NF4-D + BP 815 pcb based memory - HOW DO YOU GET INTERLEAVE TO WORK ov

[mad mikee]

1. When Trying out both PQI and PDP (both BP815 pcb) and I turn on interleaving, get massive errors in memtest #8 :eek: in DC mode (SC ok) :confused:

 

2. When trying out Gskill (BP808 pcb) no problems in test 8, but problems on 'bad' stick in tests 4, 5 (Damn can't send it back either, hunting up some more GBLE if any left )

 

Looking for anyone w/ MEM FSB OVER 250 and interleaving enabled that got it to work for BP815 memory (PQI turbo or PDP XBLK) in FULL memtest!?

 

I'm guessing since Oskar has easiest access to Gskill / OCZ (OCZ is also BP808 pcb IIRC), he of course programmed the bios for that and didn't know of this issue.

 

I can get the memory to 10 x 270 (PDP) but get less BW than my neo2 @ same settings (2.5-3-3-7) because of having to disable interleaving and I would rather not have to go backwards.

 

Please help, since I'm sure that many have both of these memories.

 

Thanks MikeE.

[mad mikee]

:confused: :eek:

[LoN3R]

might jus not work on higher HTT as seen on the newest nF3 bios

[MicroLab]

http://www.dfi-street.com/forum/showthread.php?t=5629

 

Interleave ON!

 

bye

 

MicroLab

[mad mikee]

I have seen people w/o GSkill (BP 808) (We haven't heard many OCZ TCCD owners to confirm/deny this since they also use BP 808 pcb) having very serious issues w/ their TCCD, (VERY high voltages needed to get anywhere, unstable )whereas most w/ GSkill are fine (normal TCCD voltage and good results).

I'm getting the feeling that the whole interleave / PCB thing is the root of alot of it.

Fix that and many more will beat a path here, otherwise there will be ALOT of unhappy people (myself included).

 

OSSSSKKKKKKAAAAAAARRRRRR!!!!! - HELP HELP HELP PLEASE!

[RGone]

@mad mikee, yo man the OCZ 4200EL will smoke on the NF4 but not sure how much if any memtesting has been done. Prime95 has been run for over 4 hours on AndyOcz's board with PC3200 Rev2 with tCCd. He has post in overclocking section about it. Sure his with bank interleave on.

 

AG has run his SLI board to 310x8 with 3200+ and Ocz4200EL which is tCCd. And AG is a prime freakingnutsomofo.

 

My two sticks of Patriot 400-533XBL are going for RMA as they will only now run Cas 2.0 it seems and thought that was the NF4's fault BUT they do the very same thing in the NF3 so they are shot. However they will do 260X10 at Cas2.0 but will not change Cas for anything and that ain't really fine plus they "needed" tons of voltage to run on any Nf3 or Nf4 board.

 

So not sure exactly what all is up with this or that and it maybe that some memory is programmed on the stick itself and giving fits. I am trying to find some answers my ownself. I see people beginning to say that memtest may not work well on Nf4 but I do not see how that can be. Just like prime did not work on Nf4. So verdict is still very much up in the air as to what in hale is going on.

 

Now for information purposes. I was just testing some un-named tCCd not on brainpower and it was 270x9 on my 3000winnie and throwing tons of errors in #6 / #7 and #8 with memtest1.50 of course Bank Interleave on. Now back down to 260x9 and running prime and where 270 prime failed before it has now been running 30 mins instead of 30 seconds. When it gets a little more prime time under its belt I will boot out of windows and run memtest1.50 again and see if it fails. It may be the memory itself is not holding up in my situation and not sure how it relates to yours but you now know what I know if you read what I wrote carefully.

 

Yes it has taken a huge amount of voltage to get some tCCd to run for me and the memory is about 2.5 months old and it may have come from that 6week stretch of samsung stuff that plain took a bunch of voltage to run well or it might have been pcb related then too. I cannot nail it all down as there is way too many brands involved.

 

RGone...

[RGone]

Ok here iit is about 2.5 hours later and I have done this.

1:1, CPC on, 2.5, 4, 7, 4, 7, 14, 3, 3, 2, 5, 4708, Auto, Enabled (bank interleave), Increase skew, skew value=0, 6, and all the rest bios defaults as they appear.

3000+ 90nm, 9 default multi, vCore reading 1.57 volts, LTD=1.4volts, Chipset=1.8V and Vmemory is 2.9 on some un-named non-brainpower pcb tCCd memory modules (2x512). Memtest says version 1.5 and will not recognize cpu or memory timings.

 

280x9 = fail prime fail memtest #7 & #8.

 

270x9 = fail prime fail memtest #8

 

265x9 = pass prime 12 errrors memtest #8

 

260x9 = pass prime 0 errors memtest #8.

 

So that is how mine stacks up and regardless of booting and running windows at usually 280x9 the real stability is not nearly as high as one thinks when I got down to doing all the testing.

 

RGone...

[mad mikee]

More info back - in this case memory is fine on msi neo2 up to 270 - 2.5-3-3-7

but fails w/ interleave on after 220 on NF4 - D. 2 new sticks of GSkill on the way form NewEgg. Your results and AG fit it fine w/ the theory so far.

 

Did you try same stuff w/o interleave and see what happens? I think there may be a difference?

[Angry_Games]

Did you try same stuff w/o interleave and see what happens? I think there may be a difference?

 

lol...now why would we want to run our primo TCCD (OCZ 4200EL in my case) with Interleave disabled?

 

ive got the SAME EXACT TCCD that Rgone is talking about also (the unnamed stuff) and when I get a couple customer's machines shipping back to them, I will be testing with his results and see if I get the same exact results...

 

the only other TCCD I have is some Centon stuff that rocks on Intel boards, and rocks on the NF3, but is in momma's machine so have not pulled it down to test on the new NF4.

 

However, momma is getting the Ultra-D + X850XT PE so she'll get the best possible items, hand-clocked to perfection (even with a lil help from RG since he loves momma like a daughter heh)

 

my point? I would never want to run with Interleave disabled and so therefore I will not...I'll keep testing and buying if I have to until I get the kind of clock I want with all the goodies turned on

[mad mikee]

I was just convincing my wife that switching around 'da stix' was required and 'won't hurt her system' :nod:

 

I didn't say LEAVE it off, just try it to confirm der theory sir (I don't wanna leave it off, I want a fix so that all of us that don't have memory w/ BU6808 pcbs can play too (or in my case my pdp spares for my replacement Gskill)) I just painted 'DEFECTIVE' on one of my present set (leavin the stickers on next batch for a year this time dammit!)

[tonedef42]

I got PQI Turbo 3200 512mb (2x256) 2-2-2-5 (BP815):

http://www.pqimemory.com/spec.asp?link=/dc...3200-512DPU.htm

 

I can run 267 x 9 @ 1T 1:1 with Bank Interleave on. Prime95 stable: 12 hrs, memtest stable 3 hrs.

 

Try loosening your timings? What bios are u using...?

If u use 3/16 the yellow slots @ 2.8V are good.

 

edit: btw finding the perfect settings for this ram takes awhile... I'm still trying to get 24hr stable...(well thats my goal)