[QUOTE]We noticed, while setting up a few boxes, that Mprime embeds a weird "^F" character into prime.txt if you use a proxy.

We do not know if this causes any problems or anything. It is easy to delete with vim.[/QUOTE]FWIW, the 32-bit version of Mprime does not do this.

The problems continue
 

Multithreaded, not hyper-threaded.

[code][Work thread Aug 8 19:15:25] Setting affinity to run helper thread 1 on logical CPU #2
[Work thread Aug 8 19:15:25] Starting primality test of M... using AVX FFT length 2240K, Pass1=448, Pass2=5K, 2 threads
[Work thread Aug 8 19:15:25] Iteration: 2/..., ERROR: ROUND OFF (1.770880676e+21) > 0.40[/code]

It took 15 minutes to get to iteration 642 before I noticed, and the round-off never changed from the ridiculous value there. It happened more than once for each iteration, and around 1/6 of the errors gave the reproducible/not hardware message.
[code]Linux64,Prime95,v27.7,build 2[/code]

I'll upgrade to the most recent build.

We are going to switch to one worker per core for now.

:ouch:

I ran a new test with 3200K FFT (SB 2600K with Hyper-Threading : 8 worket threds) :

CPU overclocked @ 4300MHz, Windows XP 32-bit, Prime95 1600MB tested -> Pass (running > 8h)

On Windows OS, it seems to be memory related : less memory passes but 3GB crash even at stock base clock.

[QUOTE=Laurent;307331]...
Memory is OK (8 GB physical tested with Memtest > 48h)
...

[/QUOTE]

[QUOTE=Laurent;307444]I ran a new test with 3200K FFT (SB 2600K with Hyper-Threading : 8 worket threds) :

CPU overclocked @ 4300MHz, Windows XP 32-bit, Prime95 1600MB tested -> Pass (running > 8h)

On Windows OS, it seems to be memory related : less memory passes but 3GB crash even at stock base clock.[/QUOTE]

I have a question, maybe I missed something. Originall you posted 8Gb of memory, correct? You seem to run Linux in 64 bit mode, but WinXP in 32 bit mode. With 8Gb of RAM in WinXP 32, the usable limit is around 3Gb of RAM. Maybe there is a conflict with the motherboard with 8Gb and WinXP 32? Can you install WinXP 64 or Vista 64 or Win7 64 and try it again?

[FONT=&quot]Yes, I have 8GB of physical memory installed (DDR3 certified at 1600MHz 8-8-8-24 2T).[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]I don't have access to a 64-bit version of Windows.[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]Effectively, 32-bit Microsoft systems don't use pagination extension (PAE) and physical address space is about 3.5 GB. I don't think it's the cause of the problem: it is just physical memory which is not usable (waste).[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]If I disable Hyper-threading, the 3200K FFT test passes with about 3GB of used memory.[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]Why is the 3200K test so aggressive?[/FONT]
[FONT=&quot] [/FONT]

Start Prime95 on mac automatically (before user logs in)
 

Is it possible to start the Mac version for Prime95 in the background (before any user logs on)?

[QUOTE=koekie;307903]Is it possible to start the Mac version for Prime95 in the background (before any user logs on)?[/QUOTE]

To my knowledge it is not possible. Is this important to you? If so, I can probably make an mprime for Mac version. You'd lose the GUI.

I thought there *is* a Mac version of Mprime, albeit an old one.

[QUOTE=Prime95;307910]To my knowledge it is not possible. Is this important to you? If so, I can probably make an mprime for Mac version. You'd lose the GUI.[/QUOTE]

It would certainly be useful since my macs boot to a login screen. The lack of gui is not an issue for me, in the past I used the Linux mprime quite a lot as well. So for me an mprime client for Mac would be a welcome addition, but I would say only if it's not to much work for you (unless there are more people feeling this is useful).

Thanks for your quick reaction and thinking along.

[QUOTE=koekie;308014]It would certainly be useful since my macs boot to a login screen. The lack of gui is not an issue for me, in the past I used the Linux mprime quite a lot as well. So for me an mprime client for Mac would be a welcome addition, but I would say only if it's not to much work for you.[/QUOTE]

Well, the old makefile worked. I uploaded the mprime version to [url]ftp://mersenne.org/gimps[/url]. I didn't test it, but you are welcome to try it. If you get the pesky "untrusted version" message let me know.

I got the untrusted error:
[Comm thread Aug 15 21:33] PrimeNet error 9: Access denied
[Comm thread Aug 15 21:33] Untrusted program versions currently excluded by PrimeNet

Besides that the executable appears to work.

[QUOTE=koekie;308064]I got the untrusted error[/QUOTE]

Try again

Now it works! Thank you for the quick service.

Recent versions of MacOS are based on UNIX. So it should be possible to start Prime95 from a system startup script in /etc/rc.d (exact location may vary). I'm a UNIX system admin, I'm sure I could do it if Prime95 will run without the GUI, but I don't have a Mac so can't give more precise instructions.

