Got my 170! - Now oc to 2700 - tips?

[aelfwyne]

It's CCBBE 0615EPMW stepping. Yay!

Supposedly a good core for overclocking. Will be testing it out soon on my Asus A8V Deluxe, and have a new ASRock motherboard coming next week.

Tips for this (or the new) board and this chip together?

Will be using for now air cooling, probably with the stock cooler and my Thermaltake fan (from the Venus 12)...

EDIT: On second thought... after seeing the stock fan, it is form fitted to the cooler. Doesn't look like my Thermaltake fan would fit on very well... Hrmph.

 

[Entropism]

Mine was delivered today, installing tonight, on an A8NSLI deluxe. The board got to 310 HTT stable, hoping to hit at least 2.8Ghz with air.

 

[aelfwyne]

Well so far, I'm at 2500 and haven't finished tweaking. My board (2-year old Asus A8V rev 2.0) will hit 290 HTT stable. On stock voltage, no boot at 2.9, boot at 2.5... Going to work my way up and find the max, then work on the memory....

 

[Slug]

aelfwyne,

What are your settings for a stable 290 HTT with that board? I can't get mine past about 255.

 

[LittleNemoNES]

my other rig has an opty 170 @ 2.75ghz @ 1.4v 38 idle and 50 load.

good luck!

 

[aka1nas]

Originally posted by: Slug
aelfwyne,

What are your settings for a stable 290 HTT with that board? I can't get mine past about 255.

I don't own that board, but when it initially came out I think people were saying that the board had ranges of HTT values that were unstable even though higher values would work fine. Maybe someone with more experience with that board can give more detailed information.

 

[aelfwyne]

Originally posted by: Slug
aelfwyne,

What are your settings for a stable 290 HTT with that board? I can't get mine past about 255.

Well, I'm following the step-by-step guide here, and the first step is to test the FSB...

So I tested it with the HTT multiplier set down to 3 rather than 5. Both directions set to 16 bit. I also set the memory down to 1:1 to eliminate it as the cause of any problems. As well I set the CPU multiplier down to 5. This makes sure, according to the guide, that you're testing the limits of the FSB and nothing else.

What I did was within windows I ran up the FSB by hand until it crashed the system, using ClockGen. ClockGen actually went over 300 with no problem. I then went to BIOS, set the FSB there to 300, and rebooted. It booted into windows and I got a desktop at 300, but then it bluescreened before I could do anything.

So I went back to BIOS, backed down to 290, and booted into Windows and it was then stable. I tested it with both Prime95 and SuperPI simultaneously.

Now I'm working on finding the CPU max, which entails running the FSB speed back down, setting cpu mutliplier back up, leaving memory at 1:1, and running the FSB up until I hit instability. Again, I'm testing in Windows, watching for Prime95 or SuperPI errors. I'm increasing by 5mhz steps on the FSB...

Currently I have it running at 2650 stable, so that means FSB is at 265. I don't think CPU will reach the 290 the FSB is capable of on the current cooling (stock AMD heatpipe cooler).

The obvious next step, is to check the memory max. The guide I linked to above is very useful for figuring this stuff out

Btw, I think the CPU itself must have something to do with what the HTT limit turns out to be. On my OLD cpu on this board (an Athlon64 3400+ 939 Newcastle), I couldn't get the HTT up past about 225 or it would bluescreen.

 

[aelfwyne]

Hitting 2700 with memory at 1:1 still. I'm not sure if I can go any further on current cooling, as stress test is putting the cores up to about 55-57 degrees. I'm at 1.375 volts. I'm sure that with better cooling and more voltage, there's more to be had easily.

This is what I get for using the stock AMD cooler ... I'm debating trying the Thermaltake Venus 12 to compare it to the stock... But the stock cooler is so much quieter than the Venus 12!!! I forgot what it was like to have a system under load that didn't sound like a Hoover.

 

[aelfwyne]

http://valid.x86-secret.com/show_oc?id=139068

Stable overclock... finally. It turned into a nightmare when I got to the memory. Apparently buying PC4000 memory was completely pointless, as the moment I put this memory at anything over 200, it freaked out. Major.

Therefore, I had to set the divider on memory WAY low.... so... here's what I finally got to:

CPU: 10x270 = 2705MHz. (Max 2750 stable) 1.37 volts.
FSB = 270 (Max 290 stable)
HTT: x3 = 811 MHz. (Stable max untested)
RAM: 180 MHz (Max 200 stable.... rated at 250 frickin MHz, only stable at 200), 2.5/3/3/5/8/1T (SPD timings)... (250 @ 3.0 is rated)... Max voltage per A8V (2.8v?)

I tried my best to get the memory speed up, but it isn't going to be possible due to where the dividers fall, without drastically dropping CPU speed.

This is in a closed case with the stock AMD heatpipe cooler. Idle temps are 30c, load 51c.

Any tips?

 

[Slug]

Thanks for the info. I have to run the memory on a low divider as well, even though mine is TCCD or PC4000 Ballistix. My opty 165 will do at least 2400, but I've been unlucky in the HTT dept. I'll try working on that as you suggest. I don't have any additional tips right now. I still like this board better than the MSI Neo2 platinum that it replaced.

 

[aelfwyne]

Well as for me, it turns out that my achille's heal is memory... it refuses to run much over 200, despite being rated for 250.... Apparently as well, reports that the Asus A8V gets unstable at low HTT settings may also be true, as when I set the multiplier down to 3x, it gets unstable. 2x and 4x HTT multipliers are stable, but I can't run past about 2.6ghz if I set the HTT to 4x.....

So I have a board that's capable of reaching 290, a CPU capable of 2.75 (maybe better with better cooling), but have to run it at 2.6 now due so I can have a 4xHTT and not have the memory any over 200...

2.6 is good though. My old single core was only 2.2.