RE: Mersenne: Optimising P4 usage
I would say it is defiantly worth it, since it looks like your P4 is not
doing P-1 by doing this first elsewhere you may find a factor during P-1 and
save your self a whole 33M LL test.
Quite simple, stop your clients, add the worktodo line for your next 33M
exponent to your Linux client.
I would use the undoc 'SequentialWorkToDo=0' setting on your Linux box which
causes the client to do all factoring and P-1 testing on LL exponents before
all other work.
Once P-1 has completed on your next exponent stop the client again and move
it back to the P4.
You may want to drop the number of days queued setting on the P4 so it does
not get another exponent in the meantime.
Note : If you have too much work queued to the Linux box already it may
release the exponent rather than processing it when you add it to the
worktodo, to get around this you could backup your existing worktodo and
create a new one temporarily.
This process is similar to one we are implementing for our team to
pre-factor 33M exponents for our team members with P4's which we have
discussed with George.
regards
Rob Reid
www.teamprimerib.com
-Original Message-
From: Gerry Snyder
Sent: 14 January 2002 05:38
To: mer
Subject: Mersenne: Optimizing P4 usage
{]
My question is whether it is worth the trouble to shift the trial and
P-1 factoring of the next one to one of the P3 processors (the non-LL
one). It might lead to one extra LL test in two years.
[]
_
Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Optimizing P4 usage
Hi, At 09:37 PM 1/13/2002 -0800, Gerry Snyder wrote: At home I have 2 PC's, My question is whether it is worth the trouble to shift the trial and P-1 factoring of the next one to one of the P3 processors You only need to do the trial factoring on the P3. P-1 testing on the P4 also uses the SSE2 instructions. As to whether this is worth the trouble, only you can answer that question! -- George _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: slaying cpus with prime95
here are some instances where i have damaged computers by (capriciously?) running prime95! 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. 2 - my manager at work here had a thinkpad. he ran prime95 despite my worry that it was very tough on laptops. within a few months his harddrive failed - possibly due to months of excess heat... :| this could be considered a classic Dilbertian CLM (career limiting move) on my part, but no worry since my manager is super-cool. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... 4 - i bought a 2Ghz P4 recently. despite initial worries that it was running too hot (70 C) because fan was too slow (2800 rpm), i got adventurous and clocked the cpu at 2.1 Ghz for a day. weeks later the machine started acting very badly (motherboard cpu temp alarm caused shutdown @ 90 C even without prime95 running). so i returned it to the vendor. they claimed that my overclocking it broke the P4, and that the top of the cpu was actually burnt/blackened from the heat. this is counter to my belief that improper fan/heatsink was the cause, but i can't prove it. also it runs counter to what i've read here elsewhere about the thermal-protection built into P4s 1.7Ghz or faster. they are returning the P4 to intel to see if Intel will replace it for free, but in the meantime i have to pay for a new cpu! (i'm picking 1.8Ghz this time.) so far my count is 4 for computers i've damaged with the help of the the prime95 application. but i'll keep running it because it is the coolest application around (in a hot way). /eli _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Preventing new assignments
On 12 Jan 2002, at 10:27, Paradox wrote: In about 10 days, I've got 5 Pentium4 computers that will be submitting completed LL tests (for 10,000,000 digit numbers). I used to have dozens of smaller computers working on such LL tests for years, and so I have a collection of Prime95/mprime directories which have 60% to 80% completed LL tests in them. I'm going to want to make sure that the currently running mprime's on the P4s do not get new assignments, so that I can simply remove those copies of mprime and replace it with a copy from my collection. For now, I've set the Days of Work to 1. If I were to set days of work to 0, what would happen? How can I tell it to simply submit the results to primenet, disconnect, and then exit the program when it is done? Well, if that's really what you want to do, why not try the Quit GIMPS option (bottom of Advanced menu)? In v21 (but NOT previous versions!) selecting this option allows you to complete the current assignment then quit (which is what you seem to want), or quit immediately (which you probably don't want to do since PrimeNet will think you have abandoned work before completing the assignments. You can still continue, however, report the results as they come in). If you already have new assignments which you don't want to run, they will automatically expire in about 3 months, or you could return them to the server using Unreserve Exponent in the Advanced menu (v21 only) or through the PrimeNet Manual Testing form. Regards Brian Beesley _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Optimizing P4 usage