Chris

People talking about build... 4? Where? I just found out that the last build I use is 2. I saw someone talking about 27.7.4 but I first thought it was a typo...

[QUOTE=LaurV;308244]People talking about build... 4? Where? I just found out that the last build I use is 2. I saw someone talking about 27.7.4 but I first thought it was a typo...[/QUOTE]

Look at the second post in this thread. You'll see that very, very few users would care about the 3 minor bugs fixed since build 2. In fact, build 3 and 4 may not be available for every OS.

Thanks for clarifying. Then I feel convenient with what I have.

I haven't been keeping up on the software threads lately and had a couple of questions. What's the main difference between 27.6 and 27.7 and is there a pontential problem if I continue using 27.6? Am I correct in thinking that 27 is only for IB/SB systems and should I stick with 26.6 on non IB/SB systems or upgrade to 27.7? Thanks in advance.

[QUOTE=bcp19;308274]I haven't been keeping up on the software threads lately and had a couple of questions. What's the main difference between 27.6 and 27.7 and is there a pontential problem if I continue using 27.6? Am I correct in thinking that 27 is only for IB/SB systems and should I stick with 26.6 on non IB/SB systems or upgrade to 27.7? Thanks in advance.[/QUOTE]

If in doubt, check the first post.

[QUOTE=Prime95;298381]
This version also fixes a few rare bugs in 27.6 (see [url]http://mersenneforum.org/showpost.php?p=297258&postcount=2[/url]). It also fixes a bug partly responsible for the rare false positive reports GIMPS has seen in recent years. As part of this fix, a save file that cannot be read will be renamed with a .bad extension in hopes that the file can be successfully read at a later time (yes, this can happen).
[/QUOTE]

There is a bug in the 27.7 Build 2.

Haxacore processor with hyperthreading enabled.
AffinityScramble2=4a28065b3917

The worker number two is supposed to use the cpu number "a" (10, zero based; or 11 in the GUI), insted this is what I see, and the worker jumps from one CPU to another:

[Sep 10 08:54] Waiting 5 seconds to stagger worker starts.
[Sep 10 08:54] Worker starting
[Sep 10 08:54] Setting affinity to run worker on logical CPUs Ç

That should be "A" and "B", not "a" and "b", if memory serves me.

[QUOTE=TObject;311000]There is a bug in the 27.7 Build 2.

Haxacore processor with hyperthreading enabled.
AffinityScramble2=4a28065b3917

The worker number two is supposed to use the cpu number "a" (10, zero based; or 11 in the GUI), insted this is what I see, and the worker jumps from one CPU to another:

[Sep 10 08:54] Waiting 5 seconds to stagger worker starts.
[Sep 10 08:54] Worker starting
[Sep 10 08:54] Setting affinity to run worker on logical CPUs Ç[/QUOTE]

That's [URL="http://www.mersenneforum.org/showthread.php?p=300391#post300391"]known[/URL] and [URL="http://www.mersenneforum.org/showthread.php?p=298383#post298383"]fixed[/URL]. Get the latest build.

[QUOTE=Dubslow;311017]That's [URL="http://www.mersenneforum.org/showthread.php?p=300391#post300391"]known[/URL] and [URL="http://www.mersenneforum.org/showthread.php?p=298383#post298383"]fixed[/URL]. Get the latest build.[/QUOTE]


Thank you

I downloaded again. It is the same Build 2.

Where are the new builds for Windows 64-bit?

[QUOTE=TObject;311020]Thank you

I downloaded again. It is the same Build 2.

Where are the new builds for Windows 64-bit?[/QUOTE]

The formatting changed between build 2 and build 4. p95v277.zip is no longer the correct format; instead look for p95v277.win64.zip.

(Alternately just use the links in the [URL="http://www.mersenneforum.org/showthread.php?t=16779"]head post[/URL].)

[QUOTE=TObject;311020]Thank you

I downloaded again. It is the same Build 2.

Where are the new builds for Windows 64-bit?[/QUOTE]

[url]http://mersenneforum.org/showthread.php?p=298397#post298397[/url]

edit: beat me to it

What is the file name?

p95v277.win64.zip is build 2

[QUOTE=Dubslow;311023]The formatting changed between build 2 and build 4. p95v277.zip is no longer the correct format; instead look for p95v277.win64.zip.

(Alternately just use the links in the [URL="http://www.mersenneforum.org/showthread.php?t=16779"]head post[/URL].)[/QUOTE]

Thank you, but it still says "build 2" in the Help-About Prime95... box. Is the about box wrong?

[QUOTE=TObject;311027]What is the file name?

p95v277.win64.zip is build 2[/QUOTE]

Hmm... Prime95, it appears the links need updating.

