is 90c too hot while running prime95

What CPU do you have? Should be fine if it's a Sandy or newer. Max is 100-105 before throttling down. Personally, 90c while hot, you'll likely never run that hot using the cpu under normal stress conditions.

With that said, I don't like to run past 80c. I don't like playing too close to the edge :)
 
iRUSH said:
What CPU do you have? Should be fine if it's a Sandy or newer. Max is 100-105 before throttling down. Personally, 90c while hot, you'll likely never run that hot using the cpu under normal stress conditions.

With that said, I don't like to run past 80c. I don't like playing too close to the edge :)
Click to expand...

3770k.
yeah 90 seemed high. I am waiting on my heaksink for the CPU still.
 
what clocks and voltage? ivy runs hot compared to previous cpu's
 
Then all fine bud, the stock cooler its a pile of crap..
 
That is too hot for Prime95 for that CPU IMHO. Prime95 is also outdated and not a good indicator of stability for the newer CPUs with AVX and AVX2 instruction sets. You need to use the latest version of IBT. On my 2500k at my clocks/voltage/cooling ill get low-mid 70's in P95, and near 90's with IBT with AVX support. For reference, BF3 will get it to mid-high 60's.

I realize your running stock clocks/volts/cooler, so if your 100% certain your HSF is mounted properly you may want to improve fresh airflow to the cooler, as well as improving hot air exhaust.
 
lessthanuthought said:
3770k.
yeah 90 seemed high. I am waiting on my heaksink for the CPU still.
Click to expand...



Yup, then I'd say that's totally normal. Stock coolers are so trashy they should just discount you whatever it cost AMD/Intel so you can front for a cheapo $30 cooler that will put it to absolute shame in comparisons.

On my Q9550 @ stock 2.83Ghz the stock cooler was doing 90-95 under full loads, not including Prime95 or IBT. It was pretty bad. Once you drop that aftermarket cooler with a 1Ghz OC it never reached higher than 70C under unrealistic Prime95/IBT conditions with one of the cores reaching a max of 75C (normal). Other than that full encoding of H.264 video, which is probably the best authentic way to test real world scenarios, or gaming, it would never come close to 70-75C.
 
nissanztt90 said:
That is too hot for Prime95 for that CPU IMHO. Prime95 is also outdated and not a good indicator of stability for the newer CPUs with AVX and AVX2 instruction sets. You need to use the latest version of IBT. On my 2500k at my clocks/voltage/cooling ill get low-mid 70's in P95, and near 90's with IBT with AVX support. For reference, BF3 will get it to mid-high 60's.

I realize your running stock clocks/volts/cooler, so if your 100% certain your HSF is mounted properly you may want to improve fresh airflow to the cooler, as well as improving hot air exhaust.
Click to expand...

I'll get IBT, and use that from now on thanks. I'm 99% sure the HSF is on correctly, and I have 5 noctua fans running in an Fractal R4, it seems that airflow should be good, no?

Thanks everyone for the quick responses so far.
 
Liger88 said:
Yup, then I'd say that's totally normal. Stock coolers are so trashy they should just discount you whatever it cost AMD/Intel so you can front for a cheapo $30 cooler that will put it to absolute shame in comparisons.

On my Q9550 @ stock 2.83Ghz the stock cooler was doing 90-95 under full loads, not including Prime95 or IBT. It was pretty bad. Once you drop that aftermarket cooler with a 1Ghz OC it never reached higher than 70C under unrealistic Prime95/IBT conditions with one of the cores reaching a max of 75C (normal). Other than that full encoding of H.264 video, which is probably the best authentic way to test real world scenarios, or gaming, it would never come close to 70-75C.
Click to expand...

I was just playing Crysis 3 and it only was in the mid 60s. I have some videos I have to transcode later, I'll check out the temp while doing them as well. Right now I'm idling at 39c. MB is at 31c, GPU is at 44c.
 
lessthanuthought said:
I'll get IBT, and use that from now on thanks. I'm 99% sure the HSF is on correctly, and I have 5 noctua fans running in an Fractal R4, it seems that airflow should be good, no?

Thanks everyone for the quick responses so far.
Click to expand...

