• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

AMD x6 T1100 @ 54C?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Yomama

Senior Member
Joined
Dec 18, 2000
Location
The battle born State
I just replaced my unlocked x2, which was running with all 4 cores at 3.8 GHz stable with one of the Tiger Direct T1100s. Currently I am using the stock heat sink (albeit with AS instead of the heat pad), which is way nicer (copper heatpipe model) than the one that came with the X2 and it is not overclocked and turbo is turned off. I am running rosetta so it is pretty much at full load but the temps are @ 54C - about 8C higher than my X2 with four cores at the same clock speed - I briefly o/ced it (accidentally actually) to 3.9 but temps shot up into the low 60s before I backed off. I have a better heatsink in the Corsair A70 and will probably mount that over the weekend, but I am just curious if 54C in a 65F room is a about normal or if this chip is runing too hot?

Thank you for your insights
 
That doesn't seems bad to me for the stock heatsink at full load but you may be able to reduce it until you add the A70.

Is your BIOS pushing up the vCore more than it needs to be or did you manually set it to 1.35?
You might want to consider setting the cpuNB voltage to 1.15v while you're there, too. :)
 
If the 870 is anything like my 890GX, there's no way to force a CPU voltage. Even disabling LLC allows it to creep around a bit. =(
 
Thank you for the responses. So far this processor is disappointing - got it to 3.8 without raising voltage, had to bring NB up to 1.2 though - but temps were in the low to mid 50Cs under 100% load (rosetta) even with corsair A70 and fans at 2000rpm sucking 65F air in (pushing out was even worse). I am wondering if there may be something wrong with this chip since none of the hardware monitoring utilities like CPUID HW Monitor, open HW Monitor or core temp can read the temps of the 6 cores
 
Thank you for the responses. So far this processor is disappointing - got it to 3.8 without raising voltage, had to bring NB up to 1.2 though - but temps were in the low to mid 50Cs under 100% load (rosetta) even with corsair A70 and fans at 2000rpm sucking 65F air in (pushing out was even worse). I am wondering if there may be something wrong with this chip since none of the hardware monitoring utilities like CPUID HW Monitor, open HW Monitor or core temp can read the temps of the 6 cores
Man Im building my 1st AMD build with this processor already ordered it.Dont tell me its disappointing lol lol please :)
 
- got it to 3.8 without raising voltage, had to bring NB up to 1.2 though - but temps were in the low to mid 50Cs under 100% load (rosetta) even with corsair A70 and fans at 2000rpm sucking 65F air in (pushing out was even worse). I am wondering if there may be something wrong with this chip since none of the hardware monitoring utilities like CPUID HW Monitor, open HW Monitor or core temp can read the temps of the 6 cores
Most X6's will run 4.0 GHz without too much voltage or extra cooling but all overclocking is luck of the draw. What are your other settings like RAM speed/timings/voltage and cpuNB speed/voltage? Have you increased the HT Link speed above 2000 MHz?

You mean CoreTemp doesn't show a temp at all (other than Tj.Max)? That's unusual. Do you have ACC or core unlocker or anything like that turned on in BIOS? That's the only thing I know of that stops the core temp from being reported (or reported erroneously).

If you're saying CoreTemp only reports one temp that's as it should be. Some of the other programs you mentioned report "fake" temps of six cores but in reality there is only one temp sensor on-die, which CoreTemp should report.
 
Quietice - thank you again for taking the time. Obviously this is not my first overclock :), and I had a good run with the system in my signature, which I now transplanted onto a cheap biostar mb for secondary duties (mainly rosetta), and it runs at 3.8GHz even with stock cooling. This T1100 is a different animal, I guess. I turned of the turbo, but all the standard settings (core unlock...) are on. The ram is running at 1680, HT frequency is 2100 so not a crazy step up, and I don't touch any settings related to it (oc via multiplier only for now), the issue really is the heat being emitted by this chip. I am getting info on one sensor, but none of the cpu sensors are showing anything but 0C - I seem to remember people posting individual core temps via the CPUID HWMonitor.

For now I have taken a step back - running the chip at 3.6GHz with 1.325V core and 1.15V NB - dropped temps to 46/47 with the H70 - this still seem excessive to me - I mean this is an entry level water cooler, but only a 300kHz overclock and it still is almost at 50C. I may dissassemble over the weekend and see if the preapplied heatsink compound is possibly a culprit.

Any other suggestions are welcome

Thank you
 
You need to turn core unlock, ACC, and anything else like it off. Those things only apply to X2's and X3's that have extra, non-functioning cores to unlock. Both X4's and X6's are what they are, no extra cores disabled at the factory.

HWMonitor and several other programs (NOT CoreTemp) show temps for the same number of cores as the processor has. Essentially, these programs are lying to you since there is only ONE sensor on any Phenom CPU regardless of the number of cores.

