-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
8150 OC (yeah i bought it)
- Thread starter Seth8530
- Start date
- Joined
- Mar 12, 2002
I have an 8120 on that same board. If your not already on it you need to manually download the F10-g beta bios for your board. It will allow you to kill the power managment throttling aswell as HW thermal throttling completely. The bios on the gigabyte is a bit deceptive about the vcore. So basically I have mine set to 1.36v in the bios but it idles at 1.30 and will push up to ~1.37 under heavy load.
Temps on AIR with these chips are going to be a tough one. Ive got so much raddage in my loop that the only thing that improves CPU temps is pump speed. There just comes a point where removing 192w+ of heat from the CPU wont allow for low deltas.
Temps on AIR with these chips are going to be a tough one. Ive got so much raddage in my loop that the only thing that improves CPU temps is pump speed. There just comes a point where removing 192w+ of heat from the CPU wont allow for low deltas.
Last edited:
- Thread Starter
- #23
Yeah i reckon so, I originally tried that BIOS version, but i found it a little unstable. So i backed away from the BETA BIOS. Im not getting throttled yet, so im not gona worry about that anyways. you think that mounting new fans on the heat sink might help?
- Thread Starter
- #25
There are some yes, but i got them set to full speed.
- Joined
- Oct 17, 2011
- Location
- Orange County, CA
Some of my experience:
I needed 1.41V for 4.5GHz. It's 100% stable under everything I've run (including 10-12 hours under each of the usual torture tests) with a NH-D14 with beefed up fans but throttles slightly after a couple of minutes when I run Prime95, but it never does that under normal loads. So that's the hairy edge for my CPU on air.
Through much reading I found that 99% of the info you can find on AMD CPU temps, even from AMD people, is wrong, because everybody confuses core temps with socket (CPU case) temps. See e.g. my posts http://www.overclock.net/t/1139726/amd-fx-bulldozer-owners-club/2500_100#post_16089198 and http://www.overclock.net/t/1134229/amd-cpus-max-temps/100_100#post_16216984.
The maximum core temp (rather, Tctl since "core temperature" is an Intel term) is 70 for all AMD CPUs. That is because AMD scales the value that is reported such that the maximum operating temp is always "70" (*not* C, it's an arbitrary scale.) That corresponds to a CPU case temp (i.e. motherboard socket reading) to 61 or 62C.
According to AMD's data sheets, a Tctl of 70 means "maximum cooling please" and is a perfectly acceptable operating temperature. So you're wasting a lot of OC'ing headroom if you limit yourself to 55 like many people recommend.
I needed 1.41V for 4.5GHz. It's 100% stable under everything I've run (including 10-12 hours under each of the usual torture tests) with a NH-D14 with beefed up fans but throttles slightly after a couple of minutes when I run Prime95, but it never does that under normal loads. So that's the hairy edge for my CPU on air.
Through much reading I found that 99% of the info you can find on AMD CPU temps, even from AMD people, is wrong, because everybody confuses core temps with socket (CPU case) temps. See e.g. my posts http://www.overclock.net/t/1139726/amd-fx-bulldozer-owners-club/2500_100#post_16089198 and http://www.overclock.net/t/1134229/amd-cpus-max-temps/100_100#post_16216984.
The maximum core temp (rather, Tctl since "core temperature" is an Intel term) is 70 for all AMD CPUs. That is because AMD scales the value that is reported such that the maximum operating temp is always "70" (*not* C, it's an arbitrary scale.) That corresponds to a CPU case temp (i.e. motherboard socket reading) to 61 or 62C.
According to AMD's data sheets, a Tctl of 70 means "maximum cooling please" and is a perfectly acceptable operating temperature. So you're wasting a lot of OC'ing headroom if you limit yourself to 55 like many people recommend.
Hang on.... IntelBurn (high) got my socket to 70c and Cores to 52c and it was completely stable, and no throttling.
Prime95- 9 hours, Socket 66c, Cores 49c, again completely stable.
The difference between my socket and cores is as much as 18c at the most extreme end
when playing BF3 my core never get over 40c, and socket never over 50c
When AMD talk about 62c i think they are referring to the maximum core temps, if it is the socket my CPU should have failed.... overclockers cite the core temps of 55c as maximum stable, not maximum safe tolerance.
