Know your 100% load temp? TRY THIS - S&M 1.3

[IvanAndreevich]

NEW VERSION! 1.50 now. Download here - 200 KB (http://testmem.nm.ru/snm.zip)

1) Processor test - leave ONLY CPU (FPU) checkbox checked.
2) Settings - make processor check "long" by dragging the slider left.
3) Start Testing! [60 min test]

On my system, this prog heats up the CPU 5 degrees above prime running small FFTs. Just thought many of ya could use this for stability testing. Post if you've never seen a temp this high on your system :toast:

[mastaqz]

i've seen 64° oO, with prime, i got 61°!

[macci]

Better (=hotter) than cpuburn (http://pages.sbcglobal.net/redelm/)?

[Torin]

Need for Speed Underground 2 got my CPU running hotter than prime ever did. :-/

[IvanAndreevich]

Torin
Did you try this? ;)

macci
I think so. Give it a try - can't hurt IMO :)

[mad mikee]

Danm no 'haaa hooo hootttt' smiley hehe, so :flame: and hopefully not :explode:

[stevehat1]

Tried S&M.....an appropriate name for what happens to the poor cpu :lol:
S754 3400+ Claw aircooled @2.6Ghz, 1.81vcpu, idle @ 37C, max in the 15 minutes of S&M was 47C, so 10C max rise in temp. average probably 9C.
Got to do a little off this :toast: while running ;)

[IvanAndreevich]

mad mikee
How many degrees did the temp rise from previous top temp?

stevehat1
Only 10 degrees up from IDLE? Did you do FPU only?? Cause my buddy with a 2500+ Mobile got his chip running 10 degrees ABOVE Prime95 small FFTs :)

[stevehat1]

mad mikee
How many degrees did the temp rise from previous top temp?

stevehat1
Only 10 degrees up from IDLE? Did you do FPU only?? Cause my buddy with a 2500+ Mobile got his chip running 10 degrees ABOVE Prime95 small FFTs :)
FPU only. This test seems to raise CPU temps quite well, I don't think my cpu will pass 15 min prime at the same clocks, but I don't use prime for checking stability. When this thing isn't benching it running F@H. I find what F@H runs at then back down a little and run it there 24/7

[twiggy]

Wow it heated my stuff up pretty good.

EDIT: It finds memory errors quicker then memtest too lol I dont think I like this proggy!;)

[IvanAndreevich]

twiggy
Yes, this is the kind of result I get - 5 degrees higher than prime. Memory test IS pickier than Prime @ large FFTs too :slobber:

[Shane5578]

it sure does burn my vapo'd cpu up fast.

is there an option where i can loop it?

[mad mikee]

Read title hehe.

Sures helps set paste faster, ran it a few times and lower every time. Stuck @ about 52-53 max. Priming 1024-4096 to get the memory chugging, even w/ blend only up to 48c now :D

Memory test sends me back to the minors w/in 3 minutes, so back to priming :mad:

[IvanAndreevich]

Shane5578
I believe the loop isn't included in the current release. I am willing to be it will be in something that is more final :)

mad mikee
Nice rise from Prime small FFTs :)

[STEvil]

bump

Just gave this a go, sure gets mine a lot hotter than anything else i've tried yet!

[IvanAndreevich]

STEvil
4-5 C?

[kops]

not bad, a gd 5-8° increased over prime.

[Cornelious0_0]

So on the settings screen, from left to right,

Processor check: long, everything else unchecked
Tests: CPU (FPU) is only one checked
Memory checking: long, everything else unchecked

Should I even change anything with the memory checking? I want to start using this app as my stability tester, and I just wanna know how exactly a few of these things should be set to ensure I'm getting fairly dependable results out of this program.

[STEvil]

yeah, 4-5c above prime.

[stealthbomber]

Ahh :banana: :banana: :banana: :banana: this!!!

My Athlon @2.6GHz @1.84v Primes for months non stop @~46C max, with this program it stops with the "failed" error within 3 mins.

I also notice my 12v rails jumps above 13v while running this, must replace my POS 550w Q-Tec ASAP! :eek:

[halcyon]

Comparison against CPUburn k7 and others:

http://testmem.nm.ru/snm.htm

Latest version will also be at that page (currently same as the original linked at the top of this thread).

[Cornelious0_0]

Comparison against CPUburn k7 and others:

http://testmem.nm.ru/snm.htm

Latest version will also be at that page (currently same as the original linked at the top of this thread).