On 13 Jan 2002, at 21:37, Gerry Snyder wrote: At home I have 2 PC's, both of which devote their spare cycles to GIMPS. One is a 1.3 GHz P4 running W98, and the other a dual 1 GHz P3 running linux. One of the P3's is doing LL testing, and the other is doing mostly ECM, with a little trial factoring thrown in. The only part of Prime95 that uses the SSE2 P4 instructions is the LL testing. Because of the huge speedup this gives, I would like to keep the P4 machine doing nothing but LL tests. It is now about a month away from finishing its first 10 megadigit candidate. My question is whether it is worth the trouble to shift the trial and P-1 factoring of the next one to one of the P3 processors (the non-LL one). It might lead to one extra LL test in two years. You can do even better than that: Suppose the three processors are as follows: A wants to run trial factoring - if you run LL or P-1 on this, B slows down due to memory bus contention B doesn't mind running P-1 C wants to run LL only. Firstly you want to make sure that C doesn't have SequentialWorkToDo=1 in prime.ini (else TF P-1 will be run before the current LL completes). Also make C's DaysOfWork greater than the time estimated to run TF P-1 on A B. Let C get a new assignment Test=exponent,N1,P and let N2 be the required factoring depth for that exponent. Edit the new assignment in C's worktodo.ini, replacing N1 by N2 and P by 1. If N1 N2, trial factoring is required. Edit A's worktodo.ini file, adding a new line at the top Factor=exponent,N1. Stop restart mprime on A to get the assignment started. If a factor is found, remove the assignment from C's worktodo.ini file let C get a new assignment. Whatever happens, let A report the result of the trial factoring. If A fails to find a factor and P=0, P-1 factoring is required. Edit B's worktodo.ini file, adding a new line at the top Pfactor=exponent,N2,0. Stop start mprime on B to get the assignment started. Let B report the result of the P-1 factoring. If a factor is found, remove the assignment from C's worktodo.ini file let C get a new assignment. Regards Brian Beesley _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
RE: Mersenne: slaying cpus with prime95
I'd wager that in most of those instances, the computer would have died without Prime95's help. Faulty CPU fans in particular... whether they conk out after a short while or eventually, a bad fan is a ticking time bomb. Prime95 may have shortened the life span a bit, but it was doomed. :) Any machine that lacks proper ventilation and allows excess CPU heat to affect the hard drive is buggy design if you ask me. I have a 1GHz IBM laptop that I run Prime95 on 24/7 and while the thing can get hot if I actually place it on my lap, I've had nary a problem except on a couple occasions when I inadvertently blocked the inflow air vent and the thing did overheat. When it did that, it just shut itself down. Things to look for on any computer, Prime95 running or not, are noisy cpu/case fans or fans that don't spin at all. And if you have thermal monitoring, pay attention to it. Hard drives nowadays almost all have SMART on them, so find a program that can read those stats and see if your drive is giving out warnings. Many vendors include software for monitoring your system health (I know HP and Compaq do, and I think IBM does too but I haven't looked). There are 3rd party programs for all that as well. And of course, overclocking is pushing the limits anyway, so I guess we shouldn't be too surprised to find that overclocking may ruin a CPU. I do find it curious that the thermal protection on your P4 didn't kick in, but perhaps your computer's BIOS had some default settings disabling that? Or it had been disabled? I dunno... I don't have a P4 machine (yet), so I can't say. :) In short, I've killed many machines in my life... CPU fans that stopped working and thus frying the processor, or doing silly things like plugging a 486 in the wrong way, etc. And of all the machines I've fried, NONE were running Prime95. Heck, I guess I'd have to say that running Prime95 on a machine has actually brought me better luck with them than without. :) Aaron -Original Message- From: [EMAIL PROTECTED] [mailto:mersenne-invalid- [EMAIL PROTECTED]] On Behalf Of Steve Elias Sent: Monday, January 14, 2002 11:45 AM To: [EMAIL PROTECTED] Subject: Mersenne: slaying cpus with prime95 here are some instances where i have damaged computers by (capriciously?) running prime95! 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. 2 - my manager at work here had a thinkpad. he ran prime95 despite my worry that it was very tough on laptops. within a few months his harddrive failed - possibly due to months of excess heat... :| this could be considered a classic Dilbertian CLM (career limiting move) on my part, but no worry since my manager is super-cool. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... 