Lol, I love when afterdawn turns random words into links to products/technologies when they're never going to be used in such cases... The first thing MAN springs to mind in my head is Maschinenfabrik Augsburg-Nürnberg. Guess that shows my background...
Oman7, 41-42C for the motherboard temps, is about it for x264 encoding. Normally the motherboard runs 37-38C anyway, but I've never seen it go over 42-43C yet, even stress testing. Russ
Oman7, 1.30v. I dropped it down to 3.7, so if something happens to me, it will be reliable for Russell! He wouldn't know how to change any of those settings. I could probably drop the NB voltage to 1.20v, and I may do that! 87,000+ MIPS will do just fine! Got the HD4670 card installed in it and a fresh install of XP-Pro SP3. it's good to go! Russ
Pretty sure mine is also at 1.30V. Clearly something needs to be done about the heatsink cooling. I'm not into doing that right now. I'll be opening my case again in a month or so though. Perhaps I'll do it then. I don't fancy removing the board though. Perhaps I'll just stick with my original plan, and wait to see how 1090t performs.
Oman7, I dropped mine to 1.20v and the temps are the same. That's exactly why I'm contemplating moving the Rad outside, with more room and access to cool the VRMs, and no heat at all from the Rad, what little get's into the case, if any to begin with. The Rad should cool better because the grille in the case will be removed, so no restriction of airflow caused by the holes in the grille. Russ
hy! my motherboard is MS-7525 with Phoenix-Award BIOS(v0.00pg). I really dont find any way overclocking it. please help me!
Curious, not really an overclocking question, but an underclocking question. Can the Athlon 5200 be underclocked and stable? Because my brother has one in his HTPC, and the darn CPU fan Revs up at times. I'm considering either underclocking and lowering the volts(if possible), or finding a low profile with large fan/heatsink for the CPU. Watercooling is probably not an option for this HTPC. I know of no Radiators that will work in this situation. Perhaps you could educate me though
Oman7, If the motherboard has CPU multiplier adjustment, just lower the CPU Multi. Even though it's a locked multiplier, you can usually lower it! Russ
LOL! Thanks man. The 5200 isn't black, but the gigabyte board it resides in, should be able to unlock some features I prefer not underclock, but I may do it til I can find either a low profile cooler, or water cooler with HTPC radiator in mind. Thanks both of you
Thought I'd reapply thermal past to the 5200. Well, guess what happened! The Damn CPU is stuck to the heatsink LOL! I was attempting to rock the heatsink off the cpu. The whole shebang came out. While the lever was in the locked position. Now I have to use a razor knife to get the cpu off. Hopefully the socket is ok. I didn't force, it simply came out fairly easily :S The pins thankfully did not suffer at all. I think there may have been too much paste before...
Everything is looking good I underclocked the stock 2.6ghz/1.3V/13Multi to 2.2Ghz/1.275/11multi. I then ran Prime95 64bit for 10 minutes. It never exceeded 55C. The cpu fan never exceeded 3400rpms. MUCH more agreeable. I've found if Prime 95 can run for 6 - 10 minutes, it's usually pretty stable. However, in the case of X264, it usually can find instabilities. My brother is not gonna stress this HTPC though. I don't see the need in going any further. The most I've seen his CPU hit is 80%. And that was while watching a Blu-ray. Since the media player he uses, can utilize hardware acceleration (DXVA), The cpu should really never even see that. I think he'll be quite pleased with it for a while. I still insist on increasing its cooling though. I'm very anal about hot CPU's. And 50C is my limit. And given the fairly tight quarters the HTPC resides in, I need to improve it. But at least it'll be fine for a while This is about the size of a grain of rice. I've heard people say that's usually sufficient. What do you think? The scratch was because of an imperfect razor knife. It's a trivial scratch :S I hope...