I have no idea whats going on with socket temps as they vary massively, with mine, as i said the difference is as much as 18c more, i have seen other Phenom II owners socket temp readings less at the socket then the cores.
Now, and up to a point i tend to ignore socket temps as they are so wildly different from one to the next i feel its probably the sensor on the MOBO's that should not be trusted, the core temp is from the sensor on the CPU's die.
At 4.3Ghz the cores get to 55c pretty sharpish, at which point i stop the workers, the socket balloons way past 70c but there are no signs of it getting hurt, the Cores are not error'ing out...
Prime95- 9 hours, Socket 66c, Cores 49c, again completely stable.
The difference between my socket and cores is as much as 18c at the most extreme end
when playing BF3 my core never get over 40c, and socket never over 50c
When AMD talk about 62c i think they are referring to the maximum core temps, if it is the socket my CPU should have failed.... overclockers cite the core temps of 55c as maximum stable, not maximum safe tolerance.
I have no idea whats going on with socket temps as they vary massively, with mine, as i said the difference is as much as 18c more, i have seen other Phenom II owners socket temp readings less at the socket then the cores.
Now, and up to a point i tend to ignore socket temps as they are so wildly different from one to the next i feel its probably the sensor on the MOBO's that should not be trusted, the core temp is from the sensor on the CPU's die.
At 4.3Ghz the cores get to 55c pretty sharpish, at which point i stop the workers, the socket balloons way past 70c but there are no signs of it getting hurt, the Cores are not error'ing out...
Last edited:
- Joined
- Oct 17, 2011
- Location
- Orange County, CA
I don't know if that version of HW monitor is reporting the raw Tctl values or is trying to "correct" them. I know it "corrects" Intel's values.
I use AOD, it reports the raw numbers.
- Joined
- Oct 17, 2011
- Location
- Orange County, CA
My 8150 has run for probably close to a week total at Tctl=70 to 72, 100% stable. At 73 or so it starts to throttle and is still stable.
- Joined
- Oct 17, 2011
- Location
- Orange County, CA
The Core temps do match up
Yeah in that case I agree your mobo CPU sensor reports bogus data.
- Joined
- Jul 9, 2011
so throttling is normal?
When the CPU gets to hot it throttles back to protect its self from over heating, yes its perfectly normal unless its doing this at temperatures that are not high.
I have found that it throttles the multi down to 16.5 no lower. So in order to bypass this set your multi to 16.5 and up external clock or HT for un-throttled CPU. Be warned you are circumventing the built in safety feature. But done right and keeping an eye on temps it should be ok. Only do this if you know your processors limits and are comfortable doing it.
From my findings I think it might be the MB throttling as soon as 62c is hit on the socket it throttles upto 5 cores down to 16.5. But if your multi is already set to16.5 it doesn't throttle.
Hope this helps
Sal.
From my findings I think it might be the MB throttling as soon as 62c is hit on the socket it throttles upto 5 cores down to 16.5. But if your multi is already set to16.5 it doesn't throttle.
Hope this helps
Sal.
Many boards have option to turn off throttling. There are 2-3 options like APM or most with power saving.
I can only say that cpu will turn off when you pass 100*C even if you turn off cpu overheat protection and throttling. At least that happened to me on 990FXA-UD5 and CHV ... and cpu is fine.
I can only say that cpu will turn off when you pass 100*C even if you turn off cpu overheat protection and throttling. At least that happened to me on 990FXA-UD5 and CHV ... and cpu is fine.
- Joined
- Oct 17, 2011
- Location
- Orange County, CA
Yup. AMD says Tcrit is 90, i.e. when Tctl (core temp) goes over 90 on its fake scale the CPU may turn itself off.
I have APM off /C6 powerdown Off/ C1 off/ EUP off/ C&Q off/ CPU OCP off/ CPU NB OCP off/ and it still throttles when socket temp reaches 62c exactly.. this is the MB sensor throttling thinking there is still a 1090t in the socket meh thinks..
Your a brave lad letting it run to 100c not sure I would even do that and im pretty reckless with my kit.. but im not sure I would run that hot with an intel never mind amd.
so now I just set multi to 16.5 and oc HT no throttling at all no matter what temp I reach...dangerous but fun
- Joined
- Jul 9, 2011
Similar threads
