Beta version 24.6 - Athlon users wanted - Page 3

Paulie · 2004-12-09, 14:41

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.55 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 23.7, RdtscTiming=1
Best time for 512K FFT length: 45.599 ms.
Best time for 640K FFT length: 57.303 ms.
Best time for 768K FFT length: 70.191 ms.
Best time for 896K FFT length: 84.139 ms.
Best time for 1024K FFT length: 94.577 ms.
Best time for 1280K FFT length: 122.043 ms.
Best time for 1536K FFT length: 149.250 ms.
Best time for 1792K FFT length: 177.972 ms.
Best time for 2048K FFT length: 199.492 ms.

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.51 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 23.8, RdtscTiming=1
Best time for 512K FFT length: 45.489 ms.
Best time for 640K FFT length: 57.543 ms.
Best time for 768K FFT length: 70.479 ms.
Best time for 896K FFT length: 84.352 ms.
Best time for 1024K FFT length: 95.030 ms.
Best time for 1280K FFT length: 122.612 ms.
Best time for 1536K FFT length: 149.297 ms.
Best time for 1792K FFT length: 178.057 ms.
Best time for 2048K FFT length: 199.734 ms.

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.46 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 24.6, RdtscTiming=1
Best time for 512K FFT length: 45.619 ms.
Best time for 640K FFT length: 57.735 ms.
Best time for 768K FFT length: 71.017 ms.
Best time for 896K FFT length: 89.719 ms.
Best time for 1024K FFT length: 100.478 ms.
Best time for 1280K FFT length: 130.460 ms.
Best time for 1536K FFT length: 160.626 ms.
Best time for 1792K FFT length: 189.422 ms.
Best time for 2048K FFT length: 210.682 ms.

Prime95 · 2004-12-09, 19:03

Paulie, how does the Pentium-M compare if you put "CPUSupportsSSE2=0" in local.ini?

flava · 2004-12-09, 19:10

Works like a charm on my laptop. Thank you! :)

mobile AMD Athlon(tm) XP2800+
CPU speed: 2119.96 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 256

Prime95 version 23.7, RdtscTiming=1

Best time for 512K FFT length: 45.603 ms.
Best time for 640K FFT length: 58.825 ms.
Best time for 768K FFT length: 70.577 ms.
Best time for 896K FFT length: 82.925 ms.
Best time for 1024K FFT length: 93.971 ms.
Best time for 1280K FFT length: 129.194 ms.
Best time for 1536K FFT length: 147.872 ms.
Best time for 1792K FFT length: 188.165 ms.
Best time for 2048K FFT length: 198.254 ms.

Prime95 version 24.6, RdtscTiming=1

Best time for 512K FFT length: 32.047 ms.
Best time for 640K FFT length: 42.981 ms.
Best time for 768K FFT length: 52.466 ms.
Best time for 896K FFT length: 63.637 ms.
Best time for 1024K FFT length: 71.055 ms.
Best time for 1280K FFT length: 95.804 ms.
Best time for 1536K FFT length: 114.782 ms.
Best time for 1792K FFT length: 138.144 ms.
Best time for 2048K FFT length: 154.963 ms.

SalemTheCat100 · 2004-12-09, 19:26

Quote:

Originally Posted by sdbardwick

Neat!
I'll switch my Athlon 1900MP (2x 1.6GHz) box over to double checking as soon as the current factoring assignment finishes. Probably take a little less than a month for the first results.

Given the various sizes of L1/L2 cache in the Athlon/Duron/Sempron processors, is the new code optimized for one version in particular?

Would you like benchmarks posted?

I'm sure SalemTheCat100 will be pleased

-Scott-

Thanks for the effort in supporting the AMD Athlon processors.

Salem

Wolf · 2004-12-09, 19:39

I think George is being too modest only claiming a 15% improvement!