Hehe, too bad I can't read it.....and my translator doesn't like it either. :(

[stealthbomber]

Seems my problem is purely heat related, anything at/above 50C and the test would cap out immediately (by decreasing the tornado fan speed), it passed the 1 hour FPU test @2.5GHz with 1.76v as the temp only tops out at 44C.

Gonna have to add a custom liquid cooling setup to the list... :(

[IvanAndreevich]

stealthbomber
Yes, my friend had the same problem on a aircooled XP-M 2500+. It's stable in Prime, but this prog just made it 8C hotter so his computer would crash.

[gkiing]

Burnk7 (for socket A/462) is still the best imho, it's way hotter than prime95 and probably this too. I'll do a comparison sometime.

[IvanAndreevich]

gkiing
This will pwn it. Burn will know his daddy :p There is a link above that compares this prog to all the other "burn-in / stability" progs out there.

[gkiing]

I highly doubt it. I'll run some tests at 2.2v :) I'm 100% sure it will maul this other program.

Time to fry!

[IvanAndreevich]

gkiing
Don't doubt it.. try it :)

[Shane5578]

hmm s&m reallly messes up my multitasking progs.

for example, i watch tv+prime95, to keep my vapo warm.

but when i use s&m, i get no signal from my tv card. it hogs too much cpu power. lol

[sky]

just tested it. compared to prime i got about a 1°C difference, which is strange, but my mosfet sure took a beating. they went up to new heights. ~39°C, where before they did 34-35°C in prime or any other situation. anyway my memory craps out and throws in the towel before i could do a full run, but whoa... hot stuff :)

[IvanAndreevich]

sky
How come your mosfets are running so cool in the 1st place?

[gkiing]

Thought I should add, the maximum amount of heat a cpu can put out is always 88% of its theoretical maximum.

[IvanAndreevich]

gkiing
A general statement like that can never be true, sorry ;)

[STEvil]

Well, he's right, but 88% is too general.

[IvanAndreevich]

gkiing
And how do you think is the theoretical maximum calculated, btw :) Intel shows it's TDP as 75% of the actual max.

[gkiing]

http://www.benchtest.com/calc.html

It may be different for different chips.

[DaGoof]

Burned me up

48*C @ 1.95v

[sky]

sky
How come your mosfets are running so cool in the 1st place?

drilled / cut up a former cpu-heatsink, put some tim (actually pads) on all the spots where it touches a chip and then to top it off put a 80mm fan above it all
i used to think the mosfet sensor displays the coretemp and vice versa, but no way, the temps change so fast on the cpu sensor... from idle (with disconnect enabled) to full load...2 seconds max - and you can see the temp shooting up...

first time i did use that custom mosfet-sink, i didn't pack a fan... didn't check the temps for a while.. all of a sudden 65°C and rising.. whoops... so i put a fan above and all was well...

[IvanAndreevich]

sky
Well, 65C is good. They go close or even more than 100C without sinks.

ALL
BTW, if you wanna screw your MOSFETs try the "Power Supply" mode ;) If I run that, CPU voltage becomes MORE unstable.

[IvanAndreevich]

More xposure

[docah]

S&m got my cpu 2c higher than k7burn. 4c over prime95.
I typically run p95 all day, i joined gimps. So kicked on S&m while the cpu was already hot, same for k7burn. YMMV.

[Cornelious0_0]

This is all very interesting.....because running S&M just how it has been described does not increase my temps anymore then running two copies of Prime95's torture test for a few hours does.....

[jjcom]

this tihng made my thought stable computer freeze, it reported no errors...just computer freezed when I was typing "So far my compuer seems sta-" lol I saw a new high for my PWM temp...it was something like 62degrees...CPU went up to 'round 54 before it stopped so...this board seems to read 2degrees lower than my old ASUS.

jjcom

[IvanAndreevich]

Seems like this software brings out all the temperature-related issues ;)

[jjcom]

yeah and this brings up another issue as well. If this is the only software that stresses your computer this much...then whats the point? Its like this mem test thing someone did...stress one section of the memory more than anything ever will. Prime95 always brings out the highest temps out of all the rest of the things I do...then this program shows up. I mean this might be nice, but then again if this is the only program that stresses the processor this much...

jjcom

[Cornelious0_0]

yeah and this brings up another issue as well. If this is the only software that stresses your computer this much...then whats the point? Its like this mem test thing someone did...stress one section of the memory more than anything ever will. Prime95 always brings out the highest temps out of all the rest of the things I do...then this program shows up. I mean this might be nice, but then again if this is the only program that stresses the processor this much...

jjcom

I understand where you're coming from....but it's moreso about creating or using a higher standard to "judge" our systems by so that we KNOW everything else will be fine.

