Beta version 24.6 - Athlon users wanted

[Uncwilly]

Originally Posted by renuac

Prime95 Plz where I can make download version 24.9 of Prime 95?

If you read the whole thread (in detail) you should be able to deduce where it might be.

24.6 and 24.9 are substaintial the same., AFAIK

[Prime95]

I have only emailed one copy of 24.9 to one user. It isn't QA'ed as much as I would like. There is nothing useful in it - other than the no-execute fix.

[renuac]

Ah all good if you will be able to order me in my email: renuac@somewhere for me is good, if you it will not be able not if it worries!

Sorry Bad English

[JuanTutors]

Why hasn't the links for downloading 23.8 been switched to download24.6? Also, have all these errors 24.6 has given just beeninconsequential? It seems some computers have gotten some funny results.

[Prime95]

Originally Posted by dominicanpapi82

Why hasn't the links for downloading 23.8 been switched to download24.6? Also, have all these errors 24.6 has given just beeninconsequential? It seems some computers have gotten some funny results.

Basically, I'm lazy

It is not uncommon for perfectly good versions to languish in Beta for months on end. I doubt 24.6 will ever be an official release as I'm already up to 24.10 internally. You might wonder why version numbers are jumping so. It is because version 24 has been all about ripping the FFT code cleanly out of prime95 and expanding its functionality for SoB, PFGW, LLR, and other projects. Every time I change the interface between these programs and the new gwnum library I bump the version number.

Right now I'm in the midst of porting the FFT code to 64-bit Windows. The Opteron actually passed a torture test last week! And no, it isn't any faster than 32-bit mode. First step is to port the existing assembly code, second step is to make use of the extra SSE2 registers. But right now, I'm rewriting the trial factoring code - at least that code should benefit greatly from 64-bit instructions.

As to your second question, there have been a few minor bugs found in 24.6, but none that affected LL results.

[garo]

Thanks George for the status update! Much appreciated

[renuac]

Thanks George ! Version 24.10 already this finished?

[BTTB]

AMD Athlon XP-M
CPU speed: 2204.87 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 24.6, RdtscTiming=1
Best time for 512K FFT length: 29.067 ms.
Best time for 640K FFT length: 39.908 ms.
Best time for 768K FFT length: 48.696 ms.
Best time for 896K FFT length: 59.261 ms.
Best time for 1024K FFT length: 65.797 ms.
Best time for 1280K FFT length: 84.955 ms.
Best time for 1536K FFT length: 103.051 ms.
Best time for 1792K FFT length: 124.753 ms.
Best time for 2048K FFT length: 138.413 ms.
[Thu Jan 27 00:05:12 2005]
Self-test 1024K passed!
[Thu Jan 27 00:21:47 2005]
Self-test 8K passed!
[Thu Jan 27 00:37:21 2005]
Self-test 10K passed!
[Thu Jan 27 00:54:04 2005]
Self-test 896K passed!
[Thu Jan 27 01:10:20 2005]
Self-test 768K passed!
[Thu Jan 27 01:25:43 2005]
Self-test 12K passed!
[Thu Jan 27 01:41:04 2005]
Self-test 14K passed!
[Thu Jan 27 01:57:30 2005]
Self-test 640K passed!
[Thu Jan 27 02:16:25 2005]
Self-test 512K passed!
[Thu Jan 27 02:31:32 2005]
Self-test 16K passed!
[Thu Jan 27 02:47:09 2005]
Self-test 20K passed!
[Thu Jan 27 03:02:22 2005]
Self-test 448K passed!
[Thu Jan 27 03:17:42 2005]
Self-test 384K passed!
[Thu Jan 27 03:36:44 2005]
Self-test 24K passed!
[Thu Jan 27 03:52:05 2005]
Self-test 28K passed!
[Thu Jan 27 04:07:09 2005]
Self-test 320K passed!
[Thu Jan 27 04:22:30 2005]
Self-test 256K passed!
[Thu Jan 27 04:41:50 2005]
Self-test 32K passed!
[Thu Jan 27 04:56:55 2005]
Self-test 40K passed!
[Thu Jan 27 05:19:40 2005]
Self-test 224K passed!
[Thu Jan 27 05:35:28 2005]
Self-test 192K passed!
[Thu Jan 27 06:23:44 2005]
Self-test 48K passed!
[Thu Jan 27 06:39:52 2005]
Self-test 56K passed!
[Thu Jan 27 06:55:01 2005]
Self-test 160K passed!
[Thu Jan 27 07:13:58 2005]
Self-test 128K passed!
[Thu Jan 27 07:29:37 2005]
Self-test 64K passed!
[Thu Jan 27 07:35:04 2005]
FATAL ERROR: Final result was E79462C3, expected: 455509CD.
Hardware failure detected, consult stress.txt file.

[BTTB]

AMD Athlon XP-M
CPU speed: 2260.21 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 24.6, RdtscTiming=1
Best time for 512K FFT length: 28.463 ms.
Best time for 640K FFT length: 39.035 ms.
Best time for 768K FFT length: 47.609 ms.
Best time for 896K FFT length: 57.863 ms.
Best time for 1024K FFT length: 64.360 ms.
Best time for 1280K FFT length: 83.233 ms.
Best time for 1536K FFT length: 100.096 ms.
Best time for 1792K FFT length: 122.372 ms.
Best time for 2048K FFT length: 136.048 ms.

I'm sure u know but..I failed big time under version 23.x..
But with 24.6..i pass for hours

I was going to even replace my whole system because I was failing..
Good thing I only sold my motherboard..Lol

[Ethan (EO)]

Originally Posted by Prime95

Right now I'm in the midst of porting the FFT code to 64-bit Windows. The Opteron actually passed a torture test last week! And no, it isn't any faster than 32-bit mode. First step is to port the existing assembly code, second step is to make use of the extra SSE2 registers. But right now, I'm rewriting the trial factoring code - at least that code should benefit greatly from 64-bit instructions.

Great to hear that the 64 bit development is progressing along with all the work that's been going into gwnum/FFT work.

Let me know if you need any test / double check runs on Athlon64s --
I have Server 2003 x64 and XP x64 machines eager to contribute their
XMM8 through XMM15 to the QA cause

Ethan O'Connor