4 - i bought a 2Ghz P4 recently. despite initial worries that it was running too hot (70 C) because fan was too slow (2800 rpm), i got adventurous and clocked the cpu at 2.1 Ghz for a day. weeks later the machine started acting very badly (motherboard cpu temp alarm caused shutdown @ 90 C even without prime95 running). so i returned it to the vendor. they claimed that my overclocking it broke the P4, and that the top of the cpu was actually burnt/blackened from the heat. this is counter to my belief that improper fan/heatsink was the cause, but i can't prove it. also it runs counter to what i've read here elsewhere about the thermal-protection built into P4s 1.7Ghz or faster. they are returning the P4 to intel to see if Intel will replace it for free, but in the meantime i have to pay for a new cpu! (i'm picking 1.8Ghz this time.) so far my count is 4 for computers i've damaged with the help of the the prime95 application. but i'll keep running it because it is the coolest application around (in a hot way). /eli _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: slaying cpus with prime95
On Mon, 14 Jan 2002, Steve Elias wrote: 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. Laptop fans don't seem to be very durable. I have a two year old Dell that has apparently had a broken fan for a long time, and I didn't notice it because the it (either the MB or the Celeron) throttles down automatically but continues to work and nothing I was doing was that CPU sensitive that I would notice the lack of speed. This is actually kind of neat because it leaves the computer still usable (and even at top speed with an external fan blowing in the intake port), but I do wish it would popup a notice that the CPU was being throttled. 2 - my manager at work here had a thinkpad. he ran prime95 despite my worry that it was very tough on laptops. within a few months his harddrive failed - possibly due to months of excess heat... :| this could be considered a classic Dilbertian CLM (career limiting move) on my part, but no worry since my manager is super-cool. Yes, I do not think most laptops are very durable. Companies know they will usually get a lot less use than a desktop, so they know they only need to get a few months of actual runtime to get them out of the warranty period. I think it has more to do with simply using them 24/7 (or more than a few hours a day on average) than with prime95. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... Well, I don't think that necessarily has anything to do with prime95 either. My dad had a Cyrix 200, his cpu fan started being noisy and flaking out, I advised him to get a new fan, which he did not, and eventually the system bombed hard. He was not running prime95. 4 - i bought a 2Ghz P4 recently. despite initial worries that it was running too hot (70 C) because fan was too slow (2800 rpm), i got adventurous and clocked the cpu at 2.1 Ghz for a day. weeks later the machine started acting very badly (motherboard cpu temp alarm caused shutdown @ 90 C even without prime95 running). so i returned it to the vendor. they claimed that my overclocking it broke the P4, and that the top of the cpu was actually burnt/blackened from the heat. this is counter to my belief that improper fan/heatsink was the cause, but i can't prove it. also it runs counter to what i've read here elsewhere about the thermal-protection built into P4s 1.7Ghz or faster. they are returning the P4 to intel to see if Intel will replace it for free, but in the meantime i have to pay for a new cpu! (i'm picking 1.8Ghz this time.) Thermal protection for P4's is partly a matter of the chip, and partly a matter of the motherboard/chipset. Intel built motherboards have been shown to be very paranoid about throttling the CPU compared to boards built by other vendors (and with non-Intel chipsets). As far as the CPU's own thermal protection, Intel took some heat for the fact that the thermal protection was causing performance problems compared to AMD chips. There was a rather famous test done by Tom's Hardware Guide where they removed the heatsink from test chips while they were running to simulate a catastrophic failure of the heatsink falling off the CPU. The P4 throttled perfectly with a huge performance degradation, but it kept going. So it does seem odd that your chip would have burned, but perhaps Intel backed off on the thermal diodes in the chip itself because of the performance problems. _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: slaying cpus with prime95
On Mon, Jan 14, 2002 at 02:45:21PM -0500, Steve Elias wrote: here are some instances where i have damaged computers by (capriciously?) running prime95! 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. Mmm, yes the fan on my laptop (Dell 38000) sounds like someone dumped a load of grit in it - not good :-( I stopped running prime95 (mprime actually) on it for that reason. -- Nick Craig-Wood [EMAIL PROTECTED] _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: slaying cpus with prime95
