Hi all,
This is my first post here; I figure it was about time I got some additional opinions. I'm relatively inexperienced with overclocking, although, with a Bachelors of Science in Computer Science and a system builder I do have solid ground with most of the theoretical ideas and terminologies.
Anyway, I'm a bit disappointed with what I've been able to do so far with Athlon 64 x2 5000+ Black Edition. My system specs can be found here: http://www.evga.com/community/ModsRigs/comment.asp?sysid=875
You can also see the layout of the cabling I have on the system... which isn't ideal, but, there isn't much of an alternative given the card reader/DVD drives I have taking up all the space in my case. I also have a side 80mm fan on the case cover. The airflow of the case is what you'd expect: intake at the front, exhaust at the rear. The Freezer 64 Pro blows towards the rear exhaust. The only thing that may seem a bit weird is my side fan (which blows just about behind the cpu) is actually being used as intake as opposed to exhaust. The main reason is that the PSU sucks in air from below the unit, so that keeps the air a little more fresh for the PSU.
From everything I've read up on my cpu, it seems I should be able to get it to OC to 3.2 GHz without much of a fuss. 3.2 would be a nice number for me, to keep that nice divider and memory speed at 800 MHz (the memory is running at 2.05v btw).
Currently, I have it OC'd at 2.8 GHz, which is stable and fine with p95, but, I think I'm entitled to more. That said, I think I'm going to have to look into my cooling solution to go any further, because right now things would probably get too hot regardless. Currently my CPU runs between 39-42 C at idle, 42-45 under "normal" load, and eventually makes it way up to 55 C under "prime95" load for quite some time. Note that these temps are with the cover on. So, one of the questions I had would be if people had recommendations on a better cooler for my system.
I realize I could try and lap my heatsink, but, I think I'm going to have to look at something new altogether, since, right now my Muskin memory just barely won't work in the first slot next to the cpu since the memory heatsink would just barely make contact with the freezer 64 fan. It's unfortunate that the dimm slots are so close to the CPU on this evga board. So, question #1 is thus: any good recommendations for a heatsink replacement that will lower temps and work with my board such that it will leave all memory slots available?
The second question relates to the temp monitoring I'm doing while in OS. Right now, with the 590 NVIDIA chipset, the only program that I can use that reports semi-reliable temperatures is the NVIDIA system tools. Unfortunately, while running the program, only the first sample from the CPU is reported accurately (the next sample makes the CPU jump to 23000+ C or some crazy number... I have no idea why... it does this for both XP and Vista). To work around it I have to shut the monitor off then turn it back on, and then quickly check it. Question #2 is thus: Does anybody know of a good temp monitoring program that works with the NVIDIA 590 SLI chipset? (I've tried all the standard stuff, such as TempMon, and unless some recent changes (in the last 5 months) fixed things, I just don't have that reliable of a solution I guess).
Anyway, on to the overclock! I tried to get to 3.2 GHz the simple way with the BE, that is, simply increasing the multiplier and increasing the voltage slightly.
The first thing I should mention is that CPU-Z has always reported what I would consider to be abnormal voltage levels. The x2 5000 should run at 1.35v stock, however, CPU-Z typically reports that the voltage being fed to the CPU is about .1v less than what it's set at (so, when it's set at 1.35v in the BIOS, cpu-z reports around 1.25v, at 1.45v BIOS - 1.35v cpu-z, etc.) When under load the voltage will go up slightly, around .025v or so. The more load the more the voltage will pick up (so, prime95 will make the voltage get up to 1.280v under 1.35v setting). So, I don't know what to make of that.
But, when I increased the multiplier to 16x, and set the voltage to 1.40v, everything booted fine, but, Prime95 immediately failed, giving illegal sumout errors. I deduced that this was due to lack of power being fed to the CPU. Again, cpu-z reported about .1v less than what I had set it at. So, naturally, I thought that I should increase the voltage more.
I tried increasing the voltage in the smallest increments from 1.4v, and each time Prime95 got a little further before erroring out. At 1.475v, Prime95 made it for about 2 minutes before erroring out. But, I knew at this point that my voltage would be much too high for the cooling solution I had.
Finally at about 1.5v things seemed to stabalize with Prime95. But, the temps were much too hot now, it didn't take long for the CPU temp (as reported by NVIDIA system tools anyway) to reach 69 C at which point I promptly shut prime95 off.
Surprisingly, throughout all the testing I never had a system lockup/freeze/crash, nor ever had an issue posting. But, in any regard, the results were disappointing nonetheless.
At this point, I don't know what route I should really take next. I think I want to improve my cooling solution first. But, even if I were to do so, it seems like things aren't what they should be regardless.
So, my final question #3 is: what would you recommend I do at this point? Do you think there could be a potential motherboard/PSU issue going on with the weird under-voltages?
I don't think anything is wrong with the PSU as the thing seems to run everything else absolutely fine, and it's a quality PSU.
I guess I could try messing around with the FSB and try things down that route, but, I was hoping to avoid that messiness in that regard. But, if someone else has an idea that might be somthing I'm missing, well, that's why I'm posting here.
Wow, that got pretty long! Sorry for the long post, thanks for taking the time to read (if you did), and any comments/insights/suggestions of any kind would be most appreciated!
