Official DFI LanParty UT X38-T2R(and LT) Discussion/Review/Overclock/Guide Thread

[flexy]

hm, weird, when i did this today it completed within 200 secs (i think it was 48 sec per thread) on a x3350 at 3.3ghz . Tho I guess you mightve had stuff running in background. Either that or my brain is playing tricks on me
gonna try again when I get home

well the good thing about OCCT is it does NOT take hours to test...i noticed if i am not stable i either get a problem in the first 20mins...40mins into is another milestone where it often crashes.... If it runs an hour i know i already its halfway-stable. If it runs an hour i can be pretty sure it would run 2 or 3 hours also. Just my oberservation.

[jcool]

Just tried to start systool, gives me a BSOD every time at startup (even though my rig is 24/7 primestable, lol)
Guess it doesn't like XP x64.

[t_ski]

It might have something to do with the sesnors, as I'm having trouble running it on Vista 32-bit. Unfortunately, it's been a while since W1z updated it last.

[flexy]

i had my FSB/Ram a notch up, where i *know* that it's not stable.
Systool run 30mins or so...didnt complain at all.
OCCT/MEM gave me an error 1min already...so did windows memtest...after 3 mins.
Not so sure about systool

[Rinaun]

Just tried to start systool, gives me a BSOD every time at startup (even though my rig is 24/7 primestable, lol)
Guess it doesn't like XP x64.

I get this same issue with firfox now. I think running my rig stock caused some issues installing vista, so I'll be reformatting after I'm done OCing. Anyone else have issues like window processes crashing, and nfts.sys crashing? My memory is OCCT 10 hour stable on RAM test.

[zanzabar]

u need to fix your SB voltage

[Rinaun]

sb voltage will affect this? I have it set to 1.64 but ill try the lowest.

[flexy]

I get this same issue with firfox now. I think running my rig stock caused some issues installing vista, so I'll be reformatting after I'm done OCing. Anyone else have issues like window processes crashing, and nfts.sys crashing? My memory is OCCT 10 hour stable on RAM test.

what OCCT setting? Did you run MIX, and i assume OCCT 2.0.0b.
Your problem seems to be related to NB ot VTT...so i would run OCCT at MEM setting.

Edit: hmm...you say ram-test..what does CPU test say?

[Rinaun]

I did the RAM test for infinite over 6-7 hours, and mix for over 10 (both I stopped). I'm using 2.0.0A. I usually crash hours after booting, usually only when I'm surfing or some crap. I also get windows processes randomly shutting down (Defender, host process, DMI, etc), But I can play HL2 cinemtic through the whole thing.

[zsamz_]

u need to fix your SB voltage

what you runnin you sb volts zanza?

[zanzabar]

what you runnin you sb volts zanza?

im on 1.8xV i had missing ntfs files and bad raid loads with 1.6x (with 2 drives), and i needed 2.00 when i had 5 drives connected to the ich9r

im on a tia board, not a china

[jcool]

What are you guys doing with your SB's
I'm running stock 1,5V, no problems with RAID whatsoever. Maybe it's because I don't have a CF setup (does the 2nd, 3rd etc card even go by way of the SB like with SLI on NV mobos?)

[deadman'sclick]

im also running raid0 & a single HD on the SB controler without the need to add SB voltage from stock. o/cing wise it does nothing for me.

if i remember rightly the SB voltage also affects the cpu ?

[Rinaun]

I tried stock SB and I failed OCCT Ram test on infinite in 28 minutes.

[Jor3lBR]

I tried stock SB and I failed OCCT Ram test on infinite in 28 minutes.

I dunno if you can relate that ONE experience to the sb volts, but only if you would ran a controlled x number of tests.

I've always run mine on stock and never noticed a higher vSB bring any stability.

[t_ski]

I too have random lockups eve running at stock. Usually for me it's when I'm browsing, but it's even happened while using Photoshop. CPU wasn't even loaded at the time.

BTW, I'm running a CFX system and four-drive raid array, but I've been using stock SB voltage, too.

[Hamilton]

Has anyone compared the 400/800 setting with 333/667 for 1:1 FSB:MEM ratio?
Does it change any internal timings in the NB?

[t_ski]

I too have random lockups eve running at stock. Usually for me it's when I'm browsing, but it's even happened while using Photoshop. CPU wasn't even loaded at the time.

BTW, I'm running a CFX system and four-drive raid array, but I've been using stock SB voltage, too.

Since 3DMark Vantage came out yesterday, I've been running the system overclocked to 3.6GHz. Ironically, the system seems a lot more stable and I haven't had any random lockups (knock on wood).

With four sticks I could not use the same settigns I was earlier in this thread, so all I've done is set the ram tp 4-4-4-12 and raise the voltages to what I had them at before.

[eva2000]

Some 2x2GB Corsair PC2-9136C5DF Dominator action



System specs

• Intel Core 2 Extreme QX9650 L739A640
• Thermalright Ultra 120 Extreme + 120x38mm Panaflo 103cfm (MX-2 TIM)
• DFI LP LT P38-T2R 1/11 Official bios
• 128MB Gainward FX5200 PCI
• 2x2GB Corsair PC2-9136C5DF Dominator dual channel kit (dimm 1+3 slots)
• 750GB Samsung HD753LJ
• Pioneer DVD-RW
• 620W Corsair HX620
• WinXP Pro SP2

Results:

Everest Ultimate SPD settings for modules - looks like week 12, 2008 made with a week 13 packaging date on module's Dominator heatsink label - 08130508 v1.1



I've started testing on DFI LT X38-T2R motherboard with 1/11 official bios and they overclock very easily just by setting memory to 571Mhz 5-5-5-15 and vdimm to 2.10v I was already booting and memtest86+ v2.01 stable!

@585Mhz 5-5-5-15 2.10v

Quick jump to 585Mhz 5-5-5-15 at 2.10v still was easily quad Super Pi 32M stable (hyperpi v0.99 gui wrapper) and single Super Pi 32M stable

click image for full screen shot.

CPUZ Validation





Quad Super Pi 32M



Single Super Pi 32M



Everest Ultimate Bandwidth



Memtest Windows stress test 4x 750MB




Max CPUZ Validation
4GB Corsair PC3-9136C5DF @610Mhz 5-5-5-15 at 2.23v

Using setFSB from 595Mhz 5-5-5-15 at 2.23v in windows I slowly pushed the FSB and thus the memory clocks to 610Mhz 5-5-5-15 at 2.23v for max validation

click image for full screenshot



Max Single Super Pi 32M
4GB Corsair PC3-9136C5DF @595Mhz 5-