Prime95 is still running thats why its at 55C before that it was 44C at idle.

Both Hmoniter and Everest are reading the same temps.

 

just seen your edit and stopped the test. CPU is now at 50C and dropping.

 

CPU is now back to 43C/44C at idle.

 

That's looking very good matey.

Do you have 3dMark05?

If you do, can you loop the CPU test say 3 times? It will crash if your system is at all flakey.

 

I did have 3dmark05 until i wiped the PC clean a couple of days ago to do a "fresh" install, its 180mb, so that may take a while lol, is it a must to download it again?

When i ran it everything went smooth, came back with a score of 7200

 

No, its not a must, it is just another way to test stability. However if Prime95 was still running, then that's not a problem

Just another thing, the +12v looked good in Everest, but not so good in Hmonitor. It may have been that Hmonitor was reading it wrong. But can you open both and tell me the +12v readings please, and also the bios reading if possible?

Then, go and play a game, like the one you were playing when it crashed, and tell me how you get on.

 

Everest is reading 13.07v

Hmonitor is reading 10.30v

You gonna be online for the next 20 mins or so while i test a game?

 

Should be, but could you tell me the bios reading first?

 

Yup, it was 11.932v

Il have a quick blast on BF2 and come back

 

No probs. The voltage reading is fine. Clearly Hmonitor picks it up wrong.

 

Ok just played BF2, although i got kicked from two servers for high ping everything was fine, highest temp reading i got was 53C for the CPU and 33C for the motherboard.

Although it didnt crash, sometimes i can be gaming for upto 3 hours before it does!

 

Looking good so far though. I honestly believe the temp issue was the problem.

I will watch this thread for further updates tomorrow.

Failing that, any probs, PM me.

 

Cheers for all your help, I'll keep you posted.

 

No probs, anytime

 

I wait with baited breath. Mainly because I still think my theory is more likely than yours, but I'm happy to be proved wrong if it means someone gets their computer to work. And disagreeing with you is not thread-cr*pping BTW, it's just disagreeing with you. Even on SN you are allowed to disagree with someone.

While AMD may well say the "recommended"CPU max temp is 70oC, the actual max CPU sustained temp is 85oC - it's called a margin. Those chips have to be able to work in a open office with no air-con in (say) Bangalore, where the CPU will get a lot hotter the than 55oC here. THAT is why I made my comment - to point out that the errant CPU is not too hot. If there is a an overheating problem (which I remain unconvinced about) then I'd say it's the GPU, not the CPU. CPU overheating nearly always results in lockups, not reboots. If there are BSODs before the reboots then it might be RAM, if not it's usually PSU.

Also, absolute voltage values mean little as long as the PSU remains with the ATX spec - which IIRC is +/- 10% for the 12v rail. What matters is how the rail holds up under load - particlarly the 12v rail as most motherboards nowadays use that to supply the CPU. Obviously the nominal power means nothing: there are 400W PSUs that out-perform some 500W PSUs. That said, at least one motherboard manufacturer now recommends 500W PSUs for its A64 boards, and most at least suggest it for when you are running power-hungry gfx cards like the 7800GTX.

But we shall see.


M

 

On another note, what exactly did you do Mr Manchester? As far as I can see from rereading this thread, you took the side off, took some pics, put the side back, then added some software? Even if it is an overheating issue, I seem to have missed the bit where you did something to change the situation?

(BTW, fans always blow towards the label).

M

 

These are the 2 previous posts you made in this thread. That is why I said you thread crapped. Neither of those posts was at all helpful in any way! And both were complete rubbish to boot! I have seen numerous reboots in my time due to all manner of things without BSOD. It is not exclusively PSU caused!

And the comment about the PSU being inadequate was just garbage posting too. I was trying to see if perhaps the reporting software was reading the rail wrongly, which I believe it was. But still you had to make a snide comment!

As for advising anyone that temps of 70c are fine, that is just a downright dangerous, loose comment. AMD list 70c as the max die temp for a reason. Above 70c for a prolonged period will damage the die - no question.

If you had a CPU at 70c and above for 2 years, then it was another core, most likely an XP (which had a max die temp of 90-95c). A Athlon 64 will be damaged running at 70c or higher.

I am also fully aware of the ATX spec, you are correct it is indeed +-10%. Which would effectively allow the +12v rail to drop to 10.8v before being below specs. The post on page 1 of the bios screen shows the +12v rail to be reading at 11.8, well within spec.