[***Deimos***]

yeah and this brings up another issue as well. If this is the only software that stresses your computer this much...then whats the point? Its like this mem test thing someone did...stress one section of the memory more than anything ever will. Prime95 always brings out the highest temps out of all the rest of the things I do...then this program shows up. I mean this might be nice, but then again if this is the only program that stresses the processor this much...

jjcom

now now... no need to go crazy and buy a G5 and cheer for Microsoft....

Now, whenever you need to know if someone's overclock is stable, you can ask if it is S&M stable...
afterall, if all you do all day is surf the web, and you have major overclock, then its all good and bliss, until you decide to do anything serious. You know its all fun and games until you overclock too much and have stability problems.

[jjcom]

good point, you know this could bring 'bout a nice discussion on how XS users feel about stablity...

jjcom

[Cornelious0_0]

good point, you know this could bring 'bout a nice discussion on how XS users feel about stablity...

jjcom

Yes, but the fact that this program is still kinda the "new kid on the block" isn't going to help it out in the "discussion".....cus everyone lives and dies by Prime. ;)

[jinu117]

Well I have yet to have machine that is prime 95 stable and fail on something else (granted my PSU was up to snuff at the time).
Prime 95 still will hold me for my stability test for 24/7. But I found a wonderful use for this S&M. In fact when we were live charging the Mach II, we used this. Gets to critical point faster than anything I know and will be perfect way to get good charge on a system where load capacity is so high it would otherwise take hours to find the temperature settling with baker evap :) FYI, this brings up my evap temp for another 7-10c vs prime 95 when running 1.95v about 3ghz. Good stuff indeed.

[Cornelious0_0]

Well I have yet to have machine that is prime 95 stable and fail on something else (granted my PSU was up to snuff at the time).
Prime 95 still will hold me for my stability test for 24/7. But I found a wonderful use for this S&M. In fact when we were live charging the Mach II, we used this. Gets to critical point faster than anything I know and will be perfect way to get good charge on a system where load capacity is so high it would otherwise take hours to find the temperature settling with baker evap :) FYI, this brings up my evap temp for another 7-10c vs prime 95 when running 1.95v about 3ghz. Good stuff indeed.

True.....good point.

[comment]

Running M&S for an hour gives me the usual 2-3°C temperature rise on the evap.

[v3n]

On a completly stable pc this prog hangs it in about 20 seconds on the cpu test

temp reading was about 38 degrees and ive run stable @ 55c during the summer playing games

12 hours of 2x prime stable and it runs desert combat for 6 hours which adds heat from the passive gpu cooler

is lockup a k7 thing?

[jjcom]

you know that could be...my temps were starting to get a little high...but I thought I could go for another few degrees before problems would start...the thing that really made me wonder was my PWM temps...1.8v Vcore Prime95 would take it to about 57degrees or so...this program took them to 62degrees in around a min. I guess we'll just have to wait until they release a newer version. Altho, my 2.35ghz overclock works fine...just no the 2.4ghz

jjcom

[Fried_Spam]

Tried this last night and it does get warm. SuperPi 32m run gets my processor to about 40-41 with PWM at 34, Prime95 large FFTs gets about 46-47 and about 36 on PWM, this nearly caused a meltdown at 51 on the processor and 41 on the PWM. One question though, it didn't kick out any errors but there wasn't anywhere it seemed that it would show an error. If it generated an error would it stop working like Prime and SuperPI?

[STEvil]

if it errors the window flashes a red border and it stops.

[Cornelious0_0]

if it errors the window flashes a red border and it stops.

Alright, then mine never has....but I still have yet to see S&M get my cpu temps higher then Prime95.

[IvanAndreevich]

Cornelious0_0
Is that on a P4 with HT? Maybe try runnning 2 instances.

[Cornelious0_0]

Cornelious0_0
Is that on a P4 with HT? Maybe try runnning 2 instances.

