View Full Version : 1GHz @ 1160, unstable @ 1.2GHz
ReNeG@dE
02-09-01, 10:51 AM
I've got it at 8 x 145 at 1.75V and it (so far) has been completely stable, I tried using cpu plus to get the processor to 150, it would boot into windows, was okay for about 3-5 minutes and then locked up. My temps were ~36-38C.
Then I tried bumping EVERYTHING to 150, booted up fine, but when I ran Prime95 it would give me FPU errors and 3D Mark would just close randomly. So I upped the voltage to 1.775 and tried again, was good for about 3-5 minutes, could run 3D mark and Prime95 but then it locked again.
I upped the voltage to 1.80 and tried again, this time it lasted about 10 minutes and locked up. The heat was still showing a range of 35-39C.
My room temp is about 21-22 and the case temp is the same. Now I bent the cpu temp probe up, but forgot to put some Arctic Silver on it so I'm not sure if the temps are acurate or not. Touching the heatsink, it doesn't feel very hot though... Is it temp? voltage? FSB is just too high? I haven't tried adjusting VIO at all yet...?? Thanks.
ReNeG@dE
02-09-01, 11:42 AM
w00t I just got my second star! :)
something clever
02-09-01, 12:48 PM
Hey Reneg@de,
well I don't have any bright ideas, but believe we have similar machines and could compare notes.
I just built a KT7A RAID with 1 GHz TBird a couple of weeks ago - still tweaking it. last couple of days been running at 8 X 150 with no problems whatsoever - not even a hiccup. my temps range about 36C to 44C (Alpha PAL 6035).
I HAVE NOT HAD TO INCREASE VOLTAGE - still at default! Today Ed, or Joe, or whoever has reported some users of new TBirds are having better luck without increasing voltage!?!?
my board and processor have been very cooperative, I am really surprised your's are giving you ****.
are you running cheap-ass memory?
are you running your memory out of sync with PCI bus? (supported, but not the best way to go I've read)
I removed the HS+F off the Northbridge and found a poor application of thermal grease - fixed that.
I don't know. If you really don't have a temp problem, and default voltage doesn't help - then I would have a good look at the SDRAM, and it's settings.
ReNeG@dE
02-09-01, 01:13 PM
The memory is Crucial (Micron) PC133 CAS 2 (256MB) I have it in bank 1, but I read somewhere you should start in 3... I have all the banks set to turbo, with 4-way, 4k, and latency set to 2.
what do you mean exactly, by memory out of sync with PCI, I had the FSB/PCI set to 150/37 (or whatever it is) and the FSB Plus set to 0.
I also found a ****y thermal paste app on my kt7 raid after fixing that and slowing my mushkin rev 3 down a little I can now run at 114 FSB
are you using a windows temp monitor?
I would also suggest a burn in (so does amd 48 hours) when I first assembled this system I wasn't able to reach 1.2 after the burn in I was able to push the old girl to 1233
I was also going to suggest you bump the voltage but after reading something cleavers post ill hold off on my opinion ;D
oh ya one other thing with the added pressure between the thermistor and the chip the temps are a little more accurate but not as good as they could be with the grease
ReNeG@dE
02-09-01, 03:16 PM
so just pop the pins off the back of the motherboard and the HSF for the northbridge should come right off?
Yes I'm using the VIA monitoring program as I don't really like MBM very much...
I have SETI and Prime95, I've read about CPUBurn but was unable to find it, does anyone know where to get it? Or is there another program you'd recommend for doing the burn in?
I'm a little hesitant myself with upping the voltage on stuff... I don't think Im ever going to do the voltage mod on my motherboard...
I've got a question for you Kat, I forgot the thermal paste on the temp probe (though I did bend it up), you think it makes a big difference? Also theres the fact that I bumped up the voltage two notches and there was almost no temp change.... another thing is that after running SETI for an hour, the temp goes from 35 idle, up to 38-39 and then after a little while drops back to 36-37.... As far as I can tell room temp is not changing...?
Thanks guys I really appreciate the help :)
something clever
02-09-01, 03:25 PM
I said PCI, sorry. I meant FSB. memory should run in sync with the FSB.
yea, we have the exact same memory. mine in bank 1 also. looks like same settings too - but I'm not up to turbo yet.
I'm running BIOS version WW. I seen in a FAQ somewhere that WW is a little screwed up when it comes to selecting the SDRAM speeds. for example: neither mine, nor the author of the FAQ's board would POST when "normal" is selected. furthermore, the author benchmarked each of the remaining settings and found that the options that you would expect to rock (i.e., turbo) actually performed worse than the "slower" settings. I haven't had the time to check this out myself yet - so I'm is still set at 8/10. but, at the same time, somewhere else there is a setting for CAS 2/3 (I'm at 2 here). both seem to be setting the same thing to me?
the KT7A is an asynchronous board. this means that, unlike other boards, your memory does not have to run at the same speed as the FSB. there is an option somewhere (I'm at work - can't look for it now) that lets you tweak this. it enables you to run your PC133 SDRAM at full speed, even though the the TBird runs at 100. the "FSB Plus" may or may not be it. I think that option just jacks up the processor speed (more uncertainty here: ABIT manual implies this increase affects processor only. but have read at a few places elsewhere that it takes the FSB up with it).
so maybe have a thorough look through everything to see if the memory speed and FSB (processor included) are all in sync.
did you try the stock voltages?
I run rc5 and when the pc is up loading the temp drops then will go back up again I've never look at seti maybe theres a similarity ?? when you have seti running this would be a full load temp
as to the accuracy of the temp the next time you take your chip out put the grease on it until then add 5% to your temp tbirds will run normally well into the low 60s but oc much better in the low 40s don't worry about temp until you start to approach 50 c
1.85 will not damage your chip !!! ;D
i forgot about the "burn in" this is not a program what it intail`s is this
run your pc at its highest stable speed at full load (seti rc5 folding @home) for 24hour`s and then bump and do over ;D
ReNeG@dE
02-09-01, 09:20 PM
Okay, well I upped the multiplier from 8.5 to 9 and put the FSB back to 140, this works fine, I haven't done any extensive testing, but 20 minutes of rc5 with no problems @ 39C. I then upped the FSB to 145, which worked with the 8.5 multiplier before. It boots up, but my USB mouse won't work. I tried lowering the memory to CAS 3 and upping the voltage to 1.8, same problem. I drop it back to 9 x 140 and it works again.... I've never read about mouse/USB problems when overclocking... any ideas?
ReNeG@dE
02-10-01, 09:40 AM
well I just ran RC5 for 12 hours (its still going now, but I have to do other stuff you know :) ) I'm going to try a few different things now to try and hit 1.3GHZ but I have a feeling the mouse is going to keep screwing up... If you any ideas at all I'd really appreciate it.
EDIT: Oh yeah, I tried dropping it back to 8.5 x 145 and it works fine still, so it doesn't appear to be a problem with that FSB
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.