The intel HSF are notoriously shitty for being difficult to mount. Easiest way to tell is to look at the back side of the motherboard and see that all the little black pegs have popped through all the way. Your motherboard will be bowed slightly if its installed correctly, they suck that much.

5 noctua fans...depends...are they oriented to create flow or are they just randomly blowing in and/or out? A typical flow is front to back, bottom to top.
 
nissanztt90 said:
Prime95 is also outdated and not a good indicator of stability for the newer CPUs with AVX and AVX2 instruction sets.
Click to expand...

Wrong on so many levels. Prime95 has supported AVX shortly after Sandy Bridge was released. And the author is currently implementing more instruction sets for Haswell CPUs.

And the creator of IBT hasn't updated his program for the latest Linpack. If you want to take advantage of AVX2, you need to use LinX v0.6.5.
 
nissanztt90 said:
The intel HSF are notoriously shitty for being difficult to mount. Easiest way to tell is to look at the back side of the motherboard and see that all the little black pegs have popped through all the way. Your motherboard will be bowed slightly if its installed correctly, they suck that much.

5 noctua fans...depends...are they oriented to create flow or are they just randomly blowing in and/or out? A typical flow is front to back, bottom to top.
Click to expand...

lessthanuthought said:
I was just playing Crysis 3 and it only was in the mid 60s. I have some videos I have to transcode later, I'll check out the temp while doing them as well. Right now I'm idling at 39c. MB is at 31c, GPU is at 44c.
Click to expand...

its totally fine do not worry.. you are even good if you can keep 90C on a 3770K with prime 95.

and mid 60s with stock cooler playing crysis 3? thats mean good airflow in the case, in some cases people cant even run p95 without make the CPU throttle due overheat..

SolidBladez said:
Wrong on so many levels. Prime95 has supported AVX shortly after Sandy Bridge was released. And the author is currently implementing more instruction sets for Haswell CPUs.

And the creator of IBT hasn't updated his program for the latest Linpack. If you want to take advantage of AVX2, you need to use LinX v0.6.5.
Click to expand...

lessthanuthought said:
I'll get IBT, and use that from now on thanks. I'm 99% sure the HSF is on correctly, and I have 5 noctua fans running in an Fractal R4, it seems that airflow should be good, no?

Thanks everyone for the quick responses so far.
Click to expand...

^+1

Also not the same Use prime95Blend (who does not force the FPU to the max,) than use prime 95 Small FFT.. Primer95 its still good it only have some kind of overvoltage with haswell, with sandy and ivy are fine.
 
SolidBladez said:
Wrong on so many levels. Prime95 has supported AVX shortly after Sandy Bridge was released. And the author is currently implementing more instruction sets for Haswell CPUs.

And the creator of IBT hasn't updated his program for the latest Linpack. If you want to take advantage of AVX2, you need to use LinX v0.6.5.
Click to expand...

Oh boy, nice try but im not wrong. Point to where i said Prime95 didnt support AVX. IBT with AVX tortures a CPU way more than P95, i literally downloaded the new version of P95 and tested right before i made the initial post, then ran IBTv2 with AVX. Stop trying to call people out for no reason and with no backing. The only thing i didnt know was that IBT didnt support AVX 2.0 yet, because...brace yourself...i have a SB and havent needed an update to IBT. My fucking bad.
 
nissanztt90 said:
Oh boy, nice try but im not wrong. Point to where i said Prime95 didnt support AVX.
Click to expand...

Right here:

nissanztt90 said:
Prime95 is also outdated and not a good indicator of stability for the newer CPUs with AVX and AVX2 instruction sets.
Click to expand...

You said that it is outdated, meaning it doesn't support CPUs with AVX or AVX2. Obviously that isn't true.

nissanztt90 said:
IBT with AVX tortures a CPU way more than P95, i literally downloaded the new version of P95 and tested right before i made the initial post, then ran IBTv2 with AVX. Stop trying to call people out for no reason and with no backing. The only thing i didnt know was that IBT didnt support AVX 2.0 yet, because...brace yourself...i have a SB and havent needed an update to IBT. My fucking bad.
Click to expand...

Try running Prime95 with small FFTs for about an hour or two. It should heat up just as much as any Linpack program.