No need, whether I run two instances of Prime95 to do a FULL 100% usage......or one instance of S&M (which alone does a full 100%) my temps are basically the same.

[IvanAndreevich]

So you tried running 2xS&M vs 1xS&M and it didn't affect your temps?

[LorD_Kvadd]

this is something I have to try!

never liked prime, wonder how my barton @ 2,65ghz likes it on aircooling :D burn baby burn

[Cornelious0_0]

this is something I have to try!

never liked prime, wonder how my barton @ 2,65ghz likes it on aircooling :D burn baby burn

It's nice because it seems that using S&M to call your overclocks stable would place you on a "higher level of stability" for your day to day activities.....and I like it because it's just a set, 1 hour test.....and is a little more efficient.

[IvanAndreevich]

Cornelious0_0
So you didn't answer did you run 2 of them and got the exact same temp as 1?

[NapalmV5]

Damn, what the heck is it doing to the cpu? Took a Palomino @ 1.67Ghz to 75°C in a few minutes.

[_Eduard_]

ROFL I'm running this on a Dell @ school and it is burning up... I can hear the fan going way crazy muhahahaha...

edit: it's a 2.8B

[leslens]

damn good

+5 degrees over prime

[Cornelious0_0]

Cornelious0_0
So you didn't answer did you run 2 of them and got the exact same temp as 1?

Weird, I made a seperate directory for a second copy of the app like you do with Prime, and it won't let me run two of them. Am I missing something?

[IvanAndreevich]

Cornelious0_0
I am not sure. Maybe in the manual it says it supports >1 CPUs already. I just never read manuals, as you may imagine.

Just weird - everybody getting higher temps and you aren't :confused:

_Eduard_
Let the Dell burn in HELL.

[Cornelious0_0]

Cornelious0_0
I am not sure. Maybe in the manual it says it supports >1 CPUs already. I just never read manuals, as you may imagine.

Just weird - everybody getting higher temps and you aren't :confused:

_Eduard_
Let the Dell burn in HELL.

Yeah, can't quite figure it out myself.

For now though I think I'll continue to use S&M to "dictate" my 24/7 overclock that I'll be using....as I dont game anymore, and if the comp can make it through an hour of S&M it'll be good for anything else that I'm gonna be doing with 'er.

[jjcom]

run the program, then move it to the other "processor" then start up another copy, I believe that will work

jjcom

[Cornelious0_0]

run the program, then move it to the other "processor" then start up another copy, I believe that will work

jjcom

Sry....no go.

[aCidbAbY]

:::subscribed:::

[Lith1um]

decided to run the default processor test and get something to eat.

round 5 minutes into S&M my system shuts off. no luck even posting.

take everything out of the case and start swapping parts around.

Neo 2 dead as a door nail.
fx55, xtpe, gb gskill le tccd survived so Im happy.

funny thing is my neo2 always sucked big time. when it failed the mosfet area was roasting. When I was looking around to see if it had a smoked cap I actualy burned my finger touching one of the choke coils.

Im kinda glad it died. Now I can move on to another board, pitty there is no "great" 939 agp board yet. I need a good non neo2 board to tide me over till dfi 939 agp.

[stealthbomber]

Ouch, doing the ~1hr FPU test right now on my Neo2 plat, better not go kaput!

[stealthbomber]

Damn system shutdown as well, except it's not dead, yet... :stick:

[jjcom]

Wonder if NEO2's won't do well with this program...if problems like this are happening...good luck on another S939 board

jjcom

[stealthbomber]

OK into my 2nd attemp on the same board, this time at stock 1.4v vcore @2.4GHz, temps reading ~60C ATM even with a 92mm YS-Tech on a XP90 HS...

[STEvil]

Sounds like this program could double as a review tool to stresstest motherboards and see which ones will actually stand up to some abuse :D


decided to run the default processor test and get something to eat.

round 5 minutes into S&M my system shuts off. no luck even posting.

take everything out of the case and start swapping parts around.

Neo 2 dead as a door nail.
fx55, xtpe, gb gskill le tccd survived so Im happy.

funny thing is my neo2 always sucked big time. when it failed the mosfet area was roasting. When I was looking around to see if it had a smoked cap I actualy burned my finger touching one of the choke coils.

Im kinda glad it died. Now I can move on to another board, pitty there is no "great" 939 agp board yet. I need a good non neo2 board to tide me over till dfi 939 agp.