AMD Athlon(TM) XP2100+
CPU speed: 1746.03 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 64 KB
L2 cache size: 256 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 256

Prime95 version 23.8, RdtscTiming=1

Best time for 512K FFT length: 55.348 ms.
Best time for 640K FFT length: 72.015 ms.
Best time for 768K FFT length: 85.664 ms.
Best time for 896K FFT length: 101.333 ms.
Best time for 1024K FFT length: 115.199 ms.
Best time for 1280K FFT length: 154.964 ms.
Best time for 1536K FFT length: 179.477 ms.
Best time for 1792K FFT length: 229.369 ms.
Best time for 2048K FFT length: 258.630 ms.

Prime95 version 24.6, RdtscTiming=1

Best time for 512K FFT length: 38.292 ms.
Best time for 640K FFT length: 51.586 ms.
Best time for 768K FFT length: 62.646 ms.
Best time for 896K FFT length: 76.130 ms.
Best time for 1024K FFT length: 85.223 ms.
Best time for 1280K FFT length: 115.138 ms.
Best time for 1536K FFT length: 139.228 ms.
Best time for 1792K FFT length: 167.028 ms.
Best time for 2048K FFT length: 187.977 ms.

SalemTheCat100 · 2004-12-09, 20:03

Some observations:

Version 24.6 without SSE2 is about 4.5% faster than Version 23.8.

Version 23.8 without SSE2 is slower whereas 24.6 is faster without SSE2.

Is there an optimized Athlon64 version in the works?

SALEM

-----------------------------------------------------------------------

Version 23.8.1.0

AMD Athlon(tm) 64 Processor 3400+
CPU speed: 2410.86 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 512
Prime95 version 23.8, RdtscTiming=1
Best time for 384K FFT length: 20.134 ms.
Best time for 448K FFT length: 24.076 ms.
Best time for 512K FFT length: 26.763 ms.
Best time for 640K FFT length: 33.394 ms.
Best time for 768K FFT length: 40.534 ms.
Best time for 896K FFT length: 49.038 ms.
Best time for 1024K FFT length: 55.168 ms.
Best time for 1280K FFT length: 73.898 ms.
Best time for 1536K FFT length: 88.287 ms.
Best time for 1792K FFT length: 107.433 ms.
Best time for 2048K FFT length: 121.641 ms.

No SSE2

AMD Athlon(tm) 64 Processor 3400+
CPU speed: 2410.69 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 512
Prime95 version 23.8, RdtscTiming=1
Best time for 384K FFT length: 22.612 ms.
Best time for 448K FFT length: 25.288 ms.
Best time for 512K FFT length: 27.608 ms.
Best time for 640K FFT length: 36.565 ms.
Best time for 768K FFT length: 43.933 ms.
Best time for 896K FFT length: 52.595 ms.
Best time for 1024K FFT length: 58.794 ms.
Best time for 1280K FFT length: 74.902 ms.
Best time for 1536K FFT length: 90.965 ms.
Best time for 1792K FFT length: 109.248 ms.
Best time for 2048K FFT length: 123.113 ms.

Version 24.6.1.0

AMD Athlon(tm) 64 Processor 3400+
CPU speed: 2410.87 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 512
Prime95 version 24.6, RdtscTiming=1
Best time for 512K FFT length: 26.847 ms.
Best time for 640K FFT length: 33.750 ms.
Best time for 768K FFT length: 40.709 ms.
Best time for 896K FFT length: 49.252 ms.
Best time for 1024K FFT length: 55.465 ms.
Best time for 1280K FFT length: 74.106 ms.
Best time for 1536K FFT length: 88.746 ms.
Best time for 1792K FFT length: 107.562 ms.
Best time for 2048K FFT length: 121.940 ms.

No SSE2

AMD Athlon(tm) 64 Processor 3400+
CPU speed: 2410.92 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 64 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 32
L2 TLBS: 512
Prime95 version 24.6, RdtscTiming=1
Best time for 512K FFT length: 25.206 ms.
Best time for 640K FFT length: 33.701 ms.
Best time for 768K FFT length: 40.126 ms.
Best time for 896K FFT length: 48.847 ms.
Best time for 1024K FFT length: 54.395 ms.
Best time for 1280K FFT length: 72.521 ms.
Best time for 1536K FFT length: 87.162 ms.
Best time for 1792K FFT length: 104.300 ms.
Best time for 2048K FFT length: 116.092 ms.

