OCZ 3200EL plat rev 2

[Ubermann]

No it doesn't.

Example:

230 1:1 2.5-3-3-7 Everything else on AUTO passes Test 5-7 but isn't SuperPi Stable.

Changed the TREF to 4708 Cycles and passes Test 5 and now is SuperPi Stable at 39s. And Sandra Bandwidth is at 3,509.

And When I tryed to load Windows at a Drive Strength of 4 I got "Detecting Array". So I switched it to AUTO and I now load into Windows.

I think if it passed ram test but not superpi then its you cpu that has wrong volt or is unstable in some way.

[Cohiba]

Well, I have had my hands full lately with real life, but I finally got a chance to play with the memory again. I am having some problems getting the memory stable. I cannot pass memtest with both 512 modules without getting errors. I have tried stock settings, relaxed settings, 2.5 2 2 10 with all else set to auto, and voltages between 2.7 and 2.9. The memory still errors at 200MHz and on up. Odd thing is, both sticks will pass without errors if they are used by themselves, except in the second memory slot. I cannot get the machine to boot if I have a single stick in the second slot, but 1 and 3 are fine. I have a 90mm fan blowing directly onto the sticks, so they aren't overheating. I am going to do some more tests with each stick by itself to see if maybe one of the sticks is just not getting along with the system. I do wonder though, if it might be the DFI mobo causing the issues.

Good to see that others are having some success though. Hopefully, I can get things squared away stability wise and start having some oc'ing fun too.

[ugp]

I think if it passed ram test but not superpi then its you cpu that has wrong volt or is unstable in some way.

Yeah it could be but I have a lot to play with until I finally find a stable point I am happy with. Sad part is I haven't even begun to try to find my highest CPU clock yes either.

[dav]

Well, I have had my hands full lately with real life, but I finally got a chance to play with the memory again. I am having some problems getting the memory stable. I cannot pass memtest with both 512 modules without getting errors. I have tried stock settings, relaxed settings, 2.5 2 2 10 with all else set to auto, and voltages between 2.7 and 2.9. The memory still errors at 200MHz and on up. Odd thing is, both sticks will pass without errors if they are used by themselves, except in the second memory slot. I cannot get the machine to boot if I have a single stick in the second slot, but 1 and 3 are fine. I have a 90mm fan blowing directly onto the sticks, so they aren't overheating. I am going to do some more tests with each stick by itself to see if maybe one of the sticks is just not getting along with the system. I do wonder though, if it might be the DFI mobo causing the issues.

Good to see that others are having some success though. Hopefully, I can get things squared away stability wise and start having some oc'ing fun too.

it's sad for an extreme ram to fail running at stock!
good luck and wait for your nice results soon!

[Ubermann]

Well i have been messing around for some week now and im still trying =)
Girlfriend is very angry.
But, its a hobby and fun =)

[Samael]

Well i have been messing around for some week now and im still trying =)
Girlfriend is very angry.
But, its a hobby and fun =)

Maybe she needs some oc and "tuning" as well

[Ubermann]

Will tune her later, and if shes not happy by then i need another plan.
Maybe i put some Watt in her then shes very calm.
Adjust the pots inside the bread toaster and IEEEEEKKKKK!!

[TEDY]

don't forget to vdimm her nippless :=)

[mad mikee]

I think #5 is the absolute killer for ram.
If it survives that, then it survives anything.

Checks goldmemory screen (finally passing .5 G BH-5 @ 250/3.2Vdimm)
In 17 minutes max (will run test #5 all day @250FSB/3.1 Vdimm, fails GM in 5 mins @ same, (did it w/ errs during burnin @259/3.3 ) GM (20$) seems to mimmic windows performance more accurately, however need to test w/ both since sometime memtest passes and not GM or vice versa

[ugp]

Ok so I have just tested 265 @ 3-3-3-10 @ v2.9 and Chipset @ v1.8 Times 8.

I tested SuperPi and got 41s.
I did Sandra Bandwidth and got 4,044GB/s
CPU Arithmetic scored 9,737.

I am not sure if this is stable.

But I had the usualy settings like bigtoe recommend...

TREF @ 4708 Cycles
TWCL @ 05 Cycles
DQS Skew Increase of 64
Max Async Latency of 08.0ns
Dymanic Counter Enabled
R/W Queue Bypass 04x
Bypass Max of 04x
32 Byte Granularity Disabled.

I shall run MemTest-86 and see how many errors I get. I don't exactly understand the program though. I ran it twice on a certain test and got one error and I had left the room and cam back and it didn't get any errors the second time around. Why in the hell does it do that?