I suspect, but can't be sure of course, that core unlock/ACC is part of the problem. The other part may be your heatsink. At near 1.35 vCore with reasonable cpuNB voltage like you have the load core temp shouldn't come anywhere near 50° with that cooler assuming 20-22°C ambient and good case airflow so there's no vacuum load on the fan(s).
 
ACC = Advanced Clock Calibration

Not all boards will have it labeled that way (most do) and some may not have it at all - at least not as something adjustable in BIOS.


PS
I just checked one of the ASRock manuals and it's called Active Core Calibration in there - go figure. I suppose they saved a few letters on the label that way :-/ or maybe AMD changed it with the 8-series chipsets and I missed the memo.
 
acc has an equivalent in asrock - I tured it off to no avail - pulled off the heatsink - cleaned and applied AS3 - re-installed - no change - now running stock 3.3 - no o/c at all and it's around 50c - indoor temps went up from 17c to 21 - stil this is BS - anyone have an idea how to return a chip to Tiger or is it better to address AMD directly?
 
The chip works correctly right? If that is the case you do not have a legitimate reason to return the chip due to it not overclocking well. :-/ If the 50 degrees is load temps, you have room to work with a bit.
 
QE - thank you for your continued great advice - _ i finally had time to look into the details of this oc and it turns out that the asrock utility I was using was relying on a sensor on the mb which reported 50+c temps. now - i used the same mb/tool to sucessfully unlock and o/c my be 555 to 3.8 under load and temps in the 50s indicated it was too hot, albeit they were not the core temps. via cpuid Hardware Monitor I finally found out that the core temps - at least for the hexacore are quite a bit lower - anyhow this chip is not budging much, it likes 3.8 at 1.4 with nb @1.175 BUT anything above - even with pushing nb to 1.25 is a gamble - mind you this is a w/c set-up - stock heat sink this thing would likely fry even at stock set-up
 
WC can throw the CPU socket to core temp difference off a lot. It's a matter of airflow. With air cooling you have a fan above the socket sensor stirring up the air, with WC you don't so the heat builds up in that area giving you bad readings. Stick with the load core temp (you never did say what you had for that at your current 3.8 GHz with 1.4 vCore).
 
I played with the T1100 and eventually settled for 3.9 GHz.

I am running this CPU with a modest Biostar TA870U3+ motherboard.

Setting the multiplier to 19.5 and boosting the voltage of the CPU to 1.4, the NB to 1.2 and setting the HT link to 2.4 GHz gave a nice and stable setup with the core temps maxing out at 52c (using a H60 cooler).

In trying to go to 4.0 GHz I ran into constant problems, I really had to boost the voltage to get stability but then the temps would go over 56c and while the AMD overdrive stability tests would run fine my system chocked using Vidcoder.

It seems there is simply no way to get this to run at 4.0 stable.
Suggestions are always welcome.
 
I have the same problem and everything I could find was due to my mother board not going higher then 2400 mhz for HT and NB.
Hmm, I think I can put that one higher on that MB, I thought 2.4 was ideal for 1066 memory so I did not do any tinkering with that.

Something to try. :)

Edited to add:

Tried it but it did not make any difference lowering or raising gave the same results.

Looks like 3.9 is the upper limit.
A shame because it is certainly not temperature bound.
Last night I ran another stress test with a slight lower voltage than 1.4 and that worked as well and with lower temperatures.

4.0 boots and windows works fine even AMD overdrive stress tests work as well however video conversion bombs out with the blue screen of death (usually within a minute).

Any suggestions?
 
Last edited:
Hi,
You might want to have a look here:

http://www.overclockers.com/forums/showthread.php?t=596023

This is Dolk's Guide to OC'ing Phenom II's. There's a section on Thuban chips.
Lots of information. Take note of his warnings about the Heat Wall. Thubans are pretty sensitive to heat--he says 55C is an absolute limit, I think.

Still, with WC I wonder at your temps. It's ~26C in the room right now, and my 1090T is running at 3.4GHz, 1.31 Vcore, 51C CPU temp in CPUID HW Mon, 46C core. This is on air with a CoolerMaster 212....in a warm room. I would expect better from water in a cooler environment. Any chance you could get AS5 on your CPU? That's what I'm using. I don't know how much difference it would make.

Just random thoughts here. I hope you get it worked out.

EDIT: I left out an important detail. I'm running Prime 95 on all six cores. In US terms, it's about 83F on the floor down by the computer, CPU is 123F, Core is 114F. (I'm in Chicago, but I've tried to get in the metric habit on forums.) It's hot here right now. I'm in the A/C part of the apartment. But my point is that I've run this chip harder, at higher ambient temps, and not exceeded that 55C limit with an air setup. You ought to be getting better from what you have. I think heat is your issue, not voltage. You might even be able to back Vcore off some if you took the heat off.
 
Last edited:
Back