Paulie · 2004-12-09, 20:15

Much better!!!

Intel(R) Pentium(R) M processor 1600MHz
CPU speed: 1594.81 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: 1024 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 128
Prime95 version 23.7, RdtscTiming=1
Best time for 512K FFT length: 43.403 ms.
Best time for 640K FFT length: 57.332 ms.
Best time for 768K FFT length: 69.735 ms.
Best time for 896K FFT length: 83.268 ms.
Best time for 1024K FFT length: 94.354 ms.
Best time for 1280K FFT length: 123.936 ms.
Best time for 1536K FFT length: 147.184 ms.
Best time for 1792K FFT length: 180.734 ms.
Best time for 2048K FFT length: 199.028 ms.

Intel(R) Pentium(R) M processor 1600MHz
CPU speed: 1594.67 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: 1024 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 128
Prime95 version 23.8, RdtscTiming=1
Best time for 512K FFT length: 43.555 ms.
Best time for 640K FFT length: 58.165 ms.
Best time for 768K FFT length: 70.387 ms.
Best time for 896K FFT length: 84.209 ms.
Best time for 1024K FFT length: 95.294 ms.
Best time for 1280K FFT length: 125.689 ms.
Best time for 1536K FFT length: 148.689 ms.
Best time for 1792K FFT length: 182.862 ms.
Best time for 2048K FFT length: 199.285 ms.

Intel(R) Pentium(R) M processor 1600MHz
CPU speed: 1594.73 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: 1024 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 128
Prime95 version 24.6, RdtscTiming=1
Best time for 512K FFT length: 42.945 ms.
Best time for 640K FFT length: 57.472 ms.
Best time for 768K FFT length: 70.577 ms.
Best time for 896K FFT length: 84.819 ms.
Best time for 1024K FFT length: 95.173 ms.
Best time for 1280K FFT length: 120.814 ms.
Best time for 1536K FFT length: 149.215 ms.
Best time for 1792K FFT length: 180.982 ms.
Best time for 2048K FFT length: 200.603 ms.

Paulie · 2004-12-09, 20:31

The T42 with the different CPU than my T41:

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.51 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 23.7, RdtscTiming=1
Best time for 512K FFT length: 39.832 ms.
Best time for 640K FFT length: 54.439 ms.
Best time for 768K FFT length: 66.672 ms.
Best time for 896K FFT length: 80.723 ms.
Best time for 1024K FFT length: 91.565 ms.
Best time for 1280K FFT length: 118.482 ms.
Best time for 1536K FFT length: 141.612 ms.
Best time for 1792K FFT length: 175.459 ms.
Best time for 2048K FFT length: 190.218 ms.

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.57 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 23.8, RdtscTiming=1
Best time for 512K FFT length: 39.168 ms.
Best time for 640K FFT length: 53.837 ms.
Best time for 768K FFT length: 65.463 ms.
Best time for 896K FFT length: 79.924 ms.
Best time for 1024K FFT length: 92.466 ms.
Best time for 1280K FFT length: 120.630 ms.
Best time for 1536K FFT length: 142.399 ms.
Best time for 1792K FFT length: 172.767 ms.
Best time for 2048K FFT length: 187.741 ms.