And it doesn't matter if you own an older processor or not. You can still read up on this "new" stuff.
 
nissanztt90 said:
The intel HSF are notoriously shitty for being difficult to mount. Easiest way to tell is to look at the back side of the motherboard and see that all the little black pegs have popped through all the way. Your motherboard will be bowed slightly if its installed correctly, they suck that much.

5 noctua fans...depends...are they oriented to create flow or are they just randomly blowing in and/or out? A typical flow is front to back, bottom to top.
Click to expand...

url]


The fans are two intakes in the front, one intake on the side, one exhaust on top, one exhaust in the back.
 
SolidBladez said:
Right here:



You said that it is outdated, meaning it doesn't support CPUs with AVX or AVX2. Obviously that isn't true.



Try running Prime95 with small FFTs for about an hour or two. It should heat up just as much as any Linpack program.

And it doesn't matter if you own an older processor or not. You can still read up on this "new" stuff.
Click to expand...

It is outdated, whether youd like to accept that or not is up to you. Its old. There are better stress testers like linpack now. That doesnt mean it doesnt support newer processors with AVX. If you want to misconstrue what i said, then by all means...be wrong. I can make my system P95 stable for hours, and watch it crash and burn as soon as IBT is fired up. Calling P95 a good indicator of stability these days is an exercise in futility. Again, stop trying to be mr always right on the internet. Youre wrong. I spent plenty of time making my 2500k 100% stable at 5.0ghz with a high end water setup, gonna go ahead and say i know what im doing.

You saying what "should" happen when running P95 really has no bearing on what i watch happen...so meh. But seriously...my fucking bad on not staying on top of a CPU stress testing program i havent needed in a long time. Seriously bro...im so sorry for not keeping on top of that.

lessthanuthought said:
url]


The fans are two intakes in the front, one intake on the side, one exhaust on top, one exhaust in the back.
Click to expand...

Good setup. You see how the little black pegs are sticking through far enough to have the 2 white halves be totally split? That means they are fully through and mounted correctly. You have a cable blocking the 4th one, so if thats good, then youre mounting is good.
 
Last edited:
Yes 90C is pretty damn hot! Just imagine, your only 10C away from the system boiling away.
 
nissanztt90 said:
It is outdated, whether youd like to accept that or not is up to you. Its old. There are better stress testers like linpack now. That doesnt mean it doesnt support newer processors with AVX. If you want to misconstrue what i said, then by all means...be wrong.
Click to expand...

You are wrong, plain and simple. You are basing your facts on the fact that linpack is better, when it really isn't. You can run an "old" version of Linpack/IBT and be stable, yet crash and burn with the latest binaries. The same thing goes for running an "old" version of Pime95. The author was quick to add to in AVX support, and is working on AVX2..Considering he does this on his own and isn't a huge company like Intel, then I would say that is pretty good.

Your reasoning behind Prime being inferior, is just that inferior. You are using a quad core without HT, which is MUCH easier to get 100% stable vs a cpu with HT..Those of us with HT and want TRUE stability testing with IBT/Linpack have to launch the program, go into Task Manager, lock the affinity to the proper cores, then test..This gets old quick. With Prime95, we launch the program, and let it run. Simple as that.

I have used BOTH programs, and have screenshots that show the same temperatures (+/- 2C) for short runs (10 to 30 mins). Once you start doing long term testing (100 runs of IBT on "Extreme", 24hr+ Prime95), Prime consistently produced higher temperatures.

Disclosure: I also have a "high end" water setup:p
 
ccityinstaller said:
You are wrong, plain and simple. You are basing your facts on the fact that linpack is better, when it really isn't. You can run an "old" version of Linpack/IBT and be stable, yet crash and burn with the latest binaries. The same thing goes for running an "old" version of Pime95. The author was quick to add to in AVX support, and is working on AVX2..Considering he does this on his own and isn't a huge company like Intel, then I would say that is pretty good.

Your reasoning behind Prime being inferior, is just that inferior. You are using a quad core without HT, which is MUCH easier to get 100% stable vs a cpu with HT..Those of us with HT and want TRUE stability testing with IBT/Linpack have to launch the program, go into Task Manager, lock the affinity to the proper cores, then test..This gets old quick. With Prime95, we launch the program, and let it run. Simple as that.

