Ultra-D 3700+ Overclock and 1st Post

[KandyKid]

Hey guys, I've been an avid reader of DIY since back in the days of dfistreet and finally I've mustered the finances and intestinal fortitude to tackle an overclock system I've "seen" for a while... and here it is. Comments, insults, hints are all welcome...

 

but go easy, it's my first day as an active member

DFI LANPARTY nF4 Ultra-D (623-01 BIOS)

 

AMD 3700+ single core (ADA3700DKA5CF) E6

 

2x1024MB OCZ PC3200EL (OCZ4002048ELDCPE-K)

 

EVGA 7900GS KO, 500/1380 93.71 drivers

 

2x320GB Seagate 7200.10 SATA HDD's

 

LG GDR8164B DVD-ROM

 

LG GSA-H50N DVD-RW

 

Silverstone Olypia 650W

 

==========

 

[b]Genie BIOS Settings:[/b]
[color=red]
FSB Bus Frequency.............................  -  250
LDT/FSB Frequency Ratio.......................  -  4x
CPU/FSB Frequency Ratio.......................  -  11x
PCI eXpress Frequency.........................  -  100Mhz

CPU VID StartUp Value.........................  -  Auto

CPU VID Control...............................  -  1.30v
CPU VID Special Control.......................  -  Auto
LDT Voltage Control...........................  -  1.20v 
Chip Set Voltage Control......................  -  1.50v
DRAM Voltage Control..........................  -  2.60v[/color]

[b]DRAM Configuration Settings:[/b]
[color=red]
DRAM Frequency Set............................  -  200=RAM/FSB:01/01
Command Per Clock (CPC).......................  -  1T ... BUT NOT IN BIOS
CAS Latency Control (Tcl).....................  -  2.5
RAS# to CAS# delay (Trcd).....................  -  3
Min RAS# active time (Tras)...................  -  5
Row precharge time (Trp)......................  -  3
Row Cycle time (Trc)..........................  -  8 Bus Clocks
Row refresh cyc time (Trfc)...................  -  14 Bus Clocks
Row to Row delay (Trrd).......................  -  02 Bus Clocks
Write recovery time (Twr).....................  -  02 Bus Clocks
Write to Read delay (Twtr)....................  -  02 Bus Clocks
Read to Write delay (Trwt)....................  -  02 Bus Clocks
Refresh Period (Tref).........................  -  3120 Cycles
DRAM Bank Interleave..........................  -  Enabled

DQS Skew Control..............................  -  Auto
DQS Skew Value................................  -  0
DRAM Drive Strength...........................  -  Auto
DRAM Data Drive Strength......................  -  Auto
Max Async Latency.............................  -  Auto
DRAM Response Time............................  -  Normal
Read Preamble Time............................  -  Auto
IdleCycle Limit...............................  -  256 Cycles
Dynamic Counter...............................  -  Disable
R/W Queue Bypass..............................  -  16 x
Bypass Max....................................  -  07 x
32 Byte Granularity...........................  -  Disable(4 Bursts)
[/color]

 

 

==========

 

SP2004 , cpuZ 1.39 , 3dMark2001SE:

 

Image

 

 

3dMark2003:

 

 

 

3dMark2005:

 

 

 

3dMark 2006:

 

 

 

Aquamark 3:

 

 

==========

 

short description:

250x11 @ 2750Mhz, 1.30v

2.5-3-3-5 @ 2.60v vdimm

[27floyd]

First off, welcome to the forum..

 

That is a very impressive OC you got there.. Is it prime/Orthos stable??

Also use core temp to get an acurate temp. That 37C you got there is probably 10C-15C below what it actually is..

 

Great start though.. Keep it up.

[KandyKid]

Hey 27floyd, thanks for the reply

 

I used Prime for 10hr stress test and all was good, orthos is only for dual cores no?

as for the temps, Everest gives out CPU and CPU Diode temps... both end up being the same after a few minutes @ 100% load

 

But, after all was said and done, I ran Quickpar to do a 10% paring of a set of files.... and the verification returned FAILED... uhoh... ram problem

 

went back to stock... did the same -> pass , so i knew it was an OC stability issue

 

put the ram on divider, set cpu to 2.75ghz -> pass... damn that ram

 

I've been playing with it all day today and will post when i get some settings @ 250mhz 1:1 that work without Quickpar fail

 

This got me thinking that Quickpar can be used as an excellent Ram stability verfication tool. After PI, OCCT, Prime and Memtest passed... this one failed revealing a minute instability.