Intel(R) Pentium(R) M processor 1.70GHz
CPU speed: 1698.51 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE
L1 cache size: 32 KB
L2 cache size: unknown
L1 cache line size: 64 bytes
L2 cache line size: unknown
TLBS: 128
Prime95 version 24.6, RdtscTiming=1
Best time for 512K FFT length: 37.492 ms.
Best time for 640K FFT length: 49.905 ms.
Best time for 768K FFT length: 60.386 ms.
Best time for 896K FFT length: 73.509 ms.
Best time for 1024K FFT length: 81.655 ms.
Best time for 1280K FFT length: 105.156 ms.
Best time for 1536K FFT length: 129.665 ms.
Best time for 1792K FFT length: 157.141 ms.
Best time for 2048K FFT length: 172.926 ms.

delta_t · 2004-12-09, 20:37

Interesting on the Pentium M. I know that the 23.8 version ran slower than an equally clocked Pentium 4 M. What is the reason for the improvement on the Pentium M on 24.6 (just curious)? Looks like laptops with Pentium M's should use 24.6?

Prime95 · 2004-12-09, 21:07

Quote:

Originally Posted by delta_t

Interesting on the Pentium M. I know that the 23.8 version ran slower than an equally clocked Pentium 4 M. What is the reason for the improvement on the Pentium M on 24.6 (just curious)? Looks like laptops with Pentium M's should use 24.6?

It's all about good cache usage. The P4 cache architecture, branch prediction, and out-of-order execution is so good you can pretend a Northwood is a CPU with a 512KB L1 cache.

The Athlon has substantial penalties for L2 cache accesses. So it is a CPU with a 64KB L1 cache.

The P3 and earlier have severe penalties for L2 cache access. Plus the CPU only has an 8KB or 16KB L1 cache. The Pentium-M seems to be a P3 with a 32KB L1 cache.

The size of the L1 cache governs how you should "split" a two-pass FFT. For example, a 1M FFT does 20 FFT levels. On a P3, you can do 8 levels entirely within the L1 cache. So it is best to do 12 levels in pass 1 and 8 levels in pass 2. Pass 1 won't fit in L1, but at least pass 2 will. This is what version 23 did.

In v24, the 1M FFT is done 10 levels in pass 1 and 10 levels in pass 2. The Athlon is happy because both pass 1 and pass 2 fit in the L1 cache. The P3 is unhappy because now neither fits in the L1 cache.

Your Pentium-M with a 32KB cache is like the Athlon - it likes the new 10-10 split. It doesn't like the SSE2 split because (for rather complicated reasons), SSE2 uses 2 or 4 times as much L1 cache space. The SSE2's 9 levels / 11 levels split doesn't fit in the L1 cache.

There are other factors at work too. V23 had a P3 TLB-friendly memory layout. The Athlon seems to prefer the new more linear approach. V24 supports copying pass 1 data to a contiguous scratch area, v23 only supported in-place operations.

All-in-all, use v24 with the CPUSupportsSSE2=0 option if that runs faster for you.

P.S. The Opteron remains an enigma. V24 seems to run little better than v23 with SSE2 support turned off. In fact, I think the Opteron now underperforms a similarly clocked AthlonXP.

gbvalor · 2004-12-09, 21:50

Congratulations George!,

And thanks by your superb work.

These last months I've been very busy and a bit out of Mersenne stuf and today I've seen in the forum this 'incredible and even faster prime95'.

Waiting for the Linux mprime version

.

Guillermo

2004-12-09, 14:41	#23
Paulie Aug 2002 223 Posts	Pentium M 1.7 (Thinkpad T42) (2048 cache?) Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.55 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2 L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 23.7, RdtscTiming=1 Best time for 512K FFT length: 45.599 ms. Best time for 640K FFT length: 57.303 ms. Best time for 768K FFT length: 70.191 ms. Best time for 896K FFT length: 84.139 ms. Best time for 1024K FFT length: 94.577 ms. Best time for 1280K FFT length: 122.043 ms. Best time for 1536K FFT length: 149.250 ms. Best time for 1792K FFT length: 177.972 ms. Best time for 2048K FFT length: 199.492 ms. Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.51 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2 L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 23.8, RdtscTiming=1 Best time for 512K FFT length: 45.489 ms. Best time for 640K FFT length: 57.543 ms. Best time for 768K FFT length: 70.479 ms. Best time for 896K FFT length: 84.352 ms. Best time for 1024K FFT length: 95.030 ms. Best time for 1280K FFT length: 122.612 ms. Best time for 1536K FFT length: 149.297 ms. Best time for 1792K FFT length: 178.057 ms. Best time for 2048K FFT length: 199.734 ms. Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.46 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2 L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 45.619 ms. Best time for 640K FFT length: 57.735 ms. Best time for 768K FFT length: 71.017 ms. Best time for 896K FFT length: 89.719 ms. Best time for 1024K FFT length: 100.478 ms. Best time for 1280K FFT length: 130.460 ms. Best time for 1536K FFT length: 160.626 ms. Best time for 1792K FFT length: 189.422 ms. Best time for 2048K FFT length: 210.682 ms.

