Overclocking

[eslagrassa]

Hey guys. Today i got my new psu (at my signature) , so i wanted to overclock my rig. It runs at 4.8 right now and its at 1.49. However when i tried to test it with prime 95 it probably reached to above 65 degrees and shut down itself. I just wanted to now, what is the ideal temparature for this cpu when it runs prime 95 ? How can i make that stabilized ?

Thanks for replys guys !

[Helltech]

You don't want it to go over 80c. If it shutting down at 65c then you have other issues, could be memory problems. 

[eslagrassa]

You don't want it to go over 80c. If it shutting down at 65c then you have other issues, could be memory problems. 

I started the test again with maximum fan speed. Its at 70 degrees and didnt shut down. Is that ok?

[Hamiac]

Like Helltech said, I doubt it will be temperature related.  A CPU will not usually force a shut-down unless around the 90 degree mark.    I'm not sure what changes you made to your bios so I cant advise what may be causing the instability.

[Hamiac]

Might be a good idea to revert all you settings to standard and run prime95 to see if it still shutsdown

[eslagrassa]

Might be a good idea to revert all you settings to standard and run prime95 to see if it still shutsdown

I used prime 95 before overclocking too, and it didnt fail. Rıght now its still running the test. How long should i keep that ?

[eslagrassa]

Nevermind it froze. Had to close psu. What should i do to not fail ?

[Grantelb4rt]

When your RAM runs at 2133 with your 4.8Ghz Overclock, i could imagine that this causes some Problems :P

I dont know if its the same for AMD CPUs but with Intel its pretty difficult to get those RAM speed at such an high overclock.

Set your RAM speed to 1333Mhz and test again. If you still run in to stability Problems your CPU might need more voltage ;)

[eslagrassa]

When your RAM runs at 2133 with your 4.8Ghz Overclock, i could imagine that this causes some Problems :P

I dont know if its the same for AMD CPUs but with Intel its pretty difficult to get those RAM speed at such an high overclock.

Set your RAM speed to 1333Mhz and test again. If you still run in to stability Problems your CPU might need more voltage ;)

I set the rams to 1866 before the test. Ok I ll make that 1333mhz . Hope for test does not fails :)

[Helltech]

You want to be able to run prime 95 for at least 24 hours without going over 80c to make sure its stable. 

[Opsilon]

Is your PSU starved for air?

is your PSU fan spinning?

go to some of your friends house and try the PSU on a different PC.

You should keep AMD CPU-s in 60- range, they throttle down after 68°C so i don't know why it shut down.

Your PSU should be able to take an CF 7970's and still have that overclock. try clocking the CPU to 4.6 GHZ and at least 6 h of prime and see if it fails then, freezing means instability so try lower CPU and around 1866 mhz ram speed

[eslagrassa]

You want to be able to run prime 95 for at least 24 hours without going over 80c to make sure its stable. 

Thank you !

Is your PSU starved for air?

is your PSU fan spinning?

go to some of your friends house and try the PSU on a different PC.

You should keep AMD CPU-s in 60- range, they throttle down after 68°C so i don't know why it shut down.

Your PSU should be able to take an CF 7970's and still have that overclock. try clocking the CPU to 4.6 GHZ and at least 6 h of prime and see if it fails then, freezing means instability so try lower CPU and around 1866 mhz ram speed

Case is fully open.

I dont know because it was too loud cause of h100i .

Ram is now at 1333 . Will try to test with that.

[Grantelb4rt]

Is your PSU starved for air?

is your PSU fan spinning?

go to some of your friends house and try the PSU on a different PC.

You should keep AMD CPU-s in 60- range, they throttle down after 68°C so i don't know why it shut down.

Your PSU should be able to take an CF 7970's and still have that overclock. try clocking the CPU to 4.6 GHZ and at least 6 h of prime and see if it fails then, freezing means instability so try lower CPU and around 1866 mhz ram speed

I personally dont think that the PSU is the Problem, I bet that the RAM or the CPU is causing the Problems.

