G.Skill 2GBHZ 1GX2 DDR580 PI 32M Done

[GSKILL TECH]

sorry. i didnt think this would be such a touchy topic so i'll end it. as for only RMAing stuff that is defective, thats not true. ive returned my Corsairs just because i was unsatisfied by them. talked to newegg rep and told them they didnt do what i needed them to do and they took it back even though i had them for over a month.

there are several people helping me on XS and i really appreciate it, but ive had some people tell me that my HZ were "bad". THATS why i decided to do this. didnt mean to stir up any trouble so once again, i apologize.

well it is totally not a good thing for us, it gives our rma dept extra work.......
and i already request NewEgg about in future, all rma most send to G.Skill USA and test 1st before it replace

[blazin-asian]

well it is totally not a good thing for us, it gives our rma dept extra work.......
and i already request NewEgg about in future, all rma most send to G.Skill USA and test 1st before it replace

im very sorry guys. it will never happen again.

[tHe_jOkeR]

hey guys show me some prime blend tests not superpi

[4Qman]

Super PI 32Mb pass is easy on this HZ set at 3-4-3-8 265Mhz however prime will crash after a few hours.

[tHe_jOkeR]

mine do 16m 285 3-4-4-8 32M not tested yet but prime crashs
after a few minutes...

is 16 or 32M stable enough? i dont think so

[4Qman]

Id say a good test is around 4Hrs prime or BF2.

Altho this is a 32MB Super PI topic and i can see the benifit of having it as it s a good base to have the correct timmings. But a few Hrs of BF2 or Prime is the real Tester.

[bachus_anonym]

hey guys show me some prime blend tests not superpi

If your memory can consistently run SPi32M at a given RAM clock then there is a great chance that P95 (or SP2004) will hum along for few hours as well. E.g. here:

292MHz 8-4-4-3.0-1T @ 292x10 div200 - SPi32M & 2 hours of SP2004 LargeFFT 1820MB used

292Mhz 8-4-4-3.0-1T @ 350x10 div166 - SPi32M & 2 hours of SP2004 LargeFFT 1820MB used

or from my OCing report I did 2 months ago (different CPU):

290MHz 8-4-4-3.0-1T @ 290x11 div200 - 6 hours of SP2004 LargeFFT used

I have not tried running 6hrs @ 292MHz lately but I'm pretty sure it would pass as well Also, I can't reliably complete SPi32M @ 294Mhz as It once passes and then it won't pass again. So you can see that very often with with those sticks, by maxing it out in SPi32M you can be quite sure that this would be the same spot (+/- maybe 2-4MHz) where P95 stability can be found. OCing UCCC is that straightforward

[AndreYang]

mine do 16m 285 3-4-4-8 32M not tested yet but prime crashs
after a few minutes...

is 16 or 32M stable enough? i dont think so

SP2004 Prime95 Blend 9

[4Qman]

solid as a rock mate,

Im Priming now at 272Mhz 3-4-3-8 as mine havent got the balls to run 300Mhz no chance.

Then again mine dont post with (Read Pre 5ns)

Also ive done a few test and without a 120mm fan blowing on my Sticks they crash and fail Memtest. I took the fan off they failed then added it they pass and i done this over 4 times and everytime it happened

So 120mm Fan stays and that at 2.54v

[tHe_jOkeR]

okay i see, seems like 32M is enough

[Krohling]

okay i see, seems like 32M is enough

I really don't know if 32M is enough

I've run Super PI 32M using the 2GBHZ @275 MHz, and it passed
Then next I ran Memtest86+ test #5 and in the second pass it pointed 2 errors.

Its weird. SuperPI 32M stable, but memtest points 2 errors just after the second pass.

[-.-PhanTom-.-]

Hmmmm, it would seem that these sticks generally perform better in DFI Ultra/SLI than in the DFI Expert?
Which CPU's are you guys using in your Ultra/SLI mobos? .... my bet says NOT Dualcore Opteron....
Seems DFI Expert+dualcore Opteron is a bad mix for these sticks?
Hoping that a BIOS update or custom made BIOS can fix this "issue" with this hardware combination....
Just wonder if it is Expert+RAM alone, or if the funky memory controller of the dualcore Opterons (well atleast the 170s) also plays a part....

Hmmm noone has any comment/input to this statement?
Looking at other people's overclocking results with these sticks, the peeps that get +280 are all using DFI Ultra/DR and most often single core Opteron or X2's ....hence my hypothesis..