I have used BOTH programs, and have screenshots that show the same temperatures (+/- 2C) for short runs (10 to 30 mins). Once you start doing long term testing (100 runs of IBT on "Extreme", 24hr+ Prime95), Prime consistently produced higher temperatures.

Disclosure: I also have a "high end" water setup:p
Click to expand...

You are wrong, plain and simple. I can spout most of the same shit back to you in an opposing fashion...nothing you said makes you right, nor me wrong.

Ive used both as well. They dont produce a difference of 2 degrees. Fluctuations between cores are frequently more than 5 degrees. But you would have had to play with multiple CPUs to know that. Im talking 15 degree differences between the 2 test methodologies, and can also provide screen shots. Im also talking about my CPU being Prime stable for 12 hours at 1.42v. Then firing up IBT and having it crash with in seconds. Youre just one more random nobody on the internet that is in the mood to disagree with me. Good bye.
 
nissanztt90 said:
You are wrong, plain and simple. I can spout most of the same shit back to you in an opposing fashion...nothing you said makes you right, nor me wrong.

Ive used both as well. They dont produce a difference of 2 degrees. Fluctuations between cores are frequently more than 5 degrees. But you would have had to play with multiple CPUs to know that. Im talking 15 degree differences between the 2 test methodologies, and can also provide screen shots. Im also talking about my CPU being Prime stable for 12 hours at 1.42v. Then firing up IBT and having it crash with in seconds. Youre just one more random nobody on the internet that is in the mood to disagree with me. Good bye.
Click to expand...

so.. do it, prove those 15 C difference between Prime95 27.9 Small FFT and IBT... Do it.. you can even use the new intel linpack for haswell that work pretty well with sandy and ivy...
 
Araxie said:
so.. do it, prove those 15 C difference between Prime95 27.9 Small FFT and IBT... Do it.. you can even use the new intel linpack for haswell that work pretty well with sandy and ivy...
Click to expand...

small fft give a 4-8 degree difference across cores.
 
nissanztt90 said:
They dont produce a difference of 2 degrees. Fluctuations between cores are frequently more than 5 degrees.
Click to expand...

Now we are talking about core fluctuations? I was referring to peak temperatures, since it does no good to compare the "lowest" core temperatures if the hottest core is outside the safety zone you choose.

But you would have had to play with multiple CPUs to know that.
Click to expand...

I've done 33 builds I've done over the last 18 months all O/C'd..5 have been Socket 1366 i7s, 8 have been SB, and 20 have been Ivy 3770Ks, with 15 of them using the EXACT MB..I think that qualifies me as having "played" with multiple CPUs :rolleyes:...

Im talking 15 degree differences between the 2 test methodologies, and can also provide screen shots.
Click to expand...

If you can show me a 15C difference between a 24hr run on Prime95 (Small FFT) and a 100 pass run of IBT, with guaranteed environmental variables (same fan/pump speed, same ambient temps) then I will publicly kiss your arse*..

Youre just one more random nobody on the internet that is in the mood to disagree with me.
Click to expand...

Boy, somebodies God like superiority complex is certainly running full steam:eek::eek:..

*(It has to be YOUR SS, not some random intenet grab)
 
nissanztt90 said:
It is outdated, whether youd like to accept that or not is up to you. Its old. There are better stress testers like linpack now. That doesnt mean it doesnt support newer processors with AVX. If you want to misconstrue what i said, then by all means...be wrong. I can make my system P95 stable for hours, and watch it crash and burn as soon as IBT is fired up. Calling P95 a good indicator of stability these days is an exercise in futility.
Click to expand...

And you can be IBT stable for hours, yet crash in Prime95. Thet both have its purpose and use.
 
Wierd thing happened. so I was running Prime95 in the first place testing my RAM, it ran four a couple hours, no problem. Then 2 days later my computer wont boot, just turns on and reboots. I clear the bios, and it boots. All is fine for a couple hours, then the same thing. I take one stick out and it boots fine now. So I call to RMA the RAM and then I decide to just try it in a different slot, now it has been working fine for a week. Anyone have any idea what is going on?
 
You must log in or register to reply here.
Back
Top