 

I will post updates on my newest settings as they develop. :shake:

[wardog]

You're looking for stability @ 250Mhz 1:1 using THAT ram with only 2.6v?

 

My thoughts: Wish ya the best of luck trying. Really hope ya can, but ..........

[KandyKid]

@wardog

 

The one thing I found about that ram is that it doesn't care about voltage... at all. It prefers loose timings to voltage increases, although i haven't pushed it past 2.8v during testing.

 

I also hope your not confusing it with the 3-3-3-8 ddr400 ram that ocz makes... this is the 2-3-2-5 ddr400 platinum kit

 

UPDATE//

 

2.5-3-3-5 @ 250mhz 1:1 , memtested and quickpar successful... :cool:

 

i will update the ram bios settings above to reflect this new outcome

[wardog]

Boy, were that to be the case, 2.6v at 250, I'll invite ya to put them in the OCDB

 

That would be the absolute best ever I've seen outta those sticks.

 

What revision is marked on their label?

 

No, no confusion. It's in your sig and mine too. Matter of fact I've owned 7 sets of these, v1 v1.1 and v2. I still have 3 sets and none of the 7 would come close at 2.6v.

 

Even paired with any of my 150's(10x250 sitting on its hands) I've never come close.

 

Please, I encourage you to put them in the OCDB. Many people search that Topic for help in OC'ing. These would make a great addition there.

[Guest SPQQKY]

Those kits with 2-3-2-5 timings are Mosel IC's. You are correct in that voltage will do very little in increasing clock speed. Try 2.5-4-3-8 with 2.6v and see if that gets you any further. I think the highest I got on a set of those was 250 as well, but with the more relaxed timings. Nice kit considering they're Mosel ICs.

[wardog]

Most likely Infineon, but not Mosel. Aeneon(sp?) was another used on 'em too.

 

Regardless, I'd like to see an OCDB entry for 'em. It'd help alot of folks out to view their specific settings.

[Guest SPQQKY]

Infineon were timed 2-3-3-6. Same across the board with Corsair, OCZ, Patriot, etc.. 2-3-2-5 = Mosel and 2-3-3-6 were Infineon. 2-2-2-5 or 2-2-2-6 were the TM BH5 or BH6. Lots of info at XS.

[wardog]

Eh? OCZ4002048ELDCPE-K 2Gb(2x1Gb) DC Kits SPD @ 2-3-3-6?

 

I've seen in the seven I've owned SPD to be 2-3-2-5, and on the later ones SPD is 2.5-3-2-5 for compatability reasons(easily remedied in the BIOS after a successful boot) but never 2-3-3-6.

 

Infineon on the one I dared lift the spreaders on. That was a year plus ago.

 

Not to be rude but even XS, the most widely copied chip list, is even Infineon for 1Gb sticks of OCZ EL.

 

http://xtremesystems.org/forums/showthread.php?t=50010

1GB STICKS RAM LISTING:

PC3200 Platinum EL - Infineon BE-5 + Brainpower PCB

 

And if that's not enough:

http://www.bleedinedgesupport.com/ocz/foru...ead.php?t=12861

 

http://www.bleedinedgesupport.com/ocz/foru...ead.php?t=16069

 

http://www.bleedinedgesupport.com/ocz/foru...ead.php?t=18106

 

.. and many others

[KandyKid]

well i've owned two sets of this ram... in fact still do - the other is sitting in my non-oc machine

 

the only comment I can make here is that when I physically compared the two the chip layouts on the PCB they were visibly different... leading me to believe that they are different chipsets - although i have not pulled the HS's off to find out for sure

 

for ref sake i purchased the 1st set about 1.5yrs ago and the new set about 2months ago for this build

 

@wardog

i've revamped my first post to be under OCDB guidelines, i just need to run the benchmarks over again to fill in the last bit... thanks for telling me about that

[wardog]

Those are stock settings you list out, not the 250 @ 1:1

 

When ya get the "placeholder' links straightened and you're set to move them into the OCDB proper like you might also make note that you have the newer ADA3700DKA5CF.