I remembered that the old Phenom II series had a maximum Temperature of ~64°C and that the Temperature Sensors of those CPUs were pretty fucked up...

Could you please post a screenshot of the Temperature readings? With my old 1055T i had up to a 20°C difference depending on what sensor i read the Temperature from...

[mvitkun]

I've seen multiple threads about AM3+ motherboards where overheating vrm's cause it to crash or be otherwise unstable.

 

and since you have an h100i theres no incidental airflow hitting those vrm's.

 

here's my guess at what happened

first time you ran P95 at lower fan speeds.less air flowing over the vrm's,they ran hotter,surpasses safe temps.shut down to prevent damage.

second time you ran P95 with high fan speeds.more air flowing over vrm's,ran cooler, maintained temps within safe range.did not shut down

 

what case do you have?

is there an option to mount a side panel fan,if so mount one there so that the vrm's around the cpu get some cooling and it should be fine.

[Opsilon]

I had a problem with my corsair tx650, heat sensor was bad and after a hard workload it just turns it self off, and after 2-3 secs it starts on again, the fan was also broken... i got a new one now and all is fine, so i would not rush and say that the PSU isn't a problem yet. we need more testing to see

[eslagrassa]

I personally dont think that the PSU is the Problem, I bet that the RAM or the CPU is causing the Problems.
I remembered that the old Phenom II series had a maximum Temperature of ~64°C and that the Temperature Sensors of those CPUs were pretty fucked up...
Could you please post a screenshot of the Temperature readings? With my old 1055T i had up to a 20°C difference depending on what sensor i read the Temperature from...

http://i40.tinypic.com/oho9kk.png

Btw its when nothing happening.

I've seen multiple threads about AM3+ motherboards where overheating vrm's cause it to crash or be otherwise unstable.

and since you have an h100i theres no incidental airflow hitting those vrm's.

here's my guess at what happened
first time you ran P95 at lower fan speeds.less air flowing over the vrm's,they ran hotter,surpasses safe temps.shut down to prevent damage.
second time you ran P95 with high fan speeds.more air flowing over vrm's,ran cooler, maintained temps within safe range.did not shut down

what case do you have?
is there an option to mount a side panel fan,if so mount one there so that the vrm's around the cpu get some cooling and it should be fine.

http://www.amazon.com/gp/product/B009XES5GU/ref=oh_details_o00_s00_i00?ie=UTF8&psc=1
this is my case.
I was guessing so. But i ran the test again with 1333 ram freq and it shut down . Fans were 100%

I had a problem with my corsair tx650, heat sensor was bad and after a hard workload it just turns it self off, and after 2-3 secs it starts on again, the fan was also broken... i got a new one now and all is fine, so i would not rush and say that the PSU isn't a problem yet. we need more testing to see

I changed my psu today. It was corsair hx850 but it had coil whine and weird fan problem.What should i show next ?

[Grantelb4rt]

I've seen multiple threads about AM3+ motherboards where overheating vrm's cause it to crash or be otherwise unstable.

 

and since you have an h100i theres no incidental airflow hitting those vrm's.

 

here's my guess at what happened

first time you ran P95 at lower fan speeds.less air flowing over the vrm's,they ran hotter,surpasses safe temps.shut down to prevent damage.

second time you ran P95 with high fan speeds.more air flowing over vrm's,ran cooler, maintained temps within safe range.did not shut down

 

what case do you have?

is there an option to mount a side panel fan,if so mount one there so that the vrm's around the cpu get some cooling and it should be fine.

