Jump to content

SanDeigo 250Mhz x11 4x512M UTT


Recommended Posts

It wont boot with ram at 1T, but with 2T....

 

 

250Mhz x11, 2750Mhz 8,2,2,2.0 on the ram using 4 sticks of Twinmos UTT @3.4Vdimm. Purrrrs like a kitten. My HTT is 250Mhz x3.

 

So 2Gigs of ram will work overclocked, totally stable, Vcore at 1.58V SuperPi, StressPrime, 3D 2k1-2k5, gaming for hours all stable.

 

Only issue I found was stock X850XT would croak in 3D03....needed stock clocks lowered. Ill switch over to AS3, etc and then retest.

 

Bandwidth loss over 1Gig 1T is there, but benefit in games out weighs it.

 

At 250Mhz x10, 1Gig, same timings as my infinity system the DFI SLI-D 3700+ SanDeigo X850XT is 20-25% faster. Hell of a boost for the CPU and memory controler internals vs the NForce2. Of course, when I switch it to the x11 multi....seeeeeya! :nod:

 

Im especially happy with the stability of the voltages on the board. You guys did a great job on the voltage regulators. Im using the 4V jumper, no issues at all.

 

Im a happy camper, the system has done exactly what I hoped for, 250Mhz x11, 2Gigs of ram @tight on air cooling.

 

Ill have to get a 4000+ with x12 and water cooling to go for 3Ghz unless I can master the art of ram dividers.

 

Only classy lady I have is with WesternDigital for not giving the Raptors a SataII controller... what the FK are they thinking? But ATTO puts the loaded WinXP 2x raptor raid 0 at 125MB/s write, 110MB/s read with a Gdamn fast 10Kwrite/25Kread first entry. HDtach gives 115MB/s sustained, 218MB/s burst, 8.0ms access and 10% ultization.

Share this post


Link to post
Share on other sites

New SuperSystem: Courtesy of Monarch, latest resuts.....

 

3700+ SD cabbage, 4x 512M sticks Twinmos TM400II UTT "AA4T", DFI Ultra converted to SLI-D, 5V memory jumper, OCZ 520 Powerstream, 2x 36G raptor raid0, X850XT, mod'd AMS CF1009 case, NEC 3520A 16xDVD burner.

 

Stock HSF on SD, AS3 used for all. X850XT at stock clocks. SD at 250Mhz x11, 250Mhz HTT (x3), 1.250 + 126% (1.50)Vcore, Vdimm @ 3.4V, 1.4Vldt, 1.9Vchipset.

 

2Gigs @ 250Mhz 1:1 2T 8,2,2,2.0, bankinterleave ON.

 

PCpitstop... 2465 (my Infinity is 1806)

 

32M Superpi mod 1.4 26m54.156sec... temp peaks 49C stock HSF

 

3D2K3 13075 (cpu 1356)

3D2K5 5931 (cpu 5611)... GPU Mhz dependent, Infinity sys gets 6300+

3D2K1 30159

 

HDTach 218MB/s burst, 116MB/s avg, 8.0ms access, 10% CPU utilization

Atto 0.5 10180write 27469read

Atto 1024 127522write 107440read

 

SSoft CPU 12755/5638

SSoftMEM 6004/6014.... 8000 theoretical max, 2T hurts here 75% rating

Everest 2.0 6635 read, 2300 write, 41.9ns lat

 

The new memory controller is working like a charm on 4 sticks. Strikes me X850XT's HSF should do better job, maybe I need to lap GPU mating surface, just AS3 still hits 72C GPU peak.

 

I am seeing a coldboot quirkyness. Warm boot its fine, rock solid, but unplug power, come back in a while, boot takes some trying. Maybe this needs tweaking of the startup Vcore? Any ideas? Only happens on some of the UTT sticks, the main 4 sticks I tested with (I had set aside for this system) boot fine always.

Share this post


Link to post
Share on other sites

It wont boot with ram at 1T, but with 2T....

 

Just so you know, when running 4 sticks the 2t thing is a memory controller limitation...

 

The previous 90nm Athlon 64 processor core aka Winchester had certain performance limitations when four DDR400 SDRAM modules were used. If there were four single-side DDR400 SDRAM DIMMs installed in a system with a Winchester based Athlon 64 CPU, they could only work with the 2T timing, which caused a few percent drop below the usual performance level. If there were four double-side memory modules installed, DDR400 SDRAM would not work at all and its working frequency would automatically drop down to 333MHz.

 

AMD engineers promised to eliminate this problem in the new Venice core and they did keep their word. Athlon 64 processors based on Venice can work with four single-side DDR400 SDRAM modules without any limitations, and if there are double-side DDR400 SDRAM DIMMs installed, they can work at 400MHz with 2T timing.

 

http://www.xbitlabs.com/articles/cpu/displ...4-venice_3.html

 

...and it applies to both the new Venice and San Diego cores. So, if you can find four sticks of single sided UTT then you can run them at 1t, otherwise four sticks of double sided will mean 2t :)

 

(As mentioned the limitation only applies when running four sticks, so two sticks of double sided has no limitations nor should running three sticks of double sided)

Share this post


Link to post
Share on other sites

I have 8 sticks of UTT and 2 sticks of BH5, the 4 UTT I set aside for my brother a couple months ago... 245Mhz each in NF2... work 100% PERFECTLY.

 

The 4 sticks I have... that do 245-250Mhz NF2 (2 of which have the voltage bug) have cold boot problems on NF4. The 2 sticks of BH5 also have cold boot issues.

 

Im happy I can ship out his system, but looks like Ill need to hunt down 4 more UTT sticks that parallel his sticks if I seek happiness for myself.

 

Also, the 1.5 cas option works fine on 4x 512M sticks, doesnt increase bandwidth much but works.

Share this post


Link to post
Share on other sites

ya 1.5 doesnt increase bandwith (usually). for some cpu's it increases stability.. and for most it increases the speed/latency ever so slightly.. for some it may make things worst then 2.0.

 

hmm coldboot. thats a wierd one which no one can explain (and happens as i have been noticing with other people round here). i rarely get it (has happened only twice in the 3 months i have had the board). try raising startup VID.

Share this post


Link to post
Share on other sites

uwackme... I have been searching XS and DFI Street for this exact thread.... I knew you would post it sooner or later!

 

Good sheet man, you did the ground work for this upgrade longer in advance than I have ever before... it paid off man, that rig is sick!

 

Bout the cold boot... I have a similar issue and am using CH-UTT as well... but I don't know if it's directly related to them... I mean, sheet, my VX are faultless with or without max vDimm...

 

What I DID see increased the cold boot rate was raising my 3.3vRail... and I am pretty sure your old DFI nF2 had a "tweaked" one as well, so if you have carried over that PSU or you have modded your new one's 3.3vRail this may be a contributing factor.

 

I say this because to begin with I ran my 3.3vRail @ 3.63v DMM / 3.51v SmartGuardian and I had cold boots but very rarely... now I am running my 3.3vRail @ 3.84v DMM / 3.71v SmartGuardian and the cold boots are a bit more prominent... though the thing that causes them the most is when I "Force Power Down" before I get into WinXP... EG, I hold POWER BUTTON down just after POST.

 

Anyway... congrats man, you have a power house right there!

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...