[QUOTE=TObject;311029]Thank you, but it still says "build 2" in the Help-About Prime95... box. Is the about box wrong?[/QUOTE]
Wouldn't be the first time :smile:

The following local.txt line in v27.7 build 2 does not work as expected.

MaxHighMemWorkers=0 during 8:25-20:35 else 1

Prime95 runs a high memory worker between 8:25-20:35, when I was hoping it would only run it 20:35-8:25.

[QUOTE=TObject;311678]The following local.txt line in v27.7 build 2 does not work as expected.

MaxHighMemWorkers=0 during 8:25-20:35 else 1

Prime95 runs a high memory worker between 8:25-20:35, when I was hoping it would only run it 20:35-8:25.[/QUOTE]

So then do

MaxHighMemWorkers=1 during 8:25-20:35 else 0

...it's basic logic.

The basic problem is that MaxHighmemWorkers=0 does not mean no workers, but the default number of them.

And last I checked there still was a race condition which would make Prime95 exceed the allowed number at times. All in all, you shouldn't do high memory work unless you are prepared for all workers which have such work assigned to enter high memory stage at the same time.

[QUOTE=ckdo;311711]The basic problem is that MaxHighmemWorkers=0 does not mean no workers, but the default number of them.[/QUOTE]As a workaround, in local.txt you can put, at the bottom under each [Worker #?] section:[code][Worker #1]
Memory=1 during 8:25-20:35 else 2000

[Worker #2]
Memory=1 during 8:25-20:35 else 2000

etc[/code]1MB of memory will be insufficient to run stage2 and therefore it will skip until later when it has more memory available.

Hello I know a prime exponent of which has a length of 10 characters. Is it possible to remove the restriction on the length of the exponent in Prime95 to check my number?
P.S. Sorry for my English. I am from Russia. :smile:

[QUOTE=west0293;311728]Hello I know a prime exponent of which has a length of 10 characters. Is it possible to remove the restriction on the length of the exponent in Prime95 to check my number?
P.S. Sorry for my English. I am from Russia. :smile:[/QUOTE]

The reason or the restriction is that Prime95 simply can't test numbers that large. The reason it can't test those numbers is because it would take many years (possibly >100 years or even >1000, depending) to test such a number on today's hardware.

48 cores
 

My new toy is composed of 48 (forty-eight) K10 cores and more than 100 GB of RAM, but when I'm trying to start prime95, I become:
[CODE][Comm thread Nov 7 14:28] Exchanging program options with server
[Comm thread Nov 7 14:28] PrimeNet error 7: Invalid parameter
[Comm thread Nov 7 14:28] parameter nw: Invalid int value/precision '48'
[/CODE]I think the number of cores is out of the PrimeNet authorised range. Could you change the limit? By the way, which type of work should I assign this type of machine?

Thank you,
Kyle

P-1 on all cores sounds tempting.

[QUOTE=ckdo;317412]P-1 on all cores sounds tempting.[/QUOTE]But I'd say impractical. Doing a current P-1 (~60M) can use up to about 12GB if running stage2 in one pass, so 8 instances in stage2 can fully utilize available RAM. I'd say a mix of 10x P-1 and 14x L-L (or DC) would be viable. If those are 48 "real" cores and not "half-cores" then run 2 Prime95 instances, maybe with 5x P-1 + 19x L-L/DC in each. I haven't kept up on the current consensus regarding Prime95-on-AMD.

I wouldn't be surprised if memory bandwidth limitations play a part in overall throughput limits.

[url]http://www.mersenneforum.org/showthread.php?t=16865[/url]

[QUOTE=James Heinrich;317417]But I'd say impractical. Doing a current P-1 (~60M) can use up to about 12GB if running stage2 in one pass, so 8 instances in stage2 can fully utilize available RAM. I'd say a mix of 10x P-1 and 14x L-L (or DC) would be viable. If those are 48 "real" cores and not "half-cores" then run 2 Prime95 instances, maybe with 5x P-1 + 19x L-L/DC in each. I haven't kept up on the current consensus regarding Prime95-on-AMD.

I wouldn't be surprised if memory bandwidth limitations play a part in overall throughput limits.[/QUOTE]
I launched 24 workers with multithreading on two cores: 5x P-1 and 19x L-L/DC.
This are 4 AMD Opteron 6176, so yeah, I think memory bloat will arise.
It is getting really noisy there... It needs between 0.06 and 0.12 s to make an L-L iteration... :s

[QUOTE=Kyle;317403]I think the number of cores is out of the PrimeNet authorised range. Could you change the limit? [/QUOTE]

The limit has been upped on the server.

I tried to verify the P-1 stage 1 residue with high B1 by computing the product of primes and prime powers <= B1 modulo the order of 3 in (Z/pZ)* for known prime factors p of the number. This works ok, except when B1 > 2^32. It appears that some composites are included among the "primes" used in stage 1, more specifically, composites with no prime divisors <65539. E.g., a hack attacked version of mprime that prints each prime by which it exponentiates lists
4295360521
4295622677
4295884849
4296146989
4296409193
4296540223
4296802451
4296933457
etc.