[4Qman]

Hmmm noone has any comment/input to this statement?
Looking at other people's overclocking results with these sticks, the peeps that get +280 are all using DFI Ultra/DR and most often single core Opteron or X2's ....hence my hypothesis..

Good Point mine hit the Wall at 275Mhz on a Expert.

[tHe_jOkeR]

I can do this



but prime still fails after less than one minute,
and ut2004 also freezes after 5min....

[MikalCarbine]

Just curious, I might be getting some of these myself.

What Sandra/Memtest bandwidth are you guys getting?

[4Qman]

Just curious, I might be getting some of these myself.

What Sandra/Memtest bandwidth are you guys getting?

There u are mate. 7354mb/s

[palehorse]

Mine seem to have hit a wall at 276. I was able to pass 32M at 280, but memtest #5 gave 16 errors in 40 runs at that speed. I backed them down 1 mhz at a time until I was able to run test#5 20 times, test#8 20 times, and then 10 full hours of all the tests (18 full runs) with no errors @ 276!!

So now I'm finally 32M stable, memtest stable, and hopefully BF2/WoW stable @ 2760. If anyone sees any timings I could possibly tighten without effecting my stability, please tell me. Otherwise, here ya go, perhaps these settings will help someone:

TREF = 3684
DDS/DDDS = 7/3



And then for Mikal:

[ccokeman]

I was hoping for more myself but they pass prime 95 for 24hrs at 265x11 on an opteron 148. Bought the opteron thinking it would have a better mem controller. Still get 265 stable.[IMG][/IMG]

[flesheatinvirus]

Im using Gskill HZ with an Opty 148 also on a DFI Ultra D with original Bios and so far they can do 272 almost completely stable. I do get a ton of errors though right at the end of test 5 in memtest but can run any of the 3DM benches without issues. I have heard the 704-2bta bios is very good and many people here perfer it though I have no problems with the board, might switch over anyway.

[palehorse]

Im using Gskill HZ with an Opty 148 also on a DFI Ultra D with original Bios and so far they can do 272 almost completely stable. I do get a ton of errors though right at the end of test 5 in memtest but can run any of the 3DM benches without issues. I have heard the 704-2bta bios is very good and many people here perfer it though I have no problems with the board, might switch over anyway.

ya, i just switched over myself, and it's a very stable bios version, and may be the reason I just gained 6mhz in my memory OC. (I couldnt guarentee that because I also tweaked the timings...)

[blazin-asian]

ccokeman and flesheatinvirus:

which stepping of the opteron 148 do you both have? just trying to see if my CAB2E can compete

[flesheatinvirus]

Im running a CABYE 0528GPMW 148 which I run a little over 3G at 1.45 v Max on this chip is just over 3200 so far. My other Opty is an CABNE 0528GPMW 150 which so far reached 3200 also but at higher vcore.

[scottk]

Im using Gskill HZ with an Opty 148 also on a DFI Ultra D with original Bios and so far they can do 272 almost completely stable. I do get a ton of errors though right at the end of test 5 in memtest but can run any of the 3DM benches without issues. I have heard the 704-2bta bios is very good and many people here perfer it though I have no problems with the board, might switch over anyway.

The 623-3 BIOS works better than the 704-2bta for me. Both BIOS topped out the same (11x272 memtest86+ stable), but I got lower bandwidth with 704-2bta. I'm running an opty 148 @ 11x271, 12+ hours P95 stable in blend, small fft, and large fft tests. These UCCC sticks behave different from BH5 or TCCD... my P95 stable speeds are usually about 4 or 5 HTT/FSB MHz lower than memtest86+ stable speed.

Also, my sticks scale with voltage increase up to 2.8 and get pretty hot.... so I pulled off the heatspreaders and gained 4 MHz on the HTT.

The sensitive memtest86+ tests for my config is at the end of T5 (like yours), and in T8.

[easyrider]

In memtest I only start getting errors at 40% before this its error free at 280 mhz.

I can game for hours at 270 1:1 with my opty 2.7ghz and primes for 12 hrs error free.

However I run memtest at 270 and I get errors on test 5

250 no errors like you would expect @ 2.6v
260 no errors like you would expect @ 2.6v

270 errors @ 2.6 ,2.7,2.8 v



am i being a noob?

I have tried 2.6-2.8v and no change.

all this is at stock timings

[easyrider]

Forgot to mention that I tried memtest in windows and I get errors also at 270mhz.

Could it be my cooling over the ram?