In my opinion an asus Crosshair V Formula-Z has one of the best cooling solutions for AM3+ boards, but the easiest way to find out if this causes the Problem is by putting a small Fan next to the VRM  :P

 

 

I had a problem with my corsair tx650, heat sensor was bad and after a hard workload it just turns it self off, and after 2-3 secs it starts on again, the fan was also broken... i got a new one now and all is fine, so i would not rush and say that the PSU isn't a problem yet. we need more testing to see

 

Yeah, I don't think that this is a common Problem, but you are right ^^ Better testing if it runs fine than searching for Problems everywhere else and making the discovery that in the end everything was PSU related   :rolleyes:

 

 

EDIT:

I just saw your Screenshot with the Temperatures. Same Problem with the AM3+ CPUs obiviously ^^ You have to add another 10-12 degrees to  the core temperature. I am pretty sure that the mainboard sensor usually reports a lower CPU Temperature than the Sensors on CPU die. In your case the Mainboard sensor reports 41°C, but the core temp is said to be 28°C

As I said, had the same Problems with my 1055T too :D

[eslagrassa]

Should i replace the rams ? Or just keep 1 ? I tried to run the test again and it failed.

[mvitkun]

http://i40.tinypic.com/oho9kk.png

Btw its when nothing happening.

 

http://www.amazon.com/gp/product/B009XES5GU/ref=oh_details_o00_s00_i00?ie=UTF8&psc=1

this is my case.

I was guessing so. But i ran the test again with 1333 ram freq and it shut down . Fans were 100%

 

I changed my psu today. It was corsair hx850 but it had coil whine and weird fan problem.What should i show next ?

 

yeah so the case supports 2 side panel fans,just place one fan on the higher up side panel fan mount and see if you can run p95

[eslagrassa]

yeah so the case supports 2 side panel fans,just place one fan on the higher up side panel fan mount and see if you can run p95

Actually my h100i mounted on that panel because motherboards heatsink is too big so i cant mount that to top of the case and i cant close the panel because fans are touching to 7970 's left side :D

[Opsilon]

I changed my psu today. It was corsair hx850 but it had coil whine and weird fan problem.What should i show next ?

Well now i am confused, you said you tried it with your new HX850, and that you had a bad PSU before this, well, i have a story to tell what i had with Corsair PSU-s so "hear" me out pls.

 

I bought a tx650 last year in September, the PSU was fine and all was well, but it started to whine after a month, i returned it to my store and after 2 weeks they sent me a new one. That one was screaming out of the box. On startup it screamed for 20 mins, i returned it to the store the next day, they tested it and they said it was fine... I got the PSU back and it was still screaming but i also saw that after 20 mins of work it stops, and i was like... yeah, sweet... After a week my PSU started turning it self off, and after 2-3 secs on again, and it turned out i was right from the start, the PSU was broken, returned it to the store and i got a new one 2 weeks later, this 3rd one is great, silent, powers everything i need and there was no hiccup since then, and that is now over 6 months. Sory for the long story but that is the problem i had, sometimes even a good branded PSU can be broken.

 

Note: i recommend testing for 6h or more for prime, OCCT and memtest86

Step 1 for the PSU i say, put your CPU to 4.6 ( apply a voltage that can keep that stable ), memory at 1333, and leave your gpu the way it is, run prime 95, after that for the 2nd test run prime 95 with Unigine or furmark in the same time ( OCCT has a CPU and GPU blend in the last build i believe, so try that one if you want to ), that should eliminate your PSU problem. If it fails the secound run that might mean that your gpu is broken, 3rd test try running unigine for 1 h, after that try 3dmark ( free on steam ) and last just let it run furmark, if that 3rd test passes, but the 2nd test test fails i am 90% sure it's your PSU

 

Step 2 If all ends well, download Memtest86 and test your memory, or you can just take OCCT and do ram with the CPU, but only after you tried the PSU test i gave you in step 1

 

Step 3 report what happened so we can help you more if possible or if we can't you can return your broken PC parts XD GL mate

[mvitkun]

was going to post this yesterday but lost internet.

 

since your already not in an ideal situation you could try this.