[Lith1um]

Just read this pcpc vs. ocz psu thread.

http://www.xtremesystems.org/forums/showpost.php?p=664466&postcount=95

I was powering my neo2 with an ocz 600.

now Im starting to wonder.....although the ocz is doing fine.

not going to get another neo 2, thats for sure.

edit:

Im kinda a bonehead. I have an infra red thermometer sitting here which I brought home from work. When the neo2 failed I shoulda scanned the board for hot spots and observed the vrm area temps. Never crossed my mind, I used my finger..... rof.

edit:

considering the emi issue in the linked thread, Im starting to wonder if that is why my neo2/fx55/1gb gskill le tccd combo never overclocked very well. Maybe it has someting to do with prime95's biatchyness with my combo. An emi firestorm problably wouldnt help things.

Just speculation though.

[stealthbomber]

It completed the 1 hour long FPU test on my rig, using a Tagan PSU here, tempted to swap out the YS-Tech for a 92mm Tornado @5v :stick:

[Cornelious0_0]

It completed the 1 hour long FPU test on my rig, using a Tagan PSU here, tempted to swap out the YS-Tech for a 92mm Tornado @5v :stick:

Haha, go for it.......but I'd hate to be that PSU though.....nice 'n chilly.....but bloody loud. :p

[IvanAndreevich]

The mosfets on my Neo2 do get bloody hot under substancial load. Even sinked, the sinks burn my fingers immediately if there is no air going over them.

[shadowing]

This prog. made my 64's temp all the way up to 72C. On FAH, it ran at only 60C. This is shocking...

[IvanAndreevich]

Try it :p

[Dunk]

Been running it here for the past 20mins and so far it's reached an all time high of 57C :eek:

Two instances of prime and F@H and/or superpi 32M only get it to 49-50C tops. It usually hovers between 47-49 @ full load and one or two games take it up to 51. I think I'll keep an eye on it after reading the reports there. New load temp in the mid fifties instead of the high 40's :stick:

[Cornelious0_0]

Been running it here for the past 20mins and so far it's reached an all time high of 57C :eek:

Two instances of prime and F@H and/or superpi 32M only get it to 49-50C tops. It usually hovers between 47-49 @ full load and one or two games take it up to 51. I think I'll keep an eye on it after reading the reports there. New load temp in the mid fifties instead of the high 40's :stick:

I honestly don't know what S&M could be doing differently, or why I have yet to notice a temp difference, but it's nice to see a mix of different results none-the-less.

[IvanAndreevich]

Dunk
What CPU is that?

[STEvil]

Just ran it on XP-M 2500+ @ 11.5x227.4 2.05v

56c max load

EDIT

Abit nF7, SLK-900U with 50cfm Sunon 80mm fan. 20c ambient.

[Dunk]

Winchester 3500+ on an MSI neo 2 plat xp-90 and TT ufo fan with stupid looking christmas tree decorations removed :)
2x512 pqi turbo
g.waird 6800gt gs
audigy 2 zs
antec trueblue 480w
other unimportant crap

I decided to run it overnight while I was sleeping despite hearing other peoples experiences with their neo 2's. Ran two instances of prime with small fft's and in-place large fft's with S&M running in the background set to run it's tests for as long as possible. 9.5hrs later and everything is still working, nothing has died but I've seen a new record all time high for my load temps that has me slightly worried considering it's approaching the 60C (MBM recorded max was 57C). I just hope it doesn't get there or past that during the Spring/Summer months or I might go back down to 2.6Ghz. :eek:

OC Gods, if you're listening, I could really do with that rain stopping so I can get into town and pick up a new psu/headphones :stick:
It'll be sunny again tomorrow when I'm back to work, just you wait and see..

This one's a keeper anyway, nice temps! Back to the point somebody mentioned about temps only getting so high and that only this program made temps go so high: I thought the same thing about prime95 until I started running seti/F@H which gave me the same temps as prime95. You can bet your bunions some game or something else will put as much strain on your cpu as this did and it's only a matter of time before it happens or you come across something that does. When it happens or if it can happen, I'd like to know that my system can handle it, wouldn't you?

[don_vercetti]

interesting name....i'm running it now, see how it likes my set up. 2 mins in and no errors yet!