I could take these extra values into account very easily, but the P-1 save file I started from was generated over many years, using many different versions of mprime. Have all past versions of mprime that allowed B1>2^32 chosen the stage 1 primes-or-almost-primes is this way?

[QUOTE=akruppa;318368]Have all past versions of mprime that allowed B1>2^32 chosen the stage 1 primes-or-almost-primes is this way?[/QUOTE]

Almost certainly.

Thanks, the check works when including the rough numbers.

Note that for large B1, this introduces a significant number of composite values. For example, with B1=10^11, 15% of the values by which we exponentiate in stage 1 are composites, for B1=10^12 the fraction increases to 25%, with B1=10^13 to 32%.

Question: Will P95 stop the ECM (and switch to the next assignment) if a factor is found (in stage 1? in stage 2?), or it will continue until the number of pledged curves is finished?

(edit: doing 150 curves for each of the first 5 exponents over 7M, currently at curve 136, 138, 141, 139 and 0 respectively, on a 4 core machine)

It will stop unless the factor found was supplied as part of the known_factors list for the assignment.

AffinityScramble2 doesn't work on Bulldozer
 

Hi,

Mprime v27.7 (Linux 64bit) doesn't honor AffinityScramble2 on Bulldozer CPUs? I didn't try older mprime versions

local.txt[CODE]
[...]
WorkerThreads=1
ThreadsPerTest=2
Affinity=0
AffinityScramble2=24
[...]
[/CODE]
works fine on all CPUs I've tested except Bulldozer CPUs? It works as expected on the older Opteron 61xx and on all Xeon systems I've tested.

[CODE]
Mersenne number primality test program version 27.7
Optimizing for CPU architecture: AMD Bulldozer, L2 cache size: 2 MB, L3 cache size: 12 MB
Starting worker.
Worker starting
Setting affinity to run worker on logical CPU [COLOR="Red"][B]#1[/B][/COLOR]
Optimal P-1 factoring of M59xxxxxx using up to 59615MB of memory.
Assuming no factors below 2^73 and 2 primality tests saved if a factor is found.
Optimal bounds are B1=555000, B2=11377500
Chance of finding a factor is an estimated 3.76%
Using AMD K10 FFT length 3200K, Pass1=640, Pass2=5K, 2 threads
Setting affinity to run helper thread 1 on logical CPU [COLOR="Red"][B]#2[/B][/COLOR]
[/CODE]
Any idea how to debug? I quick view into the sources didn't help, I've tried DebugAffinityScramble=2 but got no additional screen outputs.

Oliver

In my reading of the code AffinityScramble2 is only used on machine that support hyperthreading. Try setting NumCPUs=n and CpuNumHyperthreads=2
in local.txt

Adding "CpuNumHyperthreads=2" to local.txt is enough.
AffinityScramble2 is working fine on Opteron 61xx which don't have Hyperthrashing.

Oliver

[QUOTE=TheJudger;319524]
AffinityScramble2 is working fine on Opteron 61xx which don't have Hyperthrashing.[/QUOTE]

It shouldn't. Does Options/CPU report a hyperthreaded CPU?

My fault (wrong observation), on Opteron 61xx AffinityScramble2 is ignored, too. Adding CpuNumHyperthreads=2 to local.txt does the trick, too.

Oliver

26.6 and 27.7 x64 torture test difference
 

I'm using x64 versions of Prime95. There is difference in torture test memory consumption between 26.6 and 27.7.
Sometimes on 4GB Windows 7 x64 SP1 (without any updates, high perf profile) 26.6 can eat up to 3,29GB of 3,5GB usable memory (4GB installed but 512MB used by int video) in Blend test, 27.7 can occupy only 1.6GB, or 26.6 can't start the 1st of 4 threads (Starting, Self-Test, Self-test, Self-test) while using all 4 cores.
Also under 26.6 Blend test computer is barely usable, perhaps because of no free memory. I don't remember I've seen such sluggish system, it seems like frozen at first, but responds in 5-10 seconds on your actions. It can take some time to stop the test. Under 27.7 Blend test it can run unnoticed. Unfortunately it doesn't happens all the time. Frequently 26.6 starts testing consuming 1,6GB of memory and running 100% CPU unnoticed like 27.7.
I haven't found changes in whatsnew.txt regarding tests, except one: "Multi-threaded tests might be a little bit faster especially when using a lot of threads."
Do you suggest using 27.7 for AMD K10?
Can I help narrowing down the issue?

results.txt of 27.7 is:
AMD Athlon(tm) II X4 640 Processor
CPU speed: 3013.38 MHz, 4 cores
CPU features: 3DNow!, 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: 48
L2 TLBS: 512
Prime95 64-bit version 27.7, RdtscTiming=1

26.6:
AMD Athlon(tm) II X4 640 Processor
CPU speed: 8863.04 MHz, 4 cores
CPU features: 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: 48
L2 TLBS: 512
Prime95 64-bit version 26.6, RdtscTiming=1

When version 26.6 uses 3.2GB of memory and your system slows to a crawl - this is not a good thing. Windows is likely thrashing -- paging memory to and from disk. This may be stressful to the disk, but not the CPU. Thus, you are not getting a good torture test of the CPU.