 

take the plastic top off of the case,mount the H100i fans on the outside top of the case,with the rad on the inside top of the case.that should give plenty of clearance for the heatsinks while allowing you to close the case.

 

then mount some fan to the side panel window to find out if the vrm's are overheating or if there is some other issue.

[eslagrassa]

Well now i am confused, you said you tried it with your new HX850, and that you had a bad PSU before this, well, i have a story to tell what i had with Corsair PSU-s so "hear" me out pls.

 

I bought a tx650 last year in September, the PSU was fine and all was well, but it started to whine after a month, i returned it to my store and after 2 weeks they sent me a new one. That one was screaming out of the box. On startup it screamed for 20 mins, i returned it to the store the next day, they tested it and they said it was fine... I got the PSU back and it was still screaming but i also saw that after 20 mins of work it stops, and i was like... yeah, sweet... After a week my PSU started turning it self off, and after 2-3 secs on again, and it turned out i was right from the start, the PSU was broken, returned it to the store and i got a new one 2 weeks later, this 3rd one is great, silent, powers everything i need and there was no hiccup since then, and that is now over 6 months. Sory for the long story but that is the problem i had, sometimes even a good branded PSU can be broken.

 

Note: i recommend testing for 6h or more for prime, OCCT and memtest86

Step 1 for the PSU i say, put your CPU to 4.6 ( apply a voltage that can keep that stable ), memory at 1333, and leave your gpu the way it is, run prime 95, after that for the 2nd test run prime 95 with Unigine or furmark in the same time ( OCCT has a CPU and GPU blend in the last build i believe, so try that one if you want to ), that should eliminate your PSU problem. If it fails the secound run that might mean that your gpu is broken, 3rd test try running unigine for 1 h, after that try 3dmark ( free on steam ) and last just let it run furmark, if that 3rd test passes, but the 2nd test test fails i am 90% sure it's your PSU

 

Step 2 If all ends well, download Memtest86 and test your memory, or you can just take OCCT and do ram with the CPU, but only after you tried the PSU test i gave you in step 1

 

Step 3 report what happened so we can help you more if possible or if we can't you can return your broken PC parts XD GL mate

I changed rams with corsair 2x8gb 1866. changed the freq to 1333. tried to test it  and failed.

 

was going to post this yesterday but lost internet.

 

since your already not in an ideal situation you could try this.

 

take the plastic top off of the case,mount the H100i fans on the outside top of the case,with the rad on the inside top of the case.that should give plenty of clearance for the heatsinks while allowing you to close the case.

 

then mount some fan to the side panel window to find out if the vrm's are overheating or if there is some other issue.

 

There is no plastic part ?

 

 

Btw test is running flawlessly at 4,6 with 1,46 v at 55 degrees.(with the 2x8 gb rams)

[mvitkun]

I changed rams with corsair 2x8gb 1866. changed the freq to 1333. tried to test it  and failed.

There is no plastic part ?

 

 

Btw test is running flawlessly at 4,6 with 1,46 v at 55 degrees.(with the 2x8 gb rams)

my bad had two pages open and was looking at wrong one lol.

okay so just try mounting the radiator to the outer top of the case and the fans to the inner top of the case.

 

 

when you say that it runs with 2x8GB ram do you mean that your trying a different set of ram or was it a typo.

because your sig says 4x4GB

[eslagrassa]

my bad had two pages open and was looking at wrong one lol.

okay so just try mounting the radiator to the outer top of the case and the fans to the inner top of the case.

 

 

when you say that it runs with 2x8GB ram do you mean that your trying a different set of ram or was it a typo.

because your sig says 4x4GB

These are my old rams. Yeah i tried with them and nothing changed still fail at 4.8 . That means memory is not a problem.

However i can run the test  without a problem at 4.6 with 1.46 v .

I mounted the fans and the radiator at top and it blows air to inside of the case. right now i can close the case . But the motherboards back panel is open.