[i found nemo]

din't do anything diff for my cpu temp stayed at 45c and the burn in on sandra gets higher than that like 48c ....i think you guys need to re-do your heatsink compound

[Cornelious0_0]

din't do anything diff for my cpu temp stayed at 45c and the burn in on sandra gets higher than that like 48c ....i think you guys need to re-do your heatsink compound

I highly doubt THAT many people have their coolers installed incorrectly....but if you're running that system in your sig then maybe this makes sense. Maybe for some reason it doesnt effect P3's the same way it does newer chips....I have no idea...still trying to figure out what's up with MY results. :p:

[don_vercetti]

well i reacher about 4 mins and firefox started crashing, then finally i got a BSOD. Time to try upping some voltages i guess, only reached 47c though.

[IvanAndreevich]

don_vercetti
Was that prime-stable before?!

[don_vercetti]

yeah for about 10 hours (longest i tried). It's perfectly stable with everything else, far cry, 3d mark, prime, F@H. I've got plenty of temperature overhead though, so i'll raise the vcore which is only currently at 1.7v.

[Dunk]

din't do anything diff for my cpu temp stayed at 45c and the burn in on sandra gets higher than that like 48c ....i think you guys need to re-do your heatsink compound

Not the case..

[IvanAndreevich]

i found nemo
Did you even run the correct test in S&M? Doesn't seem like you selected "FPU"

[jjcom]

he won't answer, he was banned for flaming near as I can tell

jjcom

[SPL15]

-12C highest measured with Calibrated and certified production DMM and temp probe.


Normally I get -15 with farcry on multiplayer.

[IvanAndreevich]

OMG man.. 4 74GB raptors in raid. I want just 1 :) What do you get under prime?

PS everybody - seems like this prog supports HT.

[IvanAndreevich]

Try it :p

[IvanAndreevich]

NEW VERSION AVAILABLE :banana:

Voltage monitoring & other stuff. Through speedfan AFAIR right now.

[SPL15]

OMG man.. 4 74GB raptors in raid. I want just 1 :) What do you get under prime?

PS everybody - seems like this prog supports HT.


Talking about the benchmarks? or the "Blend" torture test that uses ungodly amounts of RAM?

Here's a screen shot of 282 stable with most of the main stability benches. That OCCT Program is a great little program that tests your system for 1/2 hour and graphs out all the voltages and temps of your system during the 1/2 Hour. Instead of crashing and hanging up your computer if its unstable, it will stop the test instead.

[GenTarkin]

Sweet little program although one question, is there any way to hack it to run longer than 60minutes? so I can run it overnight ....

[i found nemo]

i found nemo
Did you even run the correct test in S&M? Doesn't seem like you selected "FPU"

you're right i got 51c i din't click fpu.....oh well it's a d*mn good tes

[Walrusbonzo]

Just giving it a little run here too.

Got a heat spreaderless A64 3000+(Clawhammer with 512k l2 disabled) running at 2.5GHz with 1.8v, watercooled. Never seen more than 42C before with Prime 95 or any other software for that matter.

S&M has got me upto 45C after 10 minutes, but things are still stable :)

[kiwi]

~48C but one side of my case is always open otherwise it reaches 60C (newest beta bios)

[IvanAndreevich]

Yeah, I think there needs to be an unlimited test too. Should I contact the author and ask for that option?

[STEvil]

give it a try, would be nice.

[The Byter]

I checked the power consumption of my PSU under full load:

CPU Burn-in v1.017......................: 144W
Prime95 In-place large FFTs v23.7.1: 153W
OCCT v0.91:...............................: 153W
S&M FPU (v1.3):..........................: 160W

[saaya]

nice!

giving it a try now :)

[IvanAndreevich]

Damn it, can't find the author's email.

[STEvil]

I checked the power consumption of my PSU under full load:

CPU Burn-in v1.017......................: 144W
Prime95 In-place large FFTs v23.7.1: 153W
OCCT v0.91:...............................: 153W
S&M FPU (v1.3):..........................: 160W

Can you run those again with ACPI Multiprocessor?