Somewhere between 26.6 and 27.7 either a bug was fixed or the algorithm for computing the default amount of memory to use in a blend torture test changed.

I suggest using 27.7. If you want to try to get prime95 to use more than 1.6GB, try using a custom torture test giving prime95 more memory to use.

I am getting the following on every iteration:

[Dec 7 16:29] Iteration: 1193000 / 29810063 [4.00%]. Per iteration time: 0.055 sec.
[Dec 7 16:29] Possible hardware errors have occurred during the test!
[Dec 7 16:29] 2 ROUNDOFF > 0.4 of which 1 were repeatable (not hardware errors).
[Dec 7 16:29] Confidence in final result is fair.

Prime95 v27.7, build 2 on a box that I have not had any bad results. The other three workers are fine.

How do I specify a larger FFT? Do I stop and start over again since I'm only at 4.00%. Thanks!

Can you post the roundoff error reports from results.txt?
Some roundoff values (such as 0.4375) are really OK when running near the upper limit of an FFT size.

If you have one of the suspicious error values (0.5 or very close to it) then I'd restart the test since you're only at 4%.

Hi George,

Very strange:

Iteration: 2/29810063, ERROR: ROUND OFF (51214.66385) > 0.40
Continuing from last save file.
Iteration: 2/29810063, ERROR: ROUND OFF (51214.66385) > 0.40
Continuing from last save file.
Disregard last error. Result is reproducible and thus not a hardware problem.
For added safety, redoing iteration using a slower, more reliable method.
Continuing from last save file.

No need to restart -- the two previous restarts started from scratch.

This is another instance of the rare "huge roundoff" error. I've been unable to guess at the cause.

Ok, thanks, we'll see if it matches when complete.

[QUOTE=richs;320917]Ok, thanks, we'll see if it matches when complete.[/QUOTE]
It won't, or it will take ages to complete, almost all iteration will be done with the slower method. Are you AVX and multicore? (i.e. one worker uses many cores?). If so, try to go single core (one worker one core). And see what's happen. I had the same problem when switched to avx version (there is a discussion here around, and that time George asked me this question; going to single core workers solved the issue).

I too would like to know if you are doing a multithreaded LL test. I'm gathering clues. Your clue that the bug can happen on the very first LL iteration could be an important one.

This thread is fascinating to observe. :popcorn: Please carry on! (No disrespect intended)

It's an Intel Core i5-2500 @ 3.30GHz with AVX that I upgraded to v27.7 build 2 a couple of weeks ago. Prime95 runs 4 workers on this box. This error occurred on worker 3 and is the first new exponent that this worker has started since upgrading to v27.7.

See posts #69-#77, on the current thread. For me the problem was certainly solved by using single-core workers and disabling the HT. Never appeared since, it is related to some initialization of variables when multithreading, and it is reproducible, if I try to use "one worker 8 cores (HT)" or "one worker 4 cores (no HT)" or "2 workers 2 cores each (no HT)" the error pops up immediately. It is not temperature related. It may be OC-related, but this I can't swear. (edit: i7-2600k 4 phys cores)

We have had errors with multi-threaded P-1 and LL tests as well.

We turned in a pile of legit multi-threaded P-1 and LL (DC) results but occasionally a box would spaz and throw a ton of errors. We never tested single-thread instances but it definitely happens with multi-thread instances.

And not all of the time. We had maybe 5 errors out of 100 tests or something like that.

We think it might be related to the pause function since we noticed the error after Mprime kicked back in after being paused. But maybe that was just because we were at the console when the other job finished so it was easy to notice.

More: [URL]http://www.mersenneforum.org/showthread.php?t=17033[/URL]

Xyzzy: yepp, I think this has to do with multithreading, too.

I'm doing some multithreaded P-1 (one worker per mprime process, multiple mprime processes per machine), everytime I add new work via worktodo.add I've to check whether the workers keep going or end in a endless loop (two different symptoms [SUP]*1[/SUP]).[LIST][*]1 thread per worker: never happened[*]2 threads per worker: very, very rare[*]4 threads per worker: sometimes[*]8 threads per worker: often (>80%)[/LIST]
While mprime is running this can happen when starting a new exponent (start stage #1) or when switching from stage #1 to stage #2 (start stage #2), too.

Linux, mprime 27.7 64bit, 1 thread per core, running on a network share (NFS).
This happens with AVX code (Sandy Bridge), I can't exactly remember about SSE code.

