Hit a wall overclocking @ 2400 Mhz

[MaxPow3r]

Maybe this is the limit of my processor, but I figured I'd ask for some suggestions before giving up. My computer is currently stable at 240x10, but when I take it to 241x10 I get Prime95 rounding errors. I've taken the core voltage all the way up to 1.6 and still can't even pass Prime95. I have adjusted my LDT mult. to keep the HTT below 1000. I have tried running my RAM as close to 200 Mhz (stock timings) as well as put it on a divider with loose timings. I've increased/decreased the LDT and chipset voltage as well.

 

With my CPU voltage @ 1.6v and my processor @ 2500 Mhz, before it failed prime the temps only reached 40C, so temps don't seem to be an issue.

 

I read some people had a problem getting past 240 Mhz with SATA drives, but my hard drive is IDE so that's not a problem.

 

Any suggestions on what I should play around with next? I was hoping to reach 2600 Mhz, but I don't think that's going to happen.

[ninja_xxx]

have you tried the 9 multiplier.....

[Void]

Did you try the Special Voltage increase, to increase the voltage by percentages? Some people were saying it helped stability, also, have you tried the latest official bios? I couldn't pass 2400 either till I flashed.

[MaxPow3r]

Did you try the Special Voltage increase, to increase the voltage by percentages? Some people were saying it helped stability, also, have you tried the latest official bios? I couldn't pass 2400 either till I flashed.

 

Yeah, I use the "Special" setting to increase the voltage. Check my sig, I'm using the latest official BIOS, 623-3.

 

I tried the 9x multiplier as well and got the same result.

[newguy]

What divider you tried? If 9/10 won't work then try 5/6, also did you try using 2T? Your RAM should be able to hit at least 250MHz. What's the RAM voltage?

[MaxPow3r]

Yeah, I've tried using 2T and several different dividers. Right now my voltage is at 2.8, but I've tried 2.7-2.9 without success. I've successfully passed 8+ hours of memtest with my RAM @ 250 Mhz as well, although I know that doesn't necessarily mean it's stable in Windows.

[BlackDragon24]

Maybe this is the limit of my processor, but I figured I'd ask for some suggestions before giving up. My computer is currently stable at 240x10, but when I take it to 241x10 I get Prime95 rounding errors. I've taken the core voltage all the way up to 1.6 and still can't even pass Prime95. I have adjusted my LDT mult. to keep the HTT below 1000. I have tried running my RAM as close to 200 Mhz (stock timings) as well as put it on a divider with loose timings. I've increased/decreased the LDT and chipset voltage as well.

 

With my CPU voltage @ 1.6v and my processor @ 2500 Mhz, before it failed prime the temps only reached 40C, so temps don't seem to be an issue.

 

I read some people had a problem getting past 240 Mhz with SATA drives, but my hard drive is IDE so that's not a problem.

 

Any suggestions on what I should play around with next? I was hoping to reach 2600 Mhz, but I don't think that's going to happen.

 

Are you stable in, say superpi32M, 3dmark01 at higher frequencies? And what version of prime95 are u using....makes sure to use the 2413 version as all others have problems on A64 platforms.

 

What vlotage do you need for 10x240? Have you tried less voltage at the speeds you failed at....I know it sounds weird but some people find increasing voltage makes no difference.

 

Edit: interesting you guys had problems before flashing....I haven't found a BIOS better than the 510 series for my Venice....try 510-3 and see if that works too.

[MaxPow3r]

Are you stable in, say superpi32M, 3dmark01 at higher frequencies? And what version of prime95 are u using....makes sure to use the 2413 version as all others have problems on A64 platforms.

 

I'll check that when I get home from work today, I'm not sure what version I'm using. I never really tried superpi or occt...

 

What vlotage do you need for 10x240? Have you tried less voltage at the speeds you failed at....I know it sounds weird but some people find increasing voltage makes no difference.

 

I have it set for 1.45v and I have tried just upping the FSB without changing the core voltage, but my computer either locks up while loading Windows or when running prime95.

 

Edit: interesting you guys had problems before flashing....I haven't found a BIOS better than the 510 series for my Venice....try 510-3 and see if that works too.

 

If all else fails I may give that a try.

[BlackDragon24]

I'll check that when I get home from work today, I'm not sure what version I'm using. I never really tried superpi or occt...

 

 

 

I have it set for 1.45v and I have tried just upping the FSB without changing the core voltage, but my computer either locks up while loading Windows or when running prime95.

 

 

 

If all else fails I may give that a try.

 

Also a couple of things worth mentioning/asking...did you make sure your system was memtest stable before installing windows? Did you install window while overlocked (bad idea)?

 

Usually this is the way I measure stability:

 

1M Superpi calculation....if passes,

32M superpi calculation....if passes,

12 hour prime95 blend torture test.

Finally, once cpu is stable when tweaking the vid card, 3dmark01, 03, and 05.

 

The reason I suggest superpi first is that the 1M test takes about 30 sec, and the 32M test takes about 28 min...which is much faster than waiting hours for prime95 to fail....it also is a good measure as to whether prime95 will fail or not.....if you can;t pass it, you wont pass prime95.

 

When is it erroring out in prime95? During a large fft (say 1024, 896, etc.) or during a small fft (8, 16)? Failure during smaller fft's usually is a problem with cpu, larger with memory. Something to help you troubleshoot....

 

For example, I've been trying to stabilize my rig at 10x260....yesterday I errored out on prime95 during an 896K test, which told me mem problems. Bumped the voltage up a notch and now I am 12 hour prime95 stable.

 

We've got similar setups....PM me if you need any specifics about timings, etc. I'm thinking loosening to cas 2.5 might help you out, and also will allow you to go farther at 1:1.

[MaxPow3r]

Well I just checked and I am using Prime95 24.13, so that doesn't seem to be the issue.

 

Also a couple of things worth mentioning/asking...did you make sure your system was memtest stable before installing windows? Did you install window while overlocked (bad idea)?

 

Everything was stock when I installed Windows. I loaded optimized defaults then configured my memory timings to the stock settings and ran memtest for 8+ hours. I didn't get any errors so I installed Windows without any problems.

 

When is it erroring out in prime95? During a large fft (say 1024, 896, etc.) or during a small fft (8, 16)? Failure during smaller fft's usually is a problem with cpu, larger with memory. Something to help you troubleshoot....

 

I didn't really notice when it was erroring out, I'll try OC'ing again and pay attention to that. Thanks for the help, I appreciate it.

[MaxPow3r]

Alright, here's the settings I'm testing:

 

200

enable

2.5

3

8

3

8

16

3

3

2

3

3864

auto

enabled

auto

0

7

2

7.0

auto

256

disable

16

07

disable

 

240

4

16 16

10

100

disable

1.35v

113%

1.2v

1.5v

2.8v

 

And it appears to be failing during large FFT's, namely 1024K. So should I blame my RAM for the problem? Any suggestions as far as settings go?

[Fight Game]

I think your 200mhz ram at 240mhz is doing a fine job. Your dram drive strenth and the other one data, which you list as 7,2. Try 8,1. And you cpu voltage 1.35*113%=about 1.5. Get this up to 1.6. Try 1.325 *126%. Put your ram on a 9/10 divider- I know you've tried this already, but have you tried these other things I'm suggesting along with the divider? Just an idea. GL.

EDIT: Also try 8 ns instead of 7.0