Should I back off my Overclock?Post Date: 2016-05-20 |
Post Reply
|
Author | |
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply
Topic: Should I back off my Overclock? Posted: 20 May 2016 at 7:39pm |
I had been having some issues with my system that I thought could be related to Overclock on my Vanquish system so I removed the overclock and went with default setting.
It turned out to be the Seagate 1TB drive that was the issue which I discovered when it died on me. So I put a new HD in that I had laying around. I decided to put my Overclock back and loaded up the profile and got my 4.2 Ghz back. I had never really stressed tested it when I got it as such other than running games and a benchmarking program. So I decided to make sure it was stable. I noticed the cpu voltage was at 1.285. I hit Prime95 and watched the temps. when they hit 100C I stopped it right away. It did that within a few minutes? maybe 2-3 max. I shut the system down. Cleaned out the whole case and tried again. 100+c with in a few minutes stop Prime95 again. Started backing down voltages after that. I am down to 1.225 at the moment. It still hits 100+c in Prime95 although it seems to get there slightly slower but still too fast. The system is stable in the games I play. Wow, Marvel Heroes, Civ 5. It gets into the high 70's to low 80's in WoW and low 90's in Marvel Heroes. Fans are at max. I did add a GTX 970 but it was a blower style so likely is not adding too much heat. Think I need to lower voltages more? Just remove the overclock for safety? I believe my system came with the Corsair H60 so maybe that is just having issues keeping it cool. |
|
|
|
db188
DS Veteran Joined: 29 Jul 2014 Online Status: Offline Posts: 2115 |
Quote Reply Posted: 20 May 2016 at 8:53pm |
the H60 is only a 120mm rad cooler, so it's not going to give you a lot performance especially when oc'ing. a high end air cooler like the Noctua D14 kicks its butt.
Prime95 and IBT are really hard on your proc. they produce unrealistic heat and loads on the proc that you'll never see in any real world applications. i wouldn't use them for stability testing with your cooling setup. try Asus RealBench bench mark or Heaven instead. reset back to default frequencies and give us your base line temps and voltage vcore. can you list your system specs for us to help provide tailored advice? if you've had this system for a while you might need to take the fans off the rad and remove dust from them and the rad. you might also take the opportunity to remove the heat sink, remove old thermal paste and reapply new. once i know your system specs i can offer better oc'ing advice. a few general practices i employ are: 1) set everything to default frequencies and voltages 2) find my under volt limit for base frequencies by reducing vcore until unstable 3) raise the multiplier for all cores until bsod 4) incrementally raise vcore 1 step at a time until no bsod 5) decide whether or not to continue raising the multiplier and vcore for a higher oc 6) test for stability with the frequency and voltage that appears to be good by running an application that simulates real world usage (REalBench and Heaven are good). if any instability (bsod, freezes, stuttering, etc.) occurs during the testing, decide whether to add more voltage (up to the manufacturer's recommended limit) or back off the frequency. |
|
Aventum 3
I7-6700K Gigabyte G1 Z170X Gaming GT 16GB Corsair Dominator 3000MHz Corsair Hx1000i 1000W Samsung M.2 980 Pro 2TB;Samsung 850 EVO 1TB MSI RTX 3080 Ventus OC 10G LHR Gigabyte M28U 4K |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 20 May 2016 at 10:57pm |
Processor: Intel Core i5 3570K 3.40 GHz (Unlocked CPU) (Quad Core) Motherboard: ASUS P8Z77-V LX (Intel Z77 Chipset) System Memory: 16GB DDR3 1600MHz Corsair Vengeance Series (High-Performance) Power Supply: 600W Corsair CX Edited by Ariktu - 20 May 2016 at 10:58pm |
|
|
|
db188
DS Veteran Joined: 29 Jul 2014 Online Status: Offline Posts: 2115 |
Quote Reply Posted: 21 May 2016 at 12:29am |
Ivy Bridge is going to run hot, expect to see temps in the 70-80 range under load (gaming), especially with the limited cooling you have. if your oc is stable you can leave it alone or you can try lowering the voltage in 1 step increments to see how low you can get it stable at with that frequency. don't use Prime95 or IBT for stability testing, that's why you're seeing 100+degree temps and the cpu is probably throttling down or even shutting down at that point.
i'd expect most of those procs capable of hitting 4.2GHz @1.2V with adequate cooling. i wouldn't advise overclocking at all on a stock cooler. i wouldn't advise you go higher w/o moving to a bigger rad (240mm at least). Edited by db188 - 21 May 2016 at 12:45am |
|
Aventum 3
I7-6700K Gigabyte G1 Z170X Gaming GT 16GB Corsair Dominator 3000MHz Corsair Hx1000i 1000W Samsung M.2 980 Pro 2TB;Samsung 850 EVO 1TB MSI RTX 3080 Ventus OC 10G LHR Gigabyte M28U 4K |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 21 May 2016 at 12:39am |
I dropped the voltage down to 1.20 and ran a couple of games for a bit and ran Heaven bench once thru and it seems stable so far.
