Overclocking an amd 1100t

paradoxeternal

Distinguished
Oct 30, 2011
219
0
18,680
hi, i'm currently trying to overclock my AMD Phenom IIx6 1100t. my motherboard is a gigabyte 990fxa-ud5, my RAM kingston hyperx 12 GB at 1600mhz and my cooler the coolermaster hyper 212 evo. except for my 650 watt PSU, i'm pretty sure i've got great overclocking components.

I am pretty much a virgin when it comes to overclocking but i did my research beforehand--read the black edition AMD overclocking guide here on tom's hardware, read through threads on a few sites to see how things went for other people and looked up things like the specs for my RAM, voltages for my CPU and what certain terms and such meant.

i followed the guide and got my 1100t to my goal of 4ghz booted into the windows desktop. after running 3D mark 11 and getting it to not crash through that, i started the fine-tuning process with prime 95. here's where my problems start.

the voltage for my CPU is at 1.5v. i don't know if this is abnormally high, but the normal that my motherboard had it set to was 1.4v---just outside the range that AMD has my CPU's voltage set for. is the 1.5 voltage too high? from what i've seen on forums, most people have it somewhere in the 1.4 range, but to get my system to boot into windows at 4000 mhz i needed the 1.5 volts. is it too high? could it damage my CPU or other components?

my next question is with the RAM. my motherboard had it auto set to 11-11-11-29 at 1.5v, while kingston's site said it had been tested at 9-9-9-27 at 1.65v. this being the closest i could find to the "manufacturer's requirements" as stated in the overclocking guide, it's what i set it to. i wasn't sure why setting the timings slower and the voltage higher would be good, but it's what kingston said. is the voltage for my RAM too high? do the voltage for the RAM and the CPU have to be the same? i either drop cores or crash the prime95 blend test after about 3 minutes, could this be the problem? or would it have anything to do with the timings?

my last question has to do with the northbridge/hypertransport. my system crashed the small fft test (which i've heard tests the northbridge) so i tried bumping up the northbridge speed and voltage. i only bumped them up 1 increment, and did the same for the hypertransport since that's what my motherboard told me to do in the BIOS. could this be the problem and not my RAM? I figured all i had to do was play with the northbridge/hypertransport settings a bit but i don't know what levels are too high in terms of speed and voltage. would anybody be so kind as to pass on that knowledge as well? google searches haven't yielded much.

so, basically, are all three things (cpu voltage too high, ram odd timings/high voltage and northbridge questions) problems? Or would it seem that i am OK so far?

thank you for your time.
 
Solution
Ram with 9-9-9-27 is faster than 11-11-11-29, tighter timings equals reduced latency, improved performance, and the need for more voltage. Ram voltage and CPU voltage don't need to match, in fact if they did, most processors would be fried since Ram tends to run at higher voltages.

Most forums and overclockers consider 1.5V to be the max safe voltage for Phenom II, and to only use that much voltage if you have good cooling, so you should be fine there but I wouldn't increase it any further. Personally I run a 1090T @ 3.8Ghz using 1.4V, still testing stability to get the voltage lower. I need 1.5V for 4.0Ghz and it increases power consumption by over 40W to get that last 200mhz, and outside of a slight difference in a benchmark...

loneninja

Distinguished
Ram with 9-9-9-27 is faster than 11-11-11-29, tighter timings equals reduced latency, improved performance, and the need for more voltage. Ram voltage and CPU voltage don't need to match, in fact if they did, most processors would be fried since Ram tends to run at higher voltages.

Most forums and overclockers consider 1.5V to be the max safe voltage for Phenom II, and to only use that much voltage if you have good cooling, so you should be fine there but I wouldn't increase it any further. Personally I run a 1090T @ 3.8Ghz using 1.4V, still testing stability to get the voltage lower. I need 1.5V for 4.0Ghz and it increases power consumption by over 40W to get that last 200mhz, and outside of a slight difference in a benchmark, performance difference isn't noticeable.

I personally have little experience with tweaking northbridge voltages/speeds so I cannot help there.
 
Solution

paradoxeternal

Distinguished
Oct 30, 2011
219
0
18,680



thanks for the RAM info tho!
 

paradoxeternal

Distinguished
Oct 30, 2011
219
0
18,680


actually, i've run into a problem with my RAM. i've managed to get NB/HT speed and voltage stable, as i ran the prime95 small ffts test for a little over 6 hours fine. however when i ran the blend test, i dropped cores within minutes and minutes later the computer crashed. i've lowered the speed down to 1066 (as 1333 was also unstable) and have been fine with the blend test for about half an hour now. originally, i just tried to tighten the timings as the guide said to do, so i lowered them to 8-8-8-26, but the motherboard told me the overclocking failed so i had to stop that.

when tightning timings, is there a specific way one is supposed to go about it? also is 1066 too slow? I feel as if it is. i tried taking advantage of the easy over clocking feature in the BIOS but the motherboard told me that the system voltages were not optimized. i'm not sure what this meant.