[Liquid3D]

I'm sure you've all seen this however, I find it to be useful tool to determine how these utilities are stressig the CPU. It's how I discovered PRIME didn't tax te processor at 100% LOAD, SETI does but fluctuates, I run SETI and PRIME concurrently to get a full-load, and some serious temps;
Panopsys Throttle Watch (http://www.panopsys.com/throttlewatch.php) and you can download the Zip file here. (http://www.panopsys.com/ThrottleWatch.zip)

When you run the program open "View" and then click "Frequency Graph." Originally it only detected thermal throttling on P4, P-M and Xeon's, now it's A64 compatable. Obviosly temp is the ultimate test, this is just another tool, which shows TM1 or TM2 in real-time or logged.
SYSTEM:
P4 550 at 3.4GHz 1.373Vcore
H20 cooled Alphacool (http://www.alphacool.de/perl/news.pl?s=078c5cf506491702b0de&l=2) NexXxos BOLD for 775
Black Ice Xtreme II
Papst 2x120mm/55CFM
Alphacool 1500 (OASE adjustale pump from 900LPH ~ 1500LPH)
Cooplex 25 Res/6mm lines
Abit AA8
Corsair 4300C3PRO 1GB
ATI X800 XT
OCZ Powerstream 420

[IvanAndreevich]

I don't see the 1st two screens.

[The Byter]

Can you run those again with ACPI Multiprocessor?

I don't have Multiprocessor :\

[IvanAndreevich]

Search google for "ricky tweak". It's just installing a multiprocessor driver on a singleprocessor system. You don't need dual CPUs to use that.

[STEvil]

update processor driver in device manager list then reboot twice.

[Liquid3D]

Here's all my screenshots comparing S&M to SETI. First running at 200FSB (default speed and Vcore) then showing temps IDLE/overclocked at 250FSB (4.2GHz at 1.481Vcore) and finally LOAD/overclocked comparing both utilities.

[STEvil]

why does S&M show only 49% load?

[Liquid3D]

I have a feeling it's due to Hyperthreading since it shows two processor's O and 1 instances, if you look at the little logo in red and blue on S&M's upper LEFT corner.

I just ran the program again and it indicated 67% on processor 0 and 64% on processor 1, however; when overclocked at some points it reached 92% which is also strange. This is most likely the P4's thermal/power monitoring features working. This is explained in Intel's VRD White Paper (ftp://download.intel.com/design/Pentium4/guides/30235603.pdf) section 2:4 onwards.

[IvanAndreevich]

Liquid3D
Um you have to use FPU only for burning the CPU.

[saaya]

Originally it only detected thermal throttling on P4, P-M and Xeon's, now it's A64 compatable. Obviosly temp is the ultimate test, this is just another tool, which shows TM1 or TM2 in real-time or logged.

a64s use t1 and t2 throttling?
:confused:

[jjcom]

I believe there power saving modes

jjcom

[IvanAndreevich]

New version people.. get it!

[IvanAndreevich]

http://members.shaw.ca/andrew-olga-ivan/temp/s&m_151.rar

Version 1.51 Beta with voltage & temp monitoring :D

[nCrusader]

When i tested the program at stock speed (3.2GHz) it aborted due to low +12v voltage. Got a Antec True Power 430W. What do you think, is it my PSU that causes the fault or could it be something else? How accurate are the measurments?

[Bloody_Sorcerer]

wowzer, 65.4 C with a 3200+ newcastle @ 2475 MHz & 1.6 vcore; swiftech MCX6400-V and 75 CFM 80mm fan. thats at least 10 C hotter than my 24/7 temps ("100%" load then; F@H and S@H).

[STEvil]

nCrusader - it probably aborted due to an error.

The voltage and temperature measurements are only as accurate as the hardware on your motherboard.

[nCrusader]

The program reported a low dip on the +12v @ 11.12v, which is very low afaik. I checked MBM and it reported a low +12v @ 11.55. This damn PSU has been giving some random problems undre heavy load, mostly in prime. I measured with a multimeter on 4pin and it reports a very stable 12.05 - 12.06v on the +12v rail under full load. WTF??
I got highend watercooling and a processor at stock should not fail a test like this.

[STEvil]

got a copy of the older version still?

I would test with it to be sure its not an error.

[IvanAndreevich]

ROFL there is a guy @ overclockers.ru in the official thread on this prog who took this prog to work and ran it on all the computers in one day. He had a few mobos and PSUs burn to hell @ stock ;)

[stealthbomber]

I find it to be a very useful addition to Prime95, S&M killed my new Enermax PSU within 2 hours of installing it, such stress test conditions help identify sub standard hardware parts, many which may not fail under "typical operating conditions" (ie office work)... :D

[jjcom]

:eek: I'm almost scared to run this new version...my computer already can't pass the test...and for some odd reason my computer has been running warmer...wait...my mom turned up the heat...maybe this is why I'm haveing trouble lol

jjcom

[STEvil]

its in my programs folder for good.. great program :D

[DevilsRejection]

So Ivan any luck on contacting the authour for an option to run the program for an unlimited time frame?