I've the feeling that it happens more often on faster CPU.

Oliver

[SUP]*1[/SUP][LIST][*]endless loop: "SUMOUT error occurred"; wait 5 minutes[*]endless loop: "Possible roundoff error (<some number>), backtracking to last save file."[/LIST]

Me too....WAY, WAY > 0.4 :(
 

I upgraded to 27.7 and as soon as it started only the second core got this....

[QUOTE=Prime95;320916]No need to restart -- the two previous restarts started from scratch.

This is another instance of the rare "huge roundoff" error. I've been unable to guess at the cause.[/QUOTE]

[CODE][Mon Dec 10 08:26:05 2012]
Iteration: 22131840/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131832/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131824/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131816/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131808/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Disregard last error. Result is reproducible and thus not a hardware problem.
For added safety, redoing iteration using a slower, more reliable method.
Continuing from last save file.
Iteration: 22131840/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131832/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131824/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Iteration: 22131816/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Disregard last error. Result is reproducible and thus not a hardware problem.
For added safety, redoing iteration using a slower, more reliable method.
Continuing from last save file.
Iteration: 22131817/26871127, ERROR: ROUND OFF (3.337928738e+063) > 0.40
Continuing from last save file.
Disregard last error. Result is reproducible and thus not a hardware problem.
For added safety, redoing iteration using a slower, more reliable method.
Continuing from last save file.[/CODE]

Then I went back to the old version to finish the exponent already over 80% and got...

[CODE][Dec 10 08:30] Waiting 8 seconds to stagger worker starts.
[Dec 10 08:30] Worker starting
[Dec 10 08:30] Setting affinity to run worker on logical CPUs 2,3
[Dec 10 08:30] Setting affinity to run helper thread 1 on logical CPUs 2,3
[Dec 10 08:30] Resuming primality test of M26871127 using Core2 type-3 FFT length 1440K, Pass1=320, Pass2=4608, 2 threads
[Dec 10 08:30] Iteration: 22131819 / 26871127 [82.36%].
[Dec 10 08:30] Possible hardware errors have occurred during the test:
[Dec 10 08:30] 10 SUM(INPUTS) != SUM(OUTPUTS) of which 3 were repeatable (not hardware errors).
[Dec 10 08:30] Confidence in final result is very poor.[/CODE]

Core i5-2520M Laptop. No OC.

Not sure now which version I should use??? or is this test is doomed to be flagged "suspect" or "bad". This LapTop has not had a bad test or errors before this.

I tried to turn off Hyperthreading but using CPUs to use (Multithreading) = 1 but I still get:

[CODE][Dec 10 08:30] Waiting 8 seconds to stagger worker starts.
[Dec 10 08:30] Worker starting
[[U]Dec 10 08:30] Setting affinity to run worker on logical CPUs 2,3
[Dec 10 08:30] Setting affinity to run helper thread 1 on logical CPUs 2,3
[Dec 10 08:30] Resuming primality test of M26871127 using Core2 type-3 FFT length 1440K, Pass1=320, Pass2=4608, 2 threads[/U][Dec 10 08:30] Iteration: 22131819 / 26871127 [82.36%].
[Dec 10 08:30] Possible hardware errors have occurred during the test:
[Dec 10 08:30] 10 SUM(INPUTS) != SUM(OUTPUTS) of which 3 were repeatable (not hardware errors).
[Dec 10 08:30] Confidence in final result is very poor.[/CODE]

[QUOTE=petrw1;321186]
Not sure now which version I should use??? or is this test is doomed to be flagged "suspect" or "bad". This LapTop has not had a bad test or errors before this.[/QUOTE]

Finish off the test. I suspect the result is OK. Tell us if it successfully completes the double-check.

I'll know by morning

Assuming it tests out fine I'm really annoyed at myself!!!
I realized a few weeks after I installed V26 late June on this PC that it was a Sandy but my Frontal Lobe short circuited and forgot to tell my Cerebral Cortex that with a Sandy I need to upgrade to V27.

Well I did today and found iteration times drop from 20 to 13 ms for a DC.
So in the last 5 months I missed out on over 20 DC's. :sad:

Use the AVX but kill the helper threads. Let only so many workers as many phys core you have. The additional HT cores produce just heat and "bad confidence" stuff. At least till version 28, when George finds what's going on :smile: One worker, one physical core. This solved the issue for me.

[QUOTE=LaurV;321232]Use the AVX but kill the helper threads. Let only so many workers as many phys core you have. The additional HT cores produce just heat and "bad confidence" stuff. At least till version 28, when George finds what's going on :smile: One worker, one physical core. This solved the issue for me.[/QUOTE]

Thanks
I have 2 cores ...
2 workers
1 core each

And it says 2 logical cores make up 1 physical core and still give me a helper thread with each.

