OCCT or Furmark regarding GPU temperature?
I'm currently overclocking my GeForce GTX 560 TI (the version that came pre-overclocked to 900 - engine, and 4200/2100 memory) a little bit, and I was just wondering whether I should Use OCCT or Furmark regarding temperatur-readings. Somebody says Furmark is insanely high on temps, and that there's no point using that as a reference when gaming or using the GPU otherwise. I don't know about OCCT, but I'm guessing the max gpu temp in OCCT is a bit higher than max temp in actual use. How far would you've gone with the temp (as long as the testing is fine and stabil otherwise of course) in OCCT? I'm currently at 82 degrees in a default 1hour test with error checking. And I'm at 930 engine and 4226 memory. I was thinking about investing in a GPU-cooler, but I'm wondering if I really need to...
what's your GPU model ? all the stress test programs push your card to the limits so it's very normal to have these degrees. you can check the GPU temp while playing games, if it reaches that degree you might get an aftermarket GPU cooler.
and i kinda understood from the post that you need a temp monitor software, try GPU-Z and CPUID HW Monitor
Thank you, and yes, I needed a temp monitor that logged I've played Crysis about 10-15 minutes now, and max CPU temp was 65-67 (that's core, not the regular BIOS readings), and max GPU temperature = 69 degrees. Not bad, and a lot less than Furmark and OCCT And as far as I know, Crysis is the hardest (or one of the hardest) game on your hardware.
My GPU model is Asus ENGTX560 Ti DCII TOP/2DI/1GD5. What would you say is the "safe temp" at this card? 90 degrees max either in Furmark or OCCT?
you really don't have to test the GPU with such programs. playing games, running videos, any application that use a GPU acceleration will let you know if your card is stable or not. what you should worry about is getting 80-90 while playing games.
Don't forget that the ambient temperature are playing IMP role, if you're providing a good air flow to your case the GPU shouldn't exceed 70s while loading on it.
CPU core temps are not normal, what CPu do you have ? are you overclocking it or not ? are you running with stock cooler ?
That's kinda what I was thinking too... as long as it doesn't hit those temperatures in games etc. Anyway, I like to test it for stability before playing or something. I've figured the OCCT for a good test.
Yeah, I tried running Furmark and OCCT with the case-lid off, and it was only 2 degrees (celsius) difference. So the airflow is great
Core temps not normal? :S You mean they're too high? i7 960 3,2, not overclocked (though it shows as 3,35 or something in OCCT and realtemp) Jup, running with stock cooler, and that's why I don't overclock. The temperatures are pretty high as it is (at least when I stress-test it with Prime95, OCCT etc.). The cores, or at least one of them, goes just over 80 in Prime/OCCT. When I run Furmark/GPU stress-test and Prime95 together, the cores get around 85 celsius.
aha, stock cooler....
i guess all the people go with aftermarket CPU cooler even though they don't overclock. you get 25C core temp at IDLE and 50-55C max when overclocked and in full load. that's why you should screw up the stock cooler and get a decent one.
if the cores get to 85C then then CPU temp must exceed 90s in MOBO's sensor reading, right? 85C per core is pretty dangerous. swap the cooler immediately
No, it's the other way around... the mobo's CPU temp is lower than core temp... about 10 degrees roughly... it depends on idle vs load and temps in general. That's why I was kinda shocked when I found out that my cpu core temp was 82 or something, when it actually showed as max 73 (or around that) in BIOS readings. And I don't even wanna think about what my core temperature were when my BIOS readings were 82 at CPU (before I sat the vcore down at 1.15 - it was on automatic when it got that high). Guess I would've noticed if the CPU was damaged, right? Anyway, it seems that it's a failsafe with Prime95 or something, that I get errors if the temps are too high.
Yeah, I'm looking into coolers as we speak (it wasn't 85 on all cores btw, the other ones are two or three degrees cooler... not that it matters much) Though I can't really understand why the stock coolers are so poor? And this is AFTER I turned down the vcore from auto. If I didn't, I have the feeling it would have melted sooner or later. Well, I bought the PC pre-built, so I can't really say if the cooler is poorly installed by the people from the store, or if it's just plain poor.
yea maybe the push pins aren't probably sit correctly but whatever, that sounds normal with stock coolers but not that much (80-85C) that's why you should get an aftermarket CPU cooler. after market CPU coolers have great heatsinks, better fans with better RPMs which helps a lot cooling down the CPU Unlike the stock coolers with that small fan and useless heatsink.
if you're looking for one, get the Hyper 212+ plus as it's the best for "price per performance" it's great for both overclockers and non-overclockers.
if you were restarting your PC after the Prime test to see the BIOS reading, for sure the CPU temp would drop down some degrees, i monitor BIOS reading with ET6 that works on Gigabyte's MOBOs
No, I'm monitoring real time with Asus Pc Probe II.
I know, CPU temp is high :S But, for now, is it safe to play heavy games as Crysis with 65-67 (I'm writing both numbers because the OCCT CPU monitor shows a couple of higher degree load than Realtemp. Don't know who's more accurate)?
Heh, strange, this is the last Cooler I just checked before reading your post. I'm considering it... cheap, and it sounds good Though I have to take out the motherboard... or do I?