2004-12-09, 20:03	#28
SalemTheCat100 Oct 2002 2·13 Posts	Results for AMD Athlon 64 3400+ Some observations: Version 24.6 without SSE2 is about 4.5% faster than Version 23.8. Version 23.8 without SSE2 is slower whereas 24.6 is faster without SSE2. Is there an optimized Athlon64 version in the works? SALEM ----------------------------------------------------------------------- Version 23.8.1.0 AMD Athlon(tm) 64 Processor 3400+ CPU speed: 2410.86 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2 L1 cache size: 64 KB L2 cache size: 512 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 512 Prime95 version 23.8, RdtscTiming=1 Best time for 384K FFT length: 20.134 ms. Best time for 448K FFT length: 24.076 ms. Best time for 512K FFT length: 26.763 ms. Best time for 640K FFT length: 33.394 ms. Best time for 768K FFT length: 40.534 ms. Best time for 896K FFT length: 49.038 ms. Best time for 1024K FFT length: 55.168 ms. Best time for 1280K FFT length: 73.898 ms. Best time for 1536K FFT length: 88.287 ms. Best time for 1792K FFT length: 107.433 ms. Best time for 2048K FFT length: 121.641 ms. No SSE2 AMD Athlon(tm) 64 Processor 3400+ CPU speed: 2410.69 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 64 KB L2 cache size: 512 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 512 Prime95 version 23.8, RdtscTiming=1 Best time for 384K FFT length: 22.612 ms. Best time for 448K FFT length: 25.288 ms. Best time for 512K FFT length: 27.608 ms. Best time for 640K FFT length: 36.565 ms. Best time for 768K FFT length: 43.933 ms. Best time for 896K FFT length: 52.595 ms. Best time for 1024K FFT length: 58.794 ms. Best time for 1280K FFT length: 74.902 ms. Best time for 1536K FFT length: 90.965 ms. Best time for 1792K FFT length: 109.248 ms. Best time for 2048K FFT length: 123.113 ms. Version 24.6.1.0 AMD Athlon(tm) 64 Processor 3400+ CPU speed: 2410.87 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2 L1 cache size: 64 KB L2 cache size: 512 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 512 Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 26.847 ms. Best time for 640K FFT length: 33.750 ms. Best time for 768K FFT length: 40.709 ms. Best time for 896K FFT length: 49.252 ms. Best time for 1024K FFT length: 55.465 ms. Best time for 1280K FFT length: 74.106 ms. Best time for 1536K FFT length: 88.746 ms. Best time for 1792K FFT length: 107.562 ms. Best time for 2048K FFT length: 121.940 ms. No SSE2 AMD Athlon(tm) 64 Processor 3400+ CPU speed: 2410.92 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 64 KB L2 cache size: 512 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 512 Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 25.206 ms. Best time for 640K FFT length: 33.701 ms. Best time for 768K FFT length: 40.126 ms. Best time for 896K FFT length: 48.847 ms. Best time for 1024K FFT length: 54.395 ms. Best time for 1280K FFT length: 72.521 ms. Best time for 1536K FFT length: 87.162 ms. Best time for 1792K FFT length: 104.300 ms. Best time for 2048K FFT length: 116.092 ms.