[QUOTE=petrw1;321242]Thanks
I have 2 cores ...
2 workers
1 core each

And it says 2 logical cores make up 1 physical core and still give me a helper thread with each.[/QUOTE]
:shock::shock: Never seen that king of bug. Are you sure you set your workers right? Check the test/workers if you have "number of worker windows to run"=2, "worker number"="all workers", "CPUs to use"=1. And check if you don;t have strange settings in prime.txt and local.txt related to this.

[QUOTE=LaurV;321244]:shock::shock: Never seen that king of bug. Are you sure you set your workers right? Check the test/workers if you have "number of worker windows to run"=2, "worker number"="all workers", "CPUs to use"=1. And check if you don;t have strange settings in prime.txt and local.txt related to this.[/QUOTE]

Disable hyper-threading in BIOS and see if the problem persists.

[QUOTE=kladner;321247]Disable hyper-threading in BIOS and see if the problem persists.[/QUOTE]

That is off limits. My work laptop.

Oh well. It is understandable.

[QUOTE=Prime95;321201]Finish off the test. I suspect the result is OK. Tell us if it successfully completes the double-check.[/QUOTE]

Yup. DC residue matches. Next DC is not reporting errors.

[QUOTE=Prime95;320916]This is another instance of the rare "huge roundoff" error. I've been unable to guess at the cause.[/QUOTE]

I think I found it. Would anyone like to download version 27.9 from [url]ftp://mersenne.org/gimps[/url] and see if it is fixed? Thanks.

Yuck! :yucky:
Only my grandma is missing from that folder. Maybe some supermod can create an "old_versions" folder and move into it everything except the (all flavor of the) last version? It is very confusing for greenies, especially as the alphabetical sort does not match neither the version numbering nor the time line of development, even if the page is ftp-aware to be accessed/sorted locally with an ftp client like totalcmd... still messy.

[QUOTE=LaurV;321494]Yuck! :yucky:
Only my grandma is missing from that folder.[/QUOTE]

Look for the new naming scheme. All files begin with p95v279.

[QUOTE=Prime95;321498]Look for the new naming scheme. All files begin with p95v279.[/QUOTE]
We knew that. It was just a scream about the mess in the folder, without any intention to cloud the sky. We already installed it and it runs with no errors since we posted, 4 cores on a 50M expo. We still wait for the errors to come, just in case they are lost somewhere in the computer's intestines, but for now we would be tempted to say: Good job! What the error was? (not that I would understand, this is more a rhetorical question :razz:). This part for sure will make some laptop owners very happy:

[QUOTE=undoc.txt]WINDOWS ONLY: You can have prime95 pause when the battery level gets low.
This will allow you to charge the battery faster. Add the line
BatteryPercent=n
to prime.txt, where n is a number from 0 to 100. The default value is zero.
[/QUOTE]

[url]http://www.mersenneforum.org/gimps/[/url]

We have the option of adding text to the header. Perhaps a list of file descriptions or maybe a list of the most common files?

[QUOTE=Xyzzy;321509][url]http://www.mersenneforum.org/gimps/[/url]

We have the option of adding text to the header. Perhaps a list of file descriptions or maybe a list of the most common files?[/QUOTE]

Here's a shocking new concept: Subfolders! :shock:

[QUOTE=axn;321512]Here's a shocking new concept: Subfolders! :shock:[/QUOTE]
That is what I suggested. There is already a [sub]folder inside, so it would be no problem to create more, like [older_versions\[windows,linux,mac,sources],current[[windows,linux,mac,sources]]. I can do it in few seconds with total commander, but I have no rights (I don't want this to sound as a request, but as a proof of how easy it is :razz:)

We mirror the official site so that the forum's mirror can be used in rotation with the official mirror. (Does that make sense?)

In other words, changes to the actual structure are up to George.

That said, a note at the top might be helpful. If someone wrote one and posted it here we would see to it that it got included in the header.

:mike:

New message to me....
 

[CODE][Main thread Dec 13 08:21] Mersenne number primality test program version 27.7
[Main thread Dec 13 08:21] Optimizing for CPU architecture: Core i3/i5/i7, L2 cache size: 256 KB, L3 cache size: 3 MB
[Main thread Dec 13 08:21] Unable to detect which logical CPUs are hyperthreaded.
[Main thread Dec 13 08:21] Assuming logical CPUs 1 and 2, 3 and 4, etc. are each from one physical CPU core.
[Main thread Dec 13 08:21] To the best of my knowledge this assumption is only valid for Microsoft Windows.
[Main thread Dec 13 08:21] To override this assumption, see AffinityScramble2 in undoc.txt.
[Main thread Dec 13 08:21] Starting workers.[/CODE]

Every other time before this that it started up it said Logical cores 1 and 2 make 1 physical core and 3 and 4 makes up the other.

However, the above assumption did not affect the performance. Makes sense since there really are only 2 physical cores.