I had some fun with my brothers HTPC, Now I'm playing with mine. 4Ghz is really close to stable. Closest I've ever had! 3.4Ghz *stock* x264 HD BENCHMARK 4.0 RESULTS Please do NOT compare it with older versions of the benchmark! ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Results for x264.exe r1913 ========================== Pass 1 ------ encoded 1442 frames, 108.63 fps, 3912.32 kb/s encoded 1442 frames, 109.00 fps, 3912.32 kb/s encoded 1442 frames, 109.39 fps, 3912.32 kb/s encoded 1442 frames, 109.52 fps, 3912.32 kb/s Pass 2 ------ encoded 1442 frames, 22.93 fps, 3961.81 kb/s encoded 1442 frames, 22.94 fps, 3961.78 kb/s encoded 1442 frames, 22.95 fps, 3961.32 kb/s encoded 1442 frames, 22.89 fps, 3962.09 kb/s ---------------------------------------------------- 3.7 Ghz(CPU-V 1.35) 1.2V NB 2200Mhz HT/NB -- ~8% performance boost. x264 HD BENCHMARK 4.0 RESULTS Please do NOT compare it with older versions of the benchmark! ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Results for x264.exe r1913 ========================== Pass 1 ------ encoded 1442 frames, 117.46 fps, 3912.32 kb/s encoded 1442 frames, 118.21 fps, 3912.32 kb/s encoded 1442 frames, 118.51 fps, 3912.32 kb/s encoded 1442 frames, 119.42 fps, 3912.32 kb/s Pass 2 ------ encoded 1442 frames, 24.81 fps, 3961.27 kb/s encoded 1442 frames, 24.79 fps, 3961.44 kb/s encoded 1442 frames, 24.96 fps, 3961.86 kb/s encoded 1442 frames, 24.92 fps, 3961.75 kb/s ---------------------------------------------------- 4.0Ghz (CPU-V 1.4V) 1.2V NB 2200Mhz HT/NB -- ~5 - 6% performance boost over previous clock. 12-14% boost over stock clock. x264 HD BENCHMARK 4.0 RESULTS Please do NOT compare it with older versions of the benchmark! ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Results for x264.exe r1913 ========================== Pass 1 ------ encoded 1442 frames, 123.41 fps, 3912.32 kb/s encoded 1442 frames, 124.41 fps, 3912.32 kb/s encoded 1442 frames, 125.25 fps, 3912.32 kb/s encoded 1442 frames, 124.57 fps, 3912.32 kb/s Pass 2 ------ encoded 1442 frames, 26.44 fps, 3961.23 kb/s encoded 1442 frames, 25.58 fps, 3961.79 kb/s encoded 1442 frames, 26.63 fps, 3961.41 kb/s encoded 1442 frames, 26.71 fps, 3960.80 kb/s
Oman7, Don't rock it, press down hard enough to maintain good contact so the CPU won't pop out of the socket, and lightly rotate the coolerleft and right. What kind of paste was on it? I've had that problem with Arctic Silver, but never with Arctic MX-2. Just another good reason to use it. Russ
LOL! It may have been Arctic silver. If so, I replaced it with MX-2. It was the closest one in my desk drawer. In any case, it clearly appeared like I used too much last time. I tried both twisting and rocking lightly. Twisting is probably certainly safer than what happened. When I started rocking it lightly, that's when it gave up. It scared me when I saw no CPU in the socket LOL! I thought for sure the pins would be bent. But they took it like a champ. The CPU slipped right back in there like a glove I got lucky. I'll be running more stability tests tomorrow. If it checks out, I may just go for for 4.2Ghz. If I can stabilize 4.2, I just may survive for 6 - 8 months. On the average, a Blu-ray disc only takes 3 - 6hrs to encode. Depending on several factors of course Although if 4.2 follows the same overclock trend, It probably won't be much of a gain(It might shave 20 min off a blu-ray encode). I think memory bandwidth, and HT/NB frequency may be the next thing to tweak.
I've never seen this happen with Arctic Silver or any of the other pastes I use, but I haven't used Ceramique. That seems like it might be the most likely culprit for it. Not saying because i haven't seen it that it can't happen, of course
X264 crashed. Just increased to 1.425V. We'll see what happens. If it crashes again, I'll bring the HT/NB frequency back to stock, and see what happens. It's too close to stable to need more voltage.