V24.12 release candidate 3

[kjaget]

Originally Posted by Prime95

Can you follow the instructions in this thread, http://www.mersenneforum.org/showthread.php?t=4202, and post your Celeron D benchmarks. I thought the Celeron D would behave like my Willamette, but apparently it does not.

OK, here's the results from the run. Let me know if there is anything else I can do to help.

Kevin

[Prime95]

Originally Posted by kjaget

OK, here's the results from the run. Let me know if there is anything else I can do to help.

Kjaget, can you run wcpuid (a freeware program) and choose view / cache info. Under "L2 cache information" does it say 4-way set associative or 8-way set associative?

[kjaget]

Originally Posted by Prime95

Kjaget, can you run wcpuid (a freeware program) and choose view / cache info. Under "L2 cache information" does it say 4-way set associative or 8-way set associative?

As before, it will take a few days before I'm at the machine again. I'll run wcpuid and report the results the next time I'm there (this weekend?).

If you believe Intel's data sheet (and assuming I found the correct reference), though, it should be 4-way, 256K L2. http://download.intel.com/design/cel...s/30235305.pdf, page 9. Not sure if this matches the numbers you are seeing or not, though, so I'll get the wcpuid results to you soon.

Kevin

[akruppa]

On an Opteron 150,

AdvancedFactor=1021,200000,32,32

does not seem to find any factors. In fact, it does not seem to even look for them... however, using

AdvancedFactor=1021,200000,32,33

finds many factors smaller than 2^31.

Is there a way to make AdvancedFactor not stop at the first factor found for each number, but keep looking for all factors in the specified range?

Alex

[drew]

Originally Posted by filidoor

I just switched from RC1 to RC3 to resolve a bug, but I still have the problem, so I report it here. When starting P1-factoring stage 2, v.24.12 crashes when there is unsufficient memory. Changing DayMemory = 48 to 96 solved the problem, but the bug is still there.

How much memory does it take to gaurantee this will work? My test is due to go into stage 2 later today, and I want to make sure it doesn't cause any problems. Should I stop it and wait for RC4 to be safe?

[Prime95]

Originally Posted by akruppa

AdvancedFactor=1021,200000,32,33
finds many factors smaller than 2^31.

Is there a way to make AdvancedFactor not stop at the first factor found for each number, but keep looking for all factors in the specified range?

I don't think there is a way to find all factors below 2^32. Prime95 is not particularly efficient with these small factors. I'm guessing a GMP program would be just as fast.

[Prime95]

Originally Posted by kjaget

If you believe Intel's data sheet, though, it should be 4-way, 256K L2.

Good find. I looked at the web site and didn't find this.

I'm trying to figure out why your Celeron D and my Willamette are getting different optimal FFTs. Your 4-way and my 8-way is one difference (implying prime95 is not spreading the FFT data in the L2 cache optimally). Yours is probably DDR memory and mine is Rambus, so I have longer latencies. Yours is a Prescott core and mine is Willamette so it could be due to architectural changes.

If the problem is due to FFT data layout in the L2 cache then it could have implications for other CPUs too - making this the most serious problem I've looked into with the 24.12 release candidates.

[Prime95]

Originally Posted by drew

How much memory does it take to gaurantee this will work? My test is due to go into stage 2 later today, and I want to make sure it doesn't cause any problems.

I'm not sure. It depends on the FFT size. Probably 70MB should do the trick.

[sHORTY]

Since I've started on this new version I have been receiving 'hardware errors' once or twice a day.......

[Wed Jun 29 14:21:15 2005]
Iteration: 15130634/28114133, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 8.573808511891147e+016 != 8.573808511751453e+016
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.
[Wed Jun 29 15:33:40 2005]
Iteration: 15160329/28114133, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 7.408951430781206e+016 != 7.408951430777005e+016
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.
[Thu Jun 30 01:03:20 2005]
Iteration: 15570088/28114133, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 2656750288827555 != 2656750288652443
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.
[Thu Jun 30 08:43:36 2005]
Iteration: 15935105/28114133, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 5.794551176235642e+016 != 5.79455117624411e+016
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.
[Thu Jun 30 11:55:10 2005]
Iteration: 16067021/28114133, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 5796225627663532 != 5796225628362573
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.


Running WinXP64 - 24.12RC3 64bit
AMD Athlon(tm) 64 Processor 3200+
CPU speed: 2210.80 MHz
CPU features: RDTSC, CMOV, Prefetch, 3DNow!, 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]

Originally Posted by sHORTY

Since I've started on this new version I have been receiving 'hardware errors' once or twice a day.......

Looks like hardware trouble to me. Run a memory test, check your cooling, check your voltages, relax the memory timings, run a torture test, etc. etc. until you can find the culprit.

Note that many have reported problems with the 90nm AMD chips even running at stock speed. If you find the CPU is at fault and cannot RMA it, then you'll need to underclock it.

[sHORTY]

Odd that the problem doesn't show up until 24.12, however.

Ran for for over 2 months straight with previous builds and no problems. I will run some tests and see what I can find, however.