As others have already mentioned, those machines would probably have died even without Prime95. The way I have always looked at it, Prime95 generally causes those types of (pre-existing) problems to manifest _before_ the warranty expires, rather than after. This feature is certainly not a Bad Thing :-) Steve Harris -Original Message- From: Steve Elias [EMAIL PROTECTED] Date: Monday, January 14, 2002 1:58 PM Subject: Mersenne: slaying cpus with prime95 here are some instances where i have damaged computers by (capriciously?) running prime95! 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. 2 - my manager at work here had a thinkpad. he ran prime95 despite my worry that it was very tough on laptops. within a few months his harddrive failed - possibly due to months of excess heat... :| this could be considered a classic Dilbertian CLM (career limiting move) on my part, but no worry since my manager is super-cool. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... 4 - i bought a 2Ghz P4 recently. despite initial worries that it was running too hot (70 C) because fan was too slow (2800 rpm), i got adventurous and clocked the cpu at 2.1 Ghz for a day. weeks later the machine started acting very badly (motherboard cpu temp alarm caused shutdown @ 90 C even without prime95 running). so i returned it to the vendor. they claimed that my overclocking it broke the P4, and that the top of the cpu was actually burnt/blackened from the heat. this is counter to my belief that improper fan/heatsink was the cause, but i can't prove it. also it runs counter to what i've read here elsewhere about the thermal-protection built into P4s 1.7Ghz or faster. they are returning the P4 to intel to see if Intel will replace it for free, but in the meantime i have to pay for a new cpu! (i'm picking 1.8Ghz this time.) so far my count is 4 for computers i've damaged with the help of the the prime95 application. but i'll keep running it because it is the coolest application around (in a hot way). /eli _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
RE: Mersenne: slaying cpus with prime95
Crazy thought here, but what if prime95 only ran at x% speed? For example, put some hlts in there and see how that affects the CPU temp. I think for those of us that run Prime95 on laptops find it unnerving when the fans are running full blast all the time, and it would be cool to have a cooler running prime95 rather than no prime95 at all. Thinking about it, you could probably keep the temp down to around normal idle temp... -Jeremy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Nick Craig-Wood Sent: Monday, January 14, 2002 4:26 PM To: [EMAIL PROTECTED]; Steve Elias Subject: Re: Mersenne: slaying cpus with prime95 On Mon, Jan 14, 2002 at 02:45:21PM -0500, Steve Elias wrote: here are some instances where i have damaged computers by (capriciously?) running prime95! 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. Mmm, yes the fan on my laptop (Dell 38000) sounds like someone dumped a load of grit in it - not good :-( I stopped running prime95 (mprime actually) on it for that reason. -- Nick Craig-Wood [EMAIL PROTECTED] _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: prime95 fans slaying cpus
hi folks, thanks for all the responses regarding my Frodo-style cpu-slaying with prime95! Mary, it's great that your dell laptop throttles back the cpu when it overheats. that's smarter than my P4 desktop with Asus motherboard, and also smarter than the toshiba 'satellite' laptop. if you watch/compare the iteration time of prime95 that ought to indicate when the cpu is throttled back... when the toshiba laptop was on its last-leg i put it in the 50 degree garage so it could keep computing. i'm sorely tempted to do that again now with the working fan but i must restrain myself! maybe i need a mantra or theme song to help. sing this to the mickey mouse club theme: FFT, GCD, M-O-U-S-E. or maybe FFT, GCD, G-E-O-R-G! well, almost. maybe this sort of thing a bit too far out there for you prime readers! ok, i'll try to throttle back my juvenile sense of humor a bit... Yes, I do not think most laptops are very durable. Companies know they will usually get a lot less use than a desktop, so they know they only need to get a few months of actual runtime to get them out of the warranty period. yes this point is well taken! I think it has more to do with simply using them 24/7 (or more than a few hours a day on average) than with prime95. i doubt that to some extent - when most/many/some laptops are sitting idle, or doing something simple like a realplayer audio stream, there is no need for their cpu fan to run. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... Well, I don't think that necessarily has anything to do with prime95 either. My dad had a Cyrix 200, his cpu fan started being noisy and flaking out, I advised him to get a new fan, which he did not, and eventually the system bombed hard. He was not running prime95. i hear ya... i think that prime95 hastened the cyrix machine's departure... the amazing part is that the machine ran for nearly a year with no cpu fan, including through many 100 degree days. Thermal protection for P4's is partly a matter of the chip, and partly a matter of the motherboard/chipset. Intel built motherboards have been shown to be very paranoid about throttling the CPU compared to boards built by other vendors (and with non-Intel chipsets). As far as the CPU's own