Also, bear in mind that the PSU he is using is a dual 12v rail version. It produces 14 amps on rail 1 and 15 amps on rail 2, totalling 29 amps. The +5v rail produces 30 amps. The total power output for 12v is 384 watts sustained.

Using a PSU calculator indicates that the system Mr M has will draw 320 watts of power approx. Meaning that the 384 watts is plenty! I will also link you to a review that tests power consumption. The total under load with a P4 660 and 7800 GTX was 271 watts!

This is more than enough to power his system. Lack of power is definately not the issue here.

Oh and as for manufacturers 'suggesting it' that makes little to no odds. Nvidia recommend a 350w PSU for single 7800 GTX and 500w for dual.

And the chips do not 'have to be able to work in a open office with no air-con in (say) Bangalore'

I have never seen AMD mention that in any of their product specs or literature.

But hey, we live and learn. I was at least trying to help the guy get his system working. Maybe you could do likewise.

 

He did not have the side on the case for starters, causing an airflow issue. And also we confirmed the fans were pushing air in the correct direction.

Oh and not all fans blow towards the label. All heatsinks that I have used in the last 3 years have had a label on the top of the fan, and they most definately blew away from it. I know what you mean, but it is not quite a correct statement.

 

Meridian i know what your saying, we have installed software and nothings changed but it has ruled out the temperature and power issues.

D16GER put my mind at ease, at the the end of the day i have spent over 400 quid on this bloody thing and its worse than it was before i started! so what else can it be? faulty motherboard? graphics card? im pissed off

 

Erm, when in the bios the cpu is under full load, that's why temp seems high, so 50-55C is normal. Your temps are fine. I would suspect a gfx card / ram instability here, especially seeing as you can game 'sometimes' for 3hrs or more. If memtest reported no errors on the RAM, then it's probably the gfx drivers and a combination of something else it doesn't like to play with in your system.

 

Where did you pull that from??

That is complete and utter nonsense!! The CPU is nowhere even close to 100% load when in the bios

 

BTW Mr. M, how is the system today? Any crashes yet??

Oh and for info a graphics card problem, especially heat will cause anomalies like white dots or black artifacts long before it crashes, so if you have not noticed these, then it's fine.

As for memory errors, they normally, but not always, cause the system to lock before a reboot, which you are not experiencing.

 

lol yes it is.

 

So you are saying that the CPU is under the same load in the bios as it is under during, say Prime95??

Proof of this please? Come on, you cannot make a statement like that without backing it up.

 

Ok not 100% load but it is under some load which doesn't gives u an idle temp reading Mine is the same it shows 55C in bios, yet in Windows idle temp is 42C using MBM5.

Anyway even MBM5 or alike are inaccurate on some mobos ( and can give +5-10% readings over actual). The only way to know for sure is to use a hware monitor.

 

Seems ok today, although i havnt been on any games yet.

One thing a forgot to mention was during gaming just before it crashes it will freeze for around 4 seconds and then reboot.

 

Nope - 69oC as I look at MBM and looking good.

You are correct that BIOS is not 100% load, but the temps that show from staying in BIOS for a long time suggest it is probably about 80-90%. This is probably because the BIOS can't send an Idle command to the CPU.

Now for some more thread-cr*pping . Or disagreement as adults call it.

And overheating gfx components MAY cause artifacts, but not always. Particularly if they are not overclocked - overclocking seems to increase the likelihood of artifacts. I've certainly seen gfx cards lock up from overheating with no artifacts visible.

And I didn't suggest a temperature of 70oC was "fine" in the sense you appear to mean (a good idea), I suggested that it needs a lot higher temp than you seem to think to cause CPU faults. And as I said, CPU overheating nearly always causes lockups, plus occasionally crashing apps/games back to desktop.


M

 

Can I refer you to your earlier post....
Looks like you said fine twice in there, that's pretty conclusive from where I see it.

And as for the adult bit, catch a grip of yourself. You know full well you thread crapped, just dont want to admit that you did.

And, no this time I dont believe you crapped on the thread, and yes I am inclined to agree that artifacts are not always seen when the card overheats, but they often are.

Lets leave the argument/discussion there please.

Oh, and back to topic, Mr Manchester - one more thing I wanted to ask, did you install the AMD processor driver and enable Cool & Quiet in the bios?

If you did not, it might be an idea to do so, as it will improve the life of your CPU until you can get the cooling sorted out.

 

As with any PC world PC these days it came with no drivers or anything any idea where to get it from?

 

my brothers coming round tomorrow with a better model 500w PSU, il try that for a few days and see what it does for the PC.