I will mess around with that voltage for a bit to see if it is stable. I guess it is true that I am likely never pushing my system at 100% very often so odds are the temps will never get close to what those stress programs show. I am going to see about removing and cleaning behind the fan on the CPU cooler to see if that helps some also. |
|
|
|
db188
DS Veteran Joined: 29 Jul 2014 Online Status: Offline Posts: 2115 |
Quote Reply Posted: 21 May 2016 at 12:47am |
how have your temps responded to the lower voltage?
|
|
Aventum 3
I7-6700K Gigabyte G1 Z170X Gaming GT 16GB Corsair Dominator 3000MHz Corsair Hx1000i 1000W Samsung M.2 980 Pro 2TB;Samsung 850 EVO 1TB MSI RTX 3080 Ventus OC 10G LHR Gigabyte M28U 4K |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 21 May 2016 at 1:18am |
They are a couple of degrees lower although I have not tried any of the stress tests again.
I want to at least look at the CPU fan and radiator and see if it needs to be cleaned. Last time I did not remove the fan when I blew out the dust so I am assuming there must be dust behind the fan and on that radiator that I can not see. |
|
|
|
Meller
DS Veteran Joined: 20 Feb 2013 Online Status: Offline Posts: 1627 |
Quote Reply Posted: 21 May 2016 at 7:28am |
Don't use Heaven for your benchmark if you're OC'ing your cpu by the way. It doesn't stress your CPU to determine stability. Cinebench or ASUS RealBench will work.
Those temps are really high. I've worked with several 3570k's, and I am thinking it's not your OC that is causing everything to run so hot. That H60 is a very slim profile 120mm rad, but it should be able to handle a 4.2ghz 1.25v overclock with no issues. I'm betting that we'll still be seeing high temps even without an OC. Has it always ran this hot, since day 1? Have you tried cleaning your cpu, reapplying thermal past, reseating the cooler, and also taking some canned air to blow the radiator out? |
|
Custom PC
Ryzen 9 5950X Gigabyte X570 Aorus Xtreme 128GB DDR4 3600mhz EVGA RTX 3090 FTW3 Ultra Gaming Samsung 980 Pro 512GB m.2 Samsung 960 PRO 2TB m.2 x2 |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 21 May 2016 at 9:33am |
It ran in the mid 80's C when I got it from what I remember if the system was pushed real hard.
So it is running around 10C higher than I remember. Ran RealBench Benchmark and it completed top Temp Core 0 was 94C Core 1 was 93C Core 2 was 91C and core 3 was 84C top temp it hit. My results were Sat May 21 2016 09:26:13 Image Editing: 165632 Time: 72.377 Encoding: 99225 Time: 120.816 OpenCL: 70503 KSamples/sec: 1690 Heavy Multitasking: 84183 Time: 142.403 System Score: 90785 I am hoping a better cleaning of the fan and radiator will help. If not I guess I can buy some thermal paste and try reseating the CPU although I will admit it has been awhile since I have done that. Edited by Ariktu - 21 May 2016 at 9:34am |
|
|
|
Meller
DS Veteran Joined: 20 Feb 2013 Online Status: Offline Posts: 1627 |
Quote Reply Posted: 21 May 2016 at 9:46am |
Don't reseat the CPU. Just unscrew the cpu water block, and clean off the thermal paste on the CPU and the bottom of the block.
Then apply a nice pea size amount in the middle of the CPU, and screw back down the waterblock. No reason to bull the cpu out and reseat it also. |
|
Custom PC
Ryzen 9 5950X Gigabyte X570 Aorus Xtreme 128GB DDR4 3600mhz EVGA RTX 3090 FTW3 Ultra Gaming Samsung 980 Pro 512GB m.2 Samsung 960 PRO 2TB m.2 x2 |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 31 Jul 2016 at 2:26pm |
Finally had some time and had gathered what I needed.
Pulled off the cooler and cleaned off the CPU and cooler and applied Noctua NT-H1 paste. Ran RealBench Benchmark and it completed top Temp Core 0 was 71C Core 1 was 72C Core 2 was 69C and core 3 was 70C top temp it hit. Sun Jul 31 2016 14:20:28 Image Editing: 174049 Time: 29.233 Encoding: 102966 Time: 93.118 OpenCL: 70503 KSamples/sec: 1705 Heavy Multitasking: 97362 Time: 100.449 System Score: 97119 ---------------------------- Seems like I got some decent improvement of around 20C lower temps or so. I understand the cooler is not the best, but it seems like it will do much better with the new paste. |
|
|
|
DST4ME
DS ELITE Joined: 14 Apr 2008 Online Status: Offline Posts: 36758 |
Quote Reply Posted: 31 Jul 2016 at 3:38pm |
It was not the paste, your block was not sitting correctly on the cpu, by reseating the "block", you fixed that issue, great job
Your temps are great how are they with linx or prime? Edited by DST4ME - 31 Jul 2016 at 3:39pm |
|
Ariktu
Groupie Joined: 21 May 2013 Online Status: Offline Posts: 492 |
Quote Reply Posted: 31 Jul 2016 at 7:31pm |
It seemed like it could be the paste actually. there was none(that I could see) in the middle of the CPU or block at all.
All of the paste was at the sides of the CPU and block with most of the paste on one side of it. Almost looked like it had been lifted off and moved or something when it was first put on, but what do I know. Just glad it seems better now. |
|
|
|
Post Reply |
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |