Your overclock appears to be stable.
Your overclock appears to be stable.
Hello! I've been experimenting with overclocking lately and wanted to review my current settings to ensure everything is in order. I also had a discussion about VRM throttling, which has now been resolved by reorganizing components and adding fans. I opted for 6 cores mainly because of temperature concerns and the observation that games don't fully utilize all cores. Using Prime95, I achieved around 62°C without any errors—roughly 50°C in-game. At idle, temperatures hover around 40°C if I prevent frequency drops, but unless I'm gaming, I let it run hotter to maintain performance. This is during the summer heat. Here are my current specs and notes. Feel free to test while you can. 😛
-Antec 1100 case with fans installed
-Asrock 970 Extreme3 (learned it's just 4+1 phases, but I managed to get a fan handling VRM heat and surrounding area, which stopped throttling)
-FX-8150 paired with H55 cooler (installed in the upper slot behind the HDD trays to free up VRMs. Dual cougar 1.6mmh2o fans running push/pull. Upgraded to SP120 performance fans soon)
-4x4GB Corsair Vengeance 1600MHz tall
-Gigabyte 7870 OC Edition
-1TB Caviar Blue HDD
-Corsair TX750
Current overclock details...
CPU frequency: 200MHz
PCIe frequency: 100MHz
Spread spectrum: off
Turbo core: off
AMD APM: off
X21 4.2GHz @ 1.375V on 6 cores
NB frequency: x12 2400MHz
CPU voltage: 1.2V
HT bus speed: 2400MHz
HT bus width: 16-bit
RAM clocked up to 1866MHz (XMP profile)
9-10-9-27 load line calibration: auto
DRAM: 1.5V
NB: 1.145V
HT: 1.21V
CPU VDDA: 2.56V
PCIe VDDA: 1.81V
Cool+quiet: enabled
C1E: enabled
Secure virtual machine: enabled
C6: disabled
CPU thermal throttle: enabled
GPU also locked at 1250MHz core and 1375MHz memory @ 12.6V (still testing for stability, but it hasn't crashed yet with these settings after a few runs).
What are your thoughts? Are there any aspects that catch your eye that you should adjust or could improve? With these temperatures, should I push the CPU further to extract more performance while staying under 70 cores/60 threads without throttling again? I'm still limited by the CPU, but I know the H55 isn't a strong performer.
I never used cpu-z before, but this is actually what I was going to ask about. I was wondering if there was a way to test if it actually makes a different with higher clocks and how much so I can weigh whether the temps are even worth it lol Ive been messing around with clocks a bit and am currently at 4.3ghz 6 cores @ 1.385v. Temps get to 66*C socket and 55*C on prime with that small increase. I tried 4.4ghz, but it wanted 4.25v and when temps got to 70 I bailed out. Anyways, at 4.3ghz I seem to do around 946 single thread and 4460 multi thread. Kinda cool that I can see a reference vs an average stock clocked 8150. Not sure how well I stand against others with this kind of clock. I'm quite a ways behind you but my mobo would melt if I even tried to clock that high
😛
What is vdda and what does it do? Lol I never touched that because I was unsure what it does.
I only have the choice for auto, 2.56 and 2.7ish, anyway I think it works fine.
My score is on 6 cores and when I set it to my reference against an 8150 I notice the multithreaded score is significantly lower. However, I thought that difference comes from the 6-core versus 8-core comparison, and the overclock has improved my single-thread performance by about 130 compared to a stock 8-core setup.
My vcore does experience a slight drop when under load, though I believed it was meant to safeguard the CPU. I thought lower drops would help idle temperatures by preventing the CPU from sitting at higher voltages before the droop occurs. However, I’m unsure how this impacts performance when the system is under load since the voltage needed for the clock remains essential.
I’ll need to examine it more closely. It appears it’s better to leave it unchanged even if idle voltage is slightly higher. From what I know, the vdroop exists for a reason; unless operating at the higher idle voltage causes instability, it seems unnecessary.
For my overclocking attempts, people usually either use a multiplier or increase the FSB speed... The FSB speed is indeed the clock speed shown on my motherboard (currently 200). I also understand that this impacts other components like RAM and possibly the CPU itself. I’ve read that some users adjust the FSB, multiplier, or a middle ground based on their chip and setup. Maybe I should experiment slightly with the FSB before trying to boost the multiplier further. It would be best to keep everything at stock settings except for power-saving features, then gradually raise the FSB until stability problems appear. This way, I can determine the maximum before increasing the multiplier more.
I’m considering testing small FSB adjustments as a starting point and see if I can improve performance without overheating too much.
Also, I bought an H105 board which should arrive soon. Even if I switch chips, I can use any available brackets. I prefer the larger size for better airflow. With the H55, the fan push/pull settings are causing some noise against the card’s trim. Instead of risking that issue, I chose a thicker board and placed it in the bottom slot with only the top fan pushing, which should avoid interference if I ever need to move the case. I’m eager to see how this affects performance once the new board arrives. Unfortunately, installing it here will also increase heat in the system, but there’s still good airflow overall.
I have a query regarding overclocking with FSB speed versus adjusting multipliers. Some users report improved outcomes with one method over the other, so I decided to give it a try myself. I understand this impacts NB, RAM, and similar components. However, does it influence my graphics cards? Up until now, I've reached the point where increasing voltages is necessary. I pass basic tests when voltage is sufficient, but I notice minor flickering in core temperature and other programs occasionally. It's not severe or clearly noticeable, just occasional brief glitches. Could this be affecting the graphics card, or perhaps RAM? I'm not very familiar with RAM behavior beyond crashes, and I might be overreacting, but I want to confirm if it does impact the video card after I previously pushed it to its stable limits.