OC & Modding Post your overclocks !

Nikhil said:
@Sandy ---- SuperPi is not a true indicator of stability.

2.4 GHz on a Winnie is quite good. So, I suggest that now, you hold back on the oc'ing and concentrate in the stability.

Run MEmtest for the RAM, and Prime 95 for the CPU.

In Prime 95, the small FFTs test is the most taxing for a CPu. So, choose that test and let it keep running forever :p It should run for at least 24 hrs or more. But one thing I like abt Prime small FFTs test is that you can continue to use your comp normally. You will not feel that it is slow. You can even play games while running Prime. At those times, The game uses most of the CPU. The moment you exit the game, Prime 95 utilises 100% of the cpu again.

So, run Prime 95 for as long as possible. 2.4 GHz is a good speed and now, you have to make sure that it is stable
Click to expand...

thanks Nikhil :) :cool2: I will try them both
I have MEmtest
ehh can you tell me where to download Prime 95 from? :ashamed:

@Freaky Will set the Prority to 10 and leave running over night :eek:hyeah:
 
Last edited by a moderator:
Sandy said:
thanks Nikhil :) :cool2: I will try them both
I have MEmtest
ehh can you tell me where to download Prime 95 from? :ashamed:

@Freaky Will set the Prority to 10 and leave running over night :eek:hyeah:
Click to expand...
I prefer to use this version of Prime 95 - SP2004 (stress prime 2004)

It does the exact same tests as Prime 95, but It's got a much more user friendly interface :)

It also shows you how long its been running for unlike the standard version.

Get it here: http://sp2004.fre3.com/
 
Last edited by a moderator:
dipdude said:
@quad why is your dram idle timer set to 256 clocks, is that normal, usually it is around 16 clocks.
Click to expand...

Guys can someone please elaborate on this matter will this increase my sys
performance i.e. mem bandwidth.

I will check if its set at 256 or auto.
 
Last edited by a moderator:
DRAM Idle Timer

^ quad, this is from adrian's rojak pot

DRAM Idle Timer

Common Options : 0T, 8T, 16T, 64T, Infinite, Auto

Quick Review

This BIOS feature sets the number of idle cycles that is allowed before the memory controller forces such open pages to close and precharge.

The premise behind this BIOS feature is the concept of temporal locality. According to this concept, the longer the open page is left idle, the less likely it will be accessed again before it needs to be precharged. Therefore, it would be better to prematurely close and precharge the page so that it can be opened quickly when a data request comes along.

It can be set to a variety of clock cycles from 0T to 64T. This sets the number of clock cycles the open pages are allowed to idle before they are closed and precharged. There's also an Infinite option as well as an Auto option.

If you select 0 Cycle, then the memory controller will immediately precharge the open pages as soon as there's an idle cycle.

If you select Infinite, the memory controller will never precharge the open pages prematurely. The open pages will be left activated until they have to be precharged.

If you select Auto, the memory controller will use the manufacturer's preset default setting.

Most manufacturers use a default value of 8T which allows the memory controller to precharge the open pages once eight idle cycles have passed.

For general desktop use, it is recommended that you choose the Infinite option so that precharging can be delayed for as long as possible. This reduces the number of refreshes and increases the effective memory bandwidth.

For applications (i.e. servers) that perform a lot of random accesses, it is advisable that you select 0T as subsequent data requests would most likely be fulfilled by other pages. Closing open pages to precharge will prepare those pages for the next data request that hits them. There's also the added benefit of increased data integrity due to more frequent refreshes.
 
it does not make too much difference to your performance. Be it 256 or 16/18.

You are setting the max limit. Not the actiual idle time. So you are giving it window of 256 clocks, if ram needs 18 clocks it will operate at 18.

But if you keep it too low say 16 and ram needs 18 then you will run into stability issues.

I am 99% sure of this as i read this long time ago somewhere and practically tested this many times. Does not impact performance for me.
 
ok thanks a lot so i feel its better i dont touch that.
since currently everthing is running fine except for prime which stops after
like 2~4Mins :(

I am really clueless , feel thats its only the temps thats causing it to stop.
 
^ do run memtest too, and prime stopping within 2-4 mins does not signify a stable system, its bound to cause hard disk corruption, becareful, your system is unstable & i dont think its b'cuz of temp's.
 
guys i really need some advice now , hey dip u r making me worry.

I am working on this system for approx 2Hrs now.
even i have gamed for approx 2Hrs continuously without any hangs crash etc...
memtest i have done 2 pass without errors.
 
Simply put, If it can't prime, it aint stable no matter what anyone says.

On my previous P4 3.2 @ 3.9 GHz LGA775 rig, I could run benchmarks, NFS-U2 for like 4 hours continuously and Folding @ Home 24/7, but when I checked, it wasnt prime stable for more than 40 mins. I bumped the vcore up 0.05 v more, and it was stable for more than 24 hours.

Running prime for 24 hours is overkill TBH, but you should run it atleast 6 hours to ensure stability.

EDIT: @dipdude, that version of prime is fine, but personally, I find this more user friendly. Give it a go ;) - http://sp2004.fre3.com/
 
Last edited by a moderator:
exactly, if you are not prime stable then you rish massive data curruption at any given time. I ignored the prime for a while but one fine day i lost all of my data and since then i had been checking prime stability.

If you are getting prime failure soo soon then back down the overclock. Temperature/RAM may not be the problem but processor itself might not be able to sustain those speeds.
 
Downloading both version.
BTW currently i have passed a Vcore of 1.58 earlier it was 1.62
and prime used to run for 8 to 13Mins.

so u guys want me to pass more Vcore i will fry the processor.
Its better that i will try to underclock it to 250HTT and check.