thermal protection, Intel took some heat for the fact that the thermal protection was causing performance problems compared to AMD chips. There was a rather famous test done by Tom's Hardware Guide where they removed the heatsink from test chips while they were running to simulate a catastrophic failure of the heatsink falling off the CPU. The P4 throttled perfectly with a huge performance degradation, but it kept going. So it does seem odd that your chip would have burned, but perhaps Intel backed off on the thermal diodes in the chip itself because of the performance problems. this is all very interesting info - thank you very much! indeed i think thought that an Intel motherboard with the 2Ghz P4 would have done better than the Asus motherboard. i had seen that Tom's h/w guide review of the P4 - that's one reason why i am skeptical when the vendor said that i had burned out my P4 chip. but the data you have provided here has provided some new pieces to the jigsaw puzzle. Date: Mon, 14 Jan 2002 12:20:16 -0800 From: Aaron Blosser [EMAIL PROTECTED] Subject: RE: Mersenne: slaying cpus with prime95 I'd wager that in most of those instances, the computer would have died without Prime95's help. Faulty CPU fans in particular... whether they conk out after a short while or eventually, a bad fan is a ticking time bomb. Prime95 may have shortened the life span a bit, but it was doomed. :) hi Aaron - oh yes! have no doubt that prime95 hastened the computer/cpu's death! but maybe the cpu only quit 3 nanoseconds sooner than it would have without prime95 ;) Any machine that lacks proper ventilation and allows excess CPU heat to affect the hard drive is buggy design if you ask me. from what i've seen, many laptops are this way by design. they are too small to isolate the hard drive from the cpu heat, especially when the cpu fan quits. I have a 1GHz IBM laptop that I run Prime95 on 24/7 and while the thing can get hot if I actually place it on my lap, I've had nary a problem except on a couple occasions when I inadvertently blocked the inflow air vent and the thing did overheat. When it did that, it just shut itself down. it's pretty clear that the IBM laptop fans are more robust than the toshiba satellite-pro... go big blue! Things to look for on any computer, Prime95 running or not, are
Re: Mersenne: prime95 fans slaying cpus
On Mon, 14 Jan 2002, Steve Elias wrote: Mary, it's great that your dell laptop throttles back the cpu when it overheats. that's smarter than my P4 desktop with Asus motherboard, and also smarter than the toshiba 'satellite' laptop. if you watch/compare the iteration time of prime95 that ought to indicate when the cpu is throttled back... Actually, the iteration time does not change appreciably when the cpu throttles back, and I'm not sure why. Sometimes the external fan that I have set up (until the replacement fan parts arrive) goes a bit wobbly which really cuts down on the airflow, and the system starts heating up again. When that happens the iteration time only goes up by .001 sec (i.e. from .120 to .121), but it's obvious from looking at the timestamps that the iterations are taking a lot longer than they should (instead of every 2 minutes, it'll be 3 minutes every 4th timestamp or so). prime95 appears to be counting cycles instead of timing by the system clock, and it takes the same amount of cycles whether the cpu is running at 466 or 350. As far as the Dell goes, I've been pleased with it in spite of the fan going out. It takes a lot of hard use (even abuse one might say) even without prime95. I bought it because I'd heard the Latitudes were supposed to be fairly hardy for laptops (I hear Inspiron machines do not enjoy the same rep, however), and that's been true for me. As a bonus, I can even service it myself (the service manual is available in PDF off the Dell website), so replacing the fan will not cost me labor charges and I won't have to lose use of the machine for more than an hour or so. _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: slaying cpus with prime95
On 14 Jan 2002, at 12:18, Mary Conner wrote: Laptop fans don't seem to be very durable. Fans in general? I've had very little trouble with CPU fans (fortunately - and perhaps I'm lucky) but I find failed PSU fans and noisy (vibrating) case fans to be just about the commonest faults on desktop systems. I have a lot less to do with laptop systems, but so far (touch wood) I've no experience at all of laptop fan failure. I have a two year old Dell that has apparently had a broken fan for a long time, and I didn't notice it because the it (either the MB or the Celeron) throttles down automatically but continues to work and nothing I was doing was that CPU sensitive that I would notice the lack of speed. This is actually kind of neat because it leaves the computer still usable (and even at top speed with an external fan blowing in the intake port), but I do wish it would popup a notice that the CPU was being throttled. My Tosh 4070 makes it quite clear that the fan is running - it's LOUD, several decibels more so than the average desktop. Small diameter fan running at high speed = fast