[ugp]

Well i have been messing around for some week now and im still trying =)
Girlfriend is very angry.
But, its a hobby and fun =)

For some reason girls don't like it when we pay more attention to our computers than we do to them. They get very jealous over a piece of machinary. I don't understand it lol. Just kidding. I know how that is though man. It is very easy to get caught up in messing with a PC. I have been doing the past 3 days all day.

[Section8]

OK I have finally done some testing and I can't get this memory up past 255 1:1 with 2.5-3-3-10 1T? I did get it to boot at 260 once but it crashed soon after a Sandra run. I hit 7,200. I have the boards VDIMM maxed at 2.8? Any suggestions? I have the HT at 800 and the multiplier is at 9 to take those both out of the equation. On a side note I am running at 240 X 12 now with 2.5-3-3-10 1T and getting 7,300 in Sandra. I know I can get more out of this RAM but I don't know what is holding me back. Are there other memory settings I need to change?

[dav]

Maybe she needs some oc and "tuning" as well

i think she has already OCed too much and getting unstable!
better prime test or mem test her...

[ugp]

i think she has already OCed too much and getting unstable!
better prime test or mem test her...

He might need to give her the Touture test
Orgasm her to death and she'd be happy

[ugp]

Here is my highest clock on the memory so far.

Next I am going to try to get it with a 9 multiplier so I can hit almost 2.4Ghz. Any suggestions on how much voltage I should use?

[dav]

Here is my highest clock on the memory so far.

Next I am going to try to get it with a 9 multiplier so I can hit almost 2.4Ghz. Any suggestions on how much voltage I should use?

"single" channel? you got a stick of 1GB?

[Vitalizer]

"single" channel? you got a stick of 1GB?

no.. socket 754 doesnt have dual channel support..

ugp:i can run my mems at 265mhz with 2.5-3-3-10 and 2.8v prime95 stable..
so if u give the mems a bit more voltage u can maby run them at 270mhz.

[ugp]

no.. socket 754 doesnt have dual channel support..

ugp:i can run my mems at 265mhz with 2.5-3-3-10 and 2.8v prime95 stable..
so if u give the mems a bit more voltage u can maby run them at 270mhz.

I always run them @ v2.9 and I am still not stable. I am despiratly seeking answers to why. I have a good board I know I do. And a decent CPU. I dont understand why.

[dav]

sorry didn't notice he's on 754.. :P

[Ubermann]

Try "tref" at 4672 instead.
Seems to be more stable there.

I have never been able to run ram at 300 and cpu at 2.4 and run any 3d benches.
Now it never fails.
Using the omega drivers most lately also.
APIC enabled.

[ugp]

Ehh, I like DNA Drivers. I bench higher with them. And I shall give that a shot. I posted on OCZs forum and I have Tony aka bigtoe helping me out. He said they should do 250 easily.

http://www.bleedinedge.com/forum/showthread.php?t=4689

Might want to check it out if he gives some tips

[Ubermann]

Seems to run fine at 2.5-3-3-0 at 285FSB.
"IdleCycle limit" and "R/W queue bypass" should be put to 8 with those timings.
Gives almost same results at 3dmark01 as 2.5-4-3-0 at 300FSB.

Another thing.
When going over to 3-4-4 then "read to write delay" need to be "3".

2.9volt seems to work best with ram.
Dram level 1 can and should be used on all settings.

Tref 4672 seems more stable then 4706.

This is one a single stick.

[Vitalizer]

I always run them @ v2.9 and I am still not stable. I am despiratly seeking answers to why. I have a good board I know I do. And a decent CPU. I dont understand why.

i cant run my cpu at 2.4ghz with 2 sticks in.. only with 1.. its very strange. so i have to run the cpu at 265x8.

[ugp]

Seems to run fine at 2.5-3-3-0 at 285FSB.
"IdleCycle limit" and "R/W queue bypass" should be put to 8 with those timings.
Gives almost same results at 3dmark01 as 2.5-4-3-0 at 300FSB.

Another thing.
When going over to 3-4-4 then "read to write delay" need to be "3".

2.9volt seems to work best with ram.
Dram level 1 can and should be used on all settings.

Tref 4672 seems more stable then 4706.

This is one a single stick.

Thanks I'll give that a try real quick. I tested both DIMMs real quick and tried to max them out. I shall try it with this now.'

EDIT: I gave it a try and this is what I came up with.

I tried each DIMM by itself in Slot 3 with settings of 2.5-3-3-0. TREF at 4672 Cycles and a Drive Strength of 1. The first stick that I was able to get 290 2.5-4-3-10 with got 124 Errors on test #5. The second stick that I was only able to get 280 2.5-4-3-10 (Those were a quick test by the way) I only got 56 errors.

I do not trust MemTest86 though. My memory pasted it with flying colors and kept getting memory dump errors unless that could be a CPU problem as well.

[Ubermann]

SP2 really sux when overclocking!