Also my PCI Bus Freq is 105Mhz has it ne thing to do with this unstability.
 
funkymonkey said:
exactly, if you are not prime stable then you rish massive data curruption at any given time. I ignored the prime for a while but one fine day i lost all of my data and since then i had been checking prime stability.
If you are getting prime failure soo soon then back down the overclock. Temperature/RAM may not be the problem but processor itself might not be able to sustain those speeds.
Click to expand...

What about SuperPI 32MB that does count as stable? :S
 
@kitt, the sp2004 is based on very old source code(v23) of prime95, prime 95 has been optimised drastically for athlons in v24, here's the what's new of prime95 v24.14, so my advise stick to prime95 :cool2:

New features in Version 24.14 of prime95.exe

1) A bug that caused the torture test to hang on 256K FFTs on SSE2 machines with 128K of L2 cache was fixed.

New features in Version 24.13 of prime95.exe

1) Celeron D (256K L2 cache) and Willamette (also 256K L2 cache) now have

different implementations for several FFT sizes. This results in an improvement of several percent for the Celeron D.

2) A bug that caused some machines to generate "Error 2252" when communicating with the server was fixed.

3) SSE2 trial factoring code had a bug when factoring very large exponents.

New features in Version 24.12 of prime95.exe

1) For SSE2 machines the larger FFTs have been changed to more effectively use a wide variety of L2 cache sizes. The previous version was optimized for a 256KB L2 cache only. Depending on your CPU and FFT size, you could see an improvement of several percent.

2) As a side "benefit" even larger FFT sizes are now supported. This allows testing of exponents up to 596 million. Not recommended.

3) The factoring breakeven points have been recalculated using my 2 GHz P4. This version of prime95 will do less trial factoring.

4) Since server database crashes cause spurious error 3 messages, prime95will now ignore error 3 messages from the server for 72 hours. This should workaround the problem whereby a result is reported but no CPU credit is given and the reservation is not cleared.

5) Fixed crash bug trial factoring exponents above 286 million.

New features in Version 24.11 of prime95.exe

1) Faster FFTs for AMD64 using prefetchw in both 32-bit and 64-bit mode. You can expect about a 15% speed improvement.

2) Prime95 now detects support for 3DNow! instructions. See undoc.txt for overriding this detection with CpuSupports3DNow=n in local.ini.

3) Factoring entries in worktodo.ini now accept exponents up to 2 billion.

4) Improved checking for memory allocation errors during a torture test. Better guessing of amount of memory to use in a blend test.

5) Added timeouts to PrimeNet communications in hopes of avoiding rare hangs when contacting the PrimeNet server.

6) Fixed rare bug where P-1's GCD could miss a factor.

7) Added trial factoring to the benchmark.

8) Fixed bug in ECM when using zero-padded FFTs.

9) SSE2 macros optimized for an additional 1-3% improvement on P4 and AMD64 CPUs.

New features in Version 24.10 of prime95.exe

1) 64-bit version of prime95 available for 64-bit Windows.

New features in Version 24.9 of prime95.exe

1) Fixed crash when running on machines with "no-execute" protection.

2) Fixed mis-detection of some P3 and Celeron CPUs.

New features in Version 24.6 of prime95.exe

1) Athlon users should see a 20% to 30% speed increase.

2) Pentium, Pentium II, Pentium III users should see dramatically WORSE performance. Run version 23 instead.
Click to expand...
 
Sandy said:
What about SuperPI 32MB that does count as stable? :S
Click to expand...
cant vouch for that 100%.
If its 32M stable in superpi the system is farely stable but i have seen prime failing at speeds at which 32M passes and it has given me problems.
I dont bother running prime for 6 hours though, my machine is rarely idle for that long. For me the max i test is usually 3-4 hours. thats good enough for me.
 
Guys Latest update.







3Hrs 59Min Rock Stabble and still rocking ;)
Cabinet kept open too for prime testing.

Lowered the HTT to 270
Lowered the Vcore to 1.54V
Ram running at 221 2.5-4-4-8 @ 2.8V

This is the best Prime stable i have got after spending my whole night on this.

1 setting prime stopped at 1Hr 45Mins
1 setting stopped at 2Hrs 8Mins

After a playing a lot with BIOS this is the max i got.
Guys your opinion please.

Can i call this stable.

Now even gamed for some time all the games i have are running smoothly.
 
Quad Master said:
Guys Latest update.







3Hrs 59Min Rock Stabble and still rocking ;)
Cabinet kept open too for prime testing.

Lowered the HTT to 270
Lowered the Vcore to 1.54V
Ram running at 221 2.5-4-4-8 @ 2.8V

This is the best Prime stable i have got after spending my whole night on this.

1 setting prime stopped at 1Hr 45Mins
1 setting stopped at 2Hrs 8Mins

After a playing a lot with BIOS this is the max i got.
Guys your opinion please.

Can i call this stable.

Now even gamed for some time all the games i have are running smoothly.
Click to expand...

Good effort there quad.:clap:
"After a playing a lot with BIOS " = please elaborate in detail about this.....

I still maintain the suggestions given earlier....
1) get a better cooler asap ....(hopefully it would be done in a day or two)
2) get better quality, high speed , LL rams
3) it would take upto 2 weeks to properly fine tune the m/c as most of the memory timings/ other settings i had kept on "AUTO" for ur sake.
4)Do not get unduly perturbed about prime stability and all that for now(incl. data corruption etc...). ur cpu can easily do 285 mhz prime stable given the right vcore and adequate cooling and the right ram settings.
5) The rate at which you are going i would not be surprised if you do indeed manage to find the next highest prime number :eek:hyeah: :clap: (this was the original purpose of the program ...i think)
 
Back
Top