air flow = aerodynamic noise. Many ancient laptops had a two-colour LED which shone green for full CPU speed, red for reduced CPU speed (though you had to select the CPU speed manually, using software or by a special key combination). The extra manufacturing cost of fitting this to current systems would be very small, though the introduction of technologies like Speed Step would complicate matters to some extent. Possibly a tricolour system would be needed. C'mon, Toshiba et al... Anyway I guess that's another reason for running Prime95 - you should notice a performance loss if your processor drops into a throttled mode for some mysterious reason! Yes, I do not think most laptops are very durable. Companies know they will usually get a lot less use than a desktop, so they know they only need to get a few months of actual runtime to get them out of the warranty period. I think it has more to do with simply using them 24/7 (or more than a few hours a day on average) than with prime95. Intermittent operation costs reliability - if parts are constantly being warmed cooled they _will_ age. Also the average laptop gets dropped, rained on, moved from warm to cold environments, etc. etc., far more than the average desktop system does. Thermal protection for P4's is partly a matter of the chip, and partly a matter of the motherboard/chipset. Intel built motherboards have been shown to be very paranoid about throttling the CPU compared to boards built by other vendors (and with non-Intel chipsets). Intel's BIOS doesn't even give you the choice of overriding the factory default throttling mode / thresholds :( As far as the CPU's own thermal protection, Intel took some heat for the fact that the thermal protection was causing performance problems compared to AMD chips. This should not be a problem - provided it is actually possible to get the heat generated out of the chip without tripping thermal protection, with a properly engineered and assembled heatsink / fan / case ventilation system. There was a rather famous test done by Tom's Hardware Guide where they removed the heatsink from test chips while they were running to simulate a catastrophic failure of the heatsink falling off the CPU. Ever heard of a heatsink _accidentally_ falling off a CPU on a running system? Sounds about as probable as an airline passenger being killed in flight by a meteorite strike which doesn't bring down the whole aircraft. I thought the point of the THG test was to educate self-builders using Athlon/Duron not to test a system by applying power - even for a few seconds - if the CPU heatsink wasn't fitted. Anyway, the latest AMD processors have thermal protection against this sort of abuse. The P4 throttled perfectly with a huge performance degradation, but it kept going. So it does seem odd that your chip would have burned, but perhaps Intel backed off on the thermal diodes in the chip itself because of the performance problems. I doubt it. If they were really worried they'd just have disconnected the sensors. There seems absolutely no point at all in fiddling around with thermal protection if you leave yourself with a product which is subject to both thermal performance degradation and rapid failure due to moderate overheating. Regards Brian Beesley _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: slaying cpus with prime95
On 14 Jan 2002, at 14:45, Steve Elias wrote: 1 - i just got my wife's toshiba laptop back from toshiba warranty service. running prime95 for ~6 months on it caused the fan to die, and then the laptop would overheat shutdown even without prime95 running. apparently the heat caused lots of disk badblocks too. I've had a Tosh Sat 4070 laptop running mprime 24x7 for about 3 years. The fan runs - constantly. A fan certainly shouldn't die within one year, though some will undoubtedly do so, due to manufacturing faults. As for bad blocks on disk - I wouldn't trust a diagnosis made on a system with a cooked processor. Even if the disk has developed read problems on some blocks written when its temperature was way too high, reformatting (or block repairing, using a suitable utility) the disk when it's running properly cooled should repair them. 2 - my manager at work here had a thinkpad. he ran prime95 despite my worry that it was very tough on laptops. within a few months his harddrive failed - possibly due to months of excess heat... :| this could be considered a classic Dilbertian CLM (career limiting move) on my part, but no worry since my manager is super-cool. Blame power management. HDDs constantly starting stopping are going to age rapidly. Prime95 helps accelerate this process by writing save files every so often. Changing the disk write time to 1 hour or so seems reasonable on a laptop since the main reason you write save files is to protect against power loss; all laptops effectively contain a UPS, and should suspend without data loss even if the battery runs flat. I'd reccomend: (a) configuring power management to run the hard disk constantly when the system is on mains power, to extend the disk life; (b) configuring Prime95 to suspend itself when the system is running on battery power. This reduces HD start/stop activity and extends the usable run time by allowing the floating-point unit to shut down, reducing both power consumption and the cooling requirements considerably. 