[IvanAndreevich]

I did. He said there is a DOS version that runs off a bootable floppy like memtest does. However, it's not as convenient IMO. He never really said YES I will do it or NO I won't :-\

[DevilsRejection]

dude the floppy version, which then can be made into a bootable iso, is a great thing indeed. now does it loop forever or thats the part wjere he was flaky with you about?

[IvanAndreevich]

The DOS one loops forever. However, it's 4% less effective in heating up the CPU :(

[bass]

The DOS one loops forever. However, it's 4% less effective in heating up the CPU :(

Can I please have a link for this :)

[IvanAndreevich]

It's available in the homepage ;)

[sr4470]

Thats weird. I had cheapo 235w PSUs in duron, celeron and t-bird rigs but they only blew after 5 years of moderately stressful operation (some oc, gaming etc)

[jmke]

It's available in the homepage ;)

do you have a direct link, my russian ain't what it used to be ;)

[IvanAndreevich]

Alright, here it is -

http://www.testmem.nm.ru/sem.zip

4KB heh

[stealthbomber]

Tried the latest version of S&M on my Enermax replacement, passed two 1 hour loops of FPU and PSU tests.

[jmke]

thank you very much IvanAndreevich!!

[IvanAndreevich]

jmke
Maybe you should do the review of that DOS version. Then lots of overclockers will be aware of it's existance and the author will be inclined to improve it further :)

[IvanAndreevich]

Burn baby burn!

[Liquid3D]

Ivan check out my Mainpage www.ksbrainstorms.com simply scroll way down when you arrive, and you'll find my little blurb on comparing S&M to SETI. I use S&M for all my CPU and waterblock testing now! In fact you'll find links to it in my latest review at the [M] on the Alphacol NexXxos BOLD WB; http://www.madshrimps.be/?action=getarticle&articID=275

[IvanAndreevich]

Liquid3D
Nicely done :D

ALL
I woke up @ 7am today with an idea. What if we suggest S&M author to get together with Memtest86 author to make a utility with both of those in it? That would be THE floppy to have.

[STEvil]

How about a CD?

[IvanAndreevich]

STEvil
Sure, if they could just integrate their softwares the rest would be just little technicalities.

[EyeOfThierry]

I'm running Prime95 and get 55c VRM...
Folding gets me to 63c VRM, 50c CPUs

(w/closed case) 10c less if without.

[Sampsa]

This is insane.

Pentium EE 840 and stock cooling: Asus Probe shows CPU temp 91C and S&M crashes after 8 minutes. All this with default clocks (3,2 GHz).

If I add 120mm fan right next to CPU cooler the Asus Probe shows 79C.

:confused:

[MrSeanKon]

Current version is 1.7.0.
As about my system (A64 3200+ 754 socket) it stresses more than Prime95 and it is more sensitive.
Also CPU temperature is higher at full load.

[IvanAndreevich]

Current version is 1.7.1 ;)

[Peen]

It got my CPU temp about .5c higher then hardest prime settings, and about 1c max difference. Passed the long test with just FPU selected

[Korpse]

D20L is the best load temp "finder" if you ask me - nothing better than running that for 7 days and seeing what your temp is :p

[STEvil]

d2ol doesnt get my temps as high as Prime95, and prime is average 5c less than S&M.

[ben805]

To test the cpu maximum heat load....there's absolutely nothing can beat TOAST, it runs 7~8'C hotter than this S&M!! with the TOAST.... at one point I got my 3700 SD up to 71'C within 3 min @1.71v, while the little wimpy prime95 only managed around 54~55'C after 20min :D :D

[STEvil]

A couple other neat things on this page....

http://testmem.nm.ru/soft.htm

[NetZeroZeus]

Heh, I just tried on my Dothan laptop...I tried getting the highest temperature possible by using I8kFanGUI to turn of all the fans and watch the temperatures soar, I never could get it to fail, it reached a plateau of 87c, with no fans on and the cpu oc'd on the thing to 2.3, another example of Dothan power...they don't seem to care about temps at all......heh the memry which is under the cpu, even though it wasn't being stressed, got all the way to 64c!