2004-12-09, 20:15	#29
Paulie Aug 2002 223 Posts	George - Rerun with CPUSupportsSSE2=0 Much better!!! Intel(R) Pentium(R) M processor 1600MHz CPU speed: 1594.81 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: 1024 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes TLBS: 128 Prime95 version 23.7, RdtscTiming=1 Best time for 512K FFT length: 43.403 ms. Best time for 640K FFT length: 57.332 ms. Best time for 768K FFT length: 69.735 ms. Best time for 896K FFT length: 83.268 ms. Best time for 1024K FFT length: 94.354 ms. Best time for 1280K FFT length: 123.936 ms. Best time for 1536K FFT length: 147.184 ms. Best time for 1792K FFT length: 180.734 ms. Best time for 2048K FFT length: 199.028 ms. Intel(R) Pentium(R) M processor 1600MHz CPU speed: 1594.67 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: 1024 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes TLBS: 128 Prime95 version 23.8, RdtscTiming=1 Best time for 512K FFT length: 43.555 ms. Best time for 640K FFT length: 58.165 ms. Best time for 768K FFT length: 70.387 ms. Best time for 896K FFT length: 84.209 ms. Best time for 1024K FFT length: 95.294 ms. Best time for 1280K FFT length: 125.689 ms. Best time for 1536K FFT length: 148.689 ms. Best time for 1792K FFT length: 182.862 ms. Best time for 2048K FFT length: 199.285 ms. Intel(R) Pentium(R) M processor 1600MHz CPU speed: 1594.73 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: 1024 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes TLBS: 128 Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 42.945 ms. Best time for 640K FFT length: 57.472 ms. Best time for 768K FFT length: 70.577 ms. Best time for 896K FFT length: 84.819 ms. Best time for 1024K FFT length: 95.173 ms. Best time for 1280K FFT length: 120.814 ms. Best time for 1536K FFT length: 149.215 ms. Best time for 1792K FFT length: 180.982 ms. Best time for 2048K FFT length: 200.603 ms. Last fiddled with by Paulie on 2004-12-09 at 20:18

2004-12-09, 20:31	#30
Paulie Aug 2002 223 Posts	George - Rerun with CPUSupportsSSE2=0 The T42 with the different CPU than my T41: Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.51 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 23.7, RdtscTiming=1 Best time for 512K FFT length: 39.832 ms. Best time for 640K FFT length: 54.439 ms. Best time for 768K FFT length: 66.672 ms. Best time for 896K FFT length: 80.723 ms. Best time for 1024K FFT length: 91.565 ms. Best time for 1280K FFT length: 118.482 ms. Best time for 1536K FFT length: 141.612 ms. Best time for 1792K FFT length: 175.459 ms. Best time for 2048K FFT length: 190.218 ms. Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.57 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 23.8, RdtscTiming=1 Best time for 512K FFT length: 39.168 ms. Best time for 640K FFT length: 53.837 ms. Best time for 768K FFT length: 65.463 ms. Best time for 896K FFT length: 79.924 ms. Best time for 1024K FFT length: 92.466 ms. Best time for 1280K FFT length: 120.630 ms. Best time for 1536K FFT length: 142.399 ms. Best time for 1792K FFT length: 172.767 ms. Best time for 2048K FFT length: 187.741 ms. Intel(R) Pentium(R) M processor 1.70GHz CPU speed: 1698.51 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 32 KB L2 cache size: unknown L1 cache line size: 64 bytes L2 cache line size: unknown TLBS: 128 Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 37.492 ms. Best time for 640K FFT length: 49.905 ms. Best time for 768K FFT length: 60.386 ms. Best time for 896K FFT length: 73.509 ms. Best time for 1024K FFT length: 81.655 ms. Best time for 1280K FFT length: 105.156 ms. Best time for 1536K FFT length: 129.665 ms. Best time for 1792K FFT length: 157.141 ms. Best time for 2048K FFT length: 172.926 ms.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
LLR beta Version 3.8.13 (deprecated)	Jean Penné	Software	111	2015-01-26 21:41
Prime95 beta version 28.3	Prime95	Software	68	2014-02-23 05:42
Beta version 24.12 available	Prime95	Software	33	2005-06-14 13:19
Early Beta of version 24.11	Prime95	Software	113	2005-05-24 17:05
Beta version of PRP	Prime95	PSearch	15	2004-09-17 19:21