3 - i also ran the prime95 app for a year or so on an ancient cyrix p120+ which had a cpu-fan that stopped. after a couple months of no-cpu-fan, that cpu died completely... No surprise at all. Undercooled processors will die sooner rather than later, whatever you run on them. Meanwhile they're probably not running reliably, so whatever data you are processing aren't reliable. 4 - i bought a 2Ghz P4 recently. despite initial worries that it was running too hot (70 C) because fan was too slow (2800 rpm), i got adventurous and clocked the cpu at 2.1 Ghz for a day. weeks later the machine started acting very badly (motherboard cpu temp alarm caused shutdown @ 90 C even without prime95 running). so i returned it to the vendor. they claimed that my overclocking it broke the P4, and that the top of the cpu was actually burnt/blackened from the heat. this is counter to my belief that improper fan/heatsink was the cause, but i can't prove it. also it runs counter to what i've read here elsewhere about the thermal-protection built into P4s 1.7Ghz or faster. they are returning the P4 to intel to see if Intel will replace it for free, but in the meantime i have to pay for a new cpu! (i'm picking 1.8Ghz this time.) Again it's no surprise that, if you overclock a CPU and fail to uprate the cooling, the processor will run hot. The existing 0.18 micron 2.0 GHz P4 is pretty close to the limit of what can be done in the way of transporting heat out of the die (that's why Intel are shifting to 0.13 micron technology). Experience in the past has always been that the low-end chips in any particular family stand overclocking better than the high-end chips, though I'd be very surprised to find a x GHz rated chip overclocked to 1.2x GHz performing more stably than a 1.2x GHz rated chip of the same type. Intel's warranty seems to state very clearly that the warranty only applies to chips operated within the stated limits, so failure to honour a warranty claim on an overclocked CPU would be quite reasonable. Intel retail box processors use a larger than normal fan, so a slower rotational speed is reasonable will shift the same amount of air. The reason this is done is that aerodynamic fan noise depends on a large power of the air speed, so moving a larger slug of air more slowly is quieter than moving a smaller slug of air more quickly. 70C sounds too hot to me. But I don't know the exact way in which the sensor is incorporated into the P4 die. There should be a document on the Intel site showing the thermal constraints for each variant of the P4. Motherboard at 90C is undoubtedly fatal! Was this really due to a processor cooling problem, or is there a runaway voltage regulator? (You didn't overload the voltage regulator during your overclocking experiments...?) BTW it's fairly common for thermal tape or cheap thermal paste to crack and/or darken
Re: Mersenne: Optimizing P4 usage
- Original Message - From: Gerry Snyder [EMAIL PROTECTED] To: mer [EMAIL PROTECTED] Sent: Monday, January 14, 2002 5:37 AM Subject: Mersenne: Optimizing P4 usage At home I have 2 PC's, both of which devote their spare cycles to GIMPS. One is a 1.3 GHz P4 running W98, and the other a dual 1 GHz P3 running linux. One of the P3's is doing LL testing, and the other is doing mostly ECM, with a little trial factoring thrown in. If I'm not mistaken, you'd be better to have one of your dual processors doing just TF. [...] My question is whether it is worth the trouble to shift the trial and P-1 factoring of the next one to one of the P3 processors (the non-LL one). It might lead to one extra LL test in two years. Definitely move the TF to your P3. A P4 doing integer work is wasted. The worktodo file for the P4 has: Test=current,68,0 Why hasn't your current assignment had a P-1? You should suspend testing to do this. My system with 459MB available memory would choose B1=375000, B2=8156250 (5.17% chance of success) for a 33M first-time assignment (i.e a successful P-1 saves two LL tests) and B1=18, B2=297 (3.78%) for a 33M doublecheck assignment. Your system - with a different amount of available memory - might choose different limits, but whatever they are, you should choose limits somewhere between them, closer to the first-time, or doublecheck values according to how far through the first time check you are. Test=next,60,0 It should be noted that the W95 machine does not have enough RAM for phase 2 of the P-1 factoring, but the linux machine does. Do P-1 on whichever machine can give it the most memory, and always give it as much as you can. 340MB or thereabouts is optimal for a 6-7M exponant. GOK how much a 33M exponant would take. Any suggestions? If you've got less than 512MB on your linux box, email the exponant(s) to me, (after you've TFed it) and I'll P-1 it for you. I can do a 6-7M P-1 in about 3 hours, so I'd guess it'd take a couple of days. TIA, Gerry Daran G. _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