This is my first post here; I figure it was about time I got some additional opinions. I'm relatively inexperienced with overclocking, although, with a Bachelors of Science in Computer Science and a system builder I do have solid ground with most of the theoretical ideas and terminologies.
Anyway, I'm a bit disappointed with what I've been able to do so far with Athlon 64 x2 5000+ Black Edition. My system specs can be found here: http://www.evga.com/community/ModsRigs/comment.asp?sysid=875
You can also see the layout of the cabling I have on the system... which isn't ideal, but, there isn't much of an alternative given the card reader/DVD drives I have taking up all the space in my case. I also have a side 80mm fan on the case cover. The airflow of the case is what you'd expect: intake at the front, exhaust at the rear. The Freezer 64 Pro blows towards the rear exhaust. The only thing that may seem a bit weird is my side fan (which blows just about behind the cpu) is actually being used as intake as opposed to exhaust. The main reason is that the PSU sucks in air from below the unit, so that keeps the air a little more fresh for the PSU.
From everything I've read up on my cpu, it seems I should be able to get it to OC to 3.2 GHz without much of a fuss. 3.2 would be a nice number for me, to keep that nice divider and memory speed at 800 MHz (the memory is running at 2.05v btw).
Currently, I have it OC'd at 2.8 GHz, which is stable and fine with p95, but, I think I'm entitled to more. That said, I think I'm going to have to look into my cooling solution to go any further, because right now things would probably get too hot regardless. Currently my CPU runs between 39-42 C at idle, 42-45 under "normal" load, and eventually makes it way up to 55 C under "prime95" load for quite some time. Note that these temps are with the cover on. So, one of the questions I had would be if people had recommendations on a better cooler for my system.
I realize I could try and lap my heatsink, but, I think I'm going to have to look at something new altogether, since, right now my Muskin memory just barely won't work in the first slot next to the cpu since the memory heatsink would just barely make contact with the freezer 64 fan. It's unfortunate that the dimm slots are so close to the CPU on this evga board. So, question #1 is thus: any good recommendations for a heatsink replacement that will lower temps and work with my board such that it will leave all memory slots available?
The second question relates to the temp monitoring I'm doing while in OS. Right now, with the 590 NVIDIA chipset, the only program that I can use that reports semi-reliable temperatures is the NVIDIA system tools. Unfortunately, while running the program, only the first sample from the CPU is reported accurately (the next sample makes the CPU jump to 23000+ C or some crazy number... I have no idea why... it does this for both XP and Vista). To work around it I have to shut the monitor off then turn it back on, and then quickly check it. Question #2 is thus: Does anybody know of a good temp monitoring program that works with the NVIDIA 590 SLI chipset? (I've tried all the standard stuff, such as TempMon, and unless some recent changes (in the last 5 months) fixed things, I just don't have that reliable of a solution I guess).
Anyway, on to the overclock! I tried to get to 3.2 GHz the simple way with the BE, that is, simply increasing the multiplier and increasing the voltage slightly.
The first thing I should mention is that CPU-Z has always reported what I would consider to be abnormal voltage levels. The x2 5000 should run at 1.35v stock, however, CPU-Z typically reports that the voltage being fed to the CPU is about .1v less than what it's set at (so, when it's set at 1.35v in the BIOS, cpu-z reports around 1.25v, at 1.45v BIOS - 1.35v cpu-z, etc.) When under load the voltage will go up slightly, around .025v or so. The more load the more the voltage will pick up (so, prime95 will make the voltage get up to 1.280v under 1.35v setting). So, I don't know what to make of that.
But, when I increased the multiplier to 16x, and set the voltage to 1.40v, everything booted fine, but, Prime95 immediately failed, giving illegal sumout errors. I deduced that this was due to lack of power being fed to the CPU. Again, cpu-z reported about .1v less than what I had set it at. So, naturally, I thought that I should increase the voltage more.
I tried increasing the voltage in the smallest increments from 1.4v, and each time Prime95 got a little further before erroring out. At 1.475v, Prime95 made it for about 2 minutes before erroring out. But, I knew at this point that my voltage would be much too high for the cooling solution I had.
Finally at about 1.5v things seemed to stabalize with Prime95. But, the temps were much too hot now, it didn't take long for the CPU temp (as reported by NVIDIA system tools anyway) to reach 69 C at which point I promptly shut prime95 off.
Surprisingly, throughout all the testing I never had a system lockup/freeze/crash, nor ever had an issue posting. But, in any regard, the results were disappointing nonetheless.
At this point, I don't know what route I should really take next. I think I want to improve my cooling solution first. But, even if I were to do so, it seems like things aren't what they should be regardless.
So, my final question #3 is: what would you recommend I do at this point? Do you think there could be a potential motherboard/PSU issue going on with the weird under-voltages?
I don't think anything is wrong with the PSU as the thing seems to run everything else absolutely fine, and it's a quality PSU.
I guess I could try messing around with the FSB and try things down that route, but, I was hoping to avoid that messiness in that regard. But, if someone else has an idea that might be somthing I'm missing, well, that's why I'm posting here.
Wow, that got pretty long! Sorry for the long post, thanks for taking the time to read (if you did), and any comments/insights/suggestions of any kind would be most appreciated!