That happens for me maybe every 1 of 10 starts.
If I understood correctly, it runs a quick timing test on a few threads and compares performance to a single thread: if the second thread approximately doubles throughput then the second thread is running on a "real" core. If performance is only marginally different then it's a "fake" (hyperthreaded) core. If the results don't fall within the result margins of either scenario then you get the above error. This can happen depending on what else is running on the CPU at the time Prime95 starts up.

My test of 29810063 with the huge round-off errors completely successfully. No errors are occurring on the next exponent that worker is running.

[QUOTE=LaurV;321505]Good job! What the error was?[/QUOTE]

As always, programmer error. I did a memcpy from MAXERR (a double) rather than the correct YMM_MAXERR (4 doubles). Thus, 3 uninitialized doubles were copied.

[quote=axn]Here's a shocking new concept: Subfolders![/quote]

I'll leave the mess as is. There are a number of links around the web to these binaries. I suppose I could move all the v24 and earlier binaries to a subfolder. Anyone using those old links is probably better off getting a "not found" error.

[QUOTE=Prime95;321678]As always, programmer error. I did a memcpy from MAXERR (a double) rather than the correct YMM_MAXERR (4 doubles). Thus, 3 uninitialized doubles were copied.
[/QUOTE]

In other words no errors in tests were created in this. It was just the error reporting.

[QUOTE=henryzz;321696]In other words no errors in tests were created in this. It was just the error reporting.[/QUOTE]

Correct

[QUOTE=henryzz;321696]In other words no errors in tests were created in this. It was just the error reporting.[/QUOTE]

3 is the magic number...each time the progress updated it said something like "10 Sumout errors; 3 were not reproducible...."

I have a weird "[I]circadian[/I]"* kernel panic on my older Phenom940.
(It is quite bad because you have to restart the machine, physically: it doesn't crash completely and it freezes the intenet connections, too. See attached screenshot /literally/)

I am not sure if it is mprime or a combination of it with mysterious daemons. Other programs don't panic; month-long sieveing runs, msieve BL, pfgw, newpgen, memtest for a day... I have manipulated these possible interactions by replacing mprime 27.7 with 27.9; replaced OpenSUSE 12.1 with 12.2; ran the comp without X (i.e. init 3). George, do any mentioned in the panic dump functions make sense? Anyone else? TIA!

P.S. Just in case: the workload on which this occurs is a bunch of 831*2^n+1 PRPs on 4 threads (see the top of the snapshot for last line of output)

________
*once a day

Please help me!
When I test my PC on stability prime95 crashes () EVERY time after passes 25k test, it happens with default blend test and with custom test (max fft size 1792, mem used 4096). ALL other stability software (lynx, occt) in that time is stable, except prime95 and test after 25k, but crases after 25k stops when I increase voltage much more. Is it means my voltage is too small?

Win 7 x64, processor 3570k
event view:
[I]Faulting application name: prime95.exe, version: 27.7.1.0, time stamp: 0x4fb2d143
Faulting module name: prime95.exe, version: 27.7.1.0, time stamp: 0x4fb2d143
Exception code: 0xc0000005
Fault offset: 0x000000000014578d
Faulting process id: 0xc0c
Faulting application start time: 0x01cdecd998d17327
Faulting application path: C:\Test\p95v277.win64\prime95.exe
Faulting module path: C:\Test\p95v277.win64\prime95.exe
Report Id: c2391b26-58e2-11e2-b0fe-bc5ff447b886
[/I]

[QUOTE=el15k;323940]When I test my PC on stability prime95 crashes () EVERY time after passes 25k test, it happens with default blend test and with custom test (max fft size 1792, mem used 4096).[/QUOTE]

That is suspicious (failing in the EXACT same spot EVERY time could be a program bug). Does it fail if you run the in-place test? Does it fail with custom tests using different memory allocations?

BTW, what is the FFT size after 25K?

[QUOTE=Prime95;324014]That is suspicious (failing in the EXACT same spot EVERY time could be a program bug). Does it fail if you run the in-place test? Does it fail with custom tests using different memory allocations?

BTW, what is the FFT size after 25K?[/QUOTE]
It fail only after 25k, in-place test passes OK.
It fails after 25k in standart blend mode (I dont remember what FFT size after 25k, simething around ~800) and It fails after 25k in custom test (max fft size 1792, mem used 4096) in this test prime95 crashes after 25k on 480 fft size.

Segmentation fault
 

Hi, mprime 27.9 build 1 Linux64 "crashed" this morning on my computer. It said:
[CODE]Segmentation fault[/CODE]and stopped working. Nothing in prime.log. Invoking dmesg, I got:
[CODE]mprime[35596]: segfault at 7fbb0f2c3938 ip 00007fbc12ad4771 sp 00007fafbbffd010 error 4 in libc-2.12.so[7fbc12a5d000+186000][/CODE]I dont know if you can find out the problem with so less informations. It can't be a hardware bug, since the computer memory has 2bit-ECC and temperatures never exceed 45°C.