2004-12-09, 19:03	#24
Prime95 P90 years forever! Aug 2002 Yeehaw, FL 7,537 Posts	Paulie, how does the Pentium-M compare if you put "CPUSupportsSSE2=0" in local.ini?

2004-12-09, 19:10	#25
flava Feb 2003 2·59 Posts	Works like a charm on my laptop. Thank you! :) mobile AMD Athlon(tm) XP2800+ CPU speed: 2119.96 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 64 KB L2 cache size: 512 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 256 Prime95 version 23.7, RdtscTiming=1 Best time for 512K FFT length: 45.603 ms. Best time for 640K FFT length: 58.825 ms. Best time for 768K FFT length: 70.577 ms. Best time for 896K FFT length: 82.925 ms. Best time for 1024K FFT length: 93.971 ms. Best time for 1280K FFT length: 129.194 ms. Best time for 1536K FFT length: 147.872 ms. Best time for 1792K FFT length: 188.165 ms. Best time for 2048K FFT length: 198.254 ms. Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 32.047 ms. Best time for 640K FFT length: 42.981 ms. Best time for 768K FFT length: 52.466 ms. Best time for 896K FFT length: 63.637 ms. Best time for 1024K FFT length: 71.055 ms. Best time for 1280K FFT length: 95.804 ms. Best time for 1536K FFT length: 114.782 ms. Best time for 1792K FFT length: 138.144 ms. Best time for 2048K FFT length: 154.963 ms.

2004-12-09, 19:39	#27
Wolf Jul 2003 UK 3·17 Posts	I think George is being too modest only claiming a 15% improvement! AMD Athlon(TM) XP2100+ CPU speed: 1746.03 MHz CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE L1 cache size: 64 KB L2 cache size: 256 KB L1 cache line size: 64 bytes L2 cache line size: 64 bytes L1 TLBS: 32 L2 TLBS: 256 Prime95 version 23.8, RdtscTiming=1 Best time for 512K FFT length: 55.348 ms. Best time for 640K FFT length: 72.015 ms. Best time for 768K FFT length: 85.664 ms. Best time for 896K FFT length: 101.333 ms. Best time for 1024K FFT length: 115.199 ms. Best time for 1280K FFT length: 154.964 ms. Best time for 1536K FFT length: 179.477 ms. Best time for 1792K FFT length: 229.369 ms. Best time for 2048K FFT length: 258.630 ms. Prime95 version 24.6, RdtscTiming=1 Best time for 512K FFT length: 38.292 ms. Best time for 640K FFT length: 51.586 ms. Best time for 768K FFT length: 62.646 ms. Best time for 896K FFT length: 76.130 ms. Best time for 1024K FFT length: 85.223 ms. Best time for 1280K FFT length: 115.138 ms. Best time for 1536K FFT length: 139.228 ms. Best time for 1792K FFT length: 167.028 ms. Best time for 2048K FFT length: 187.977 ms.

2004-12-09, 20:37	#31
delta_t Nov 2002 Anchorage, AK 3·7·17 Posts	Interesting on the Pentium M. I know that the 23.8 version ran slower than an equally clocked Pentium 4 M. What is the reason for the improvement on the Pentium M on 24.6 (just curious)? Looks like laptops with Pentium M's should use 24.6?

2004-12-09, 21:50	#33
gbvalor Aug 2002 3×37 Posts	Congratulations George!, And thanks by your superb work. These last months I've been very busy and a bit out of Mersenne stuf and today I've seen in the forum this 'incredible and even faster prime95'. Waiting for the Linux mprime version . Guillermo