
152 |
|
86 |
|
INTEL PENTIUM |
133 |
104 |
|
59 |
|
SGI R4400/R4010 |
150 |
93 |
|
53 |
|
INTEL PENTIUM OVERDRIVE MMX |
200 |
75 |
|
42 |
|
INTEL PENTIUM PRO |
200 |
48 |
|
24 |
|
CRAY 1 |
|
48 |
|
24 |
|
INTEL PENTIUM II MMX |
300 |
35 |
|
20 |
|
SGI R8000 |
150 |
34 |
|
19 |
|
CRAY XMP |
|
23.7 |
|
14 |
|
SGI R10000 |
195 |
15.7 |
|
8.9 |
|
CRAY YMP |
|
15.2 |
|
8.6 |
|
PENTIUM III |
700 |
15 |
|
8.5 |
|
PENTIUM III |
1000 |
8.4 |
68.9 |
4.6 |
|
PENTIUM III - Intel compiler |
1000 |
|
50.4 |
3.4 |
|
Pentium 4 M Laptop |
1500 |
4.9 |
40.3 |
2.7 |
|
Pentium 4 M Laptop - Intel compiler |
1500 |
|
24.7 |
1.7 |
|
Pentium 4 M Laptop |
1600 |
4.6 |
37.9 |
2.6 |
|
Pentium 4 M Laptop - Intel compiler |
1600 |
|
23.2 |
1.6 |
|
INTEL PENTIUM 4 |
3000 |
4.1 |
32.3 |
2.2 |
|
INTEL PENTIUM 4 - Intel compiler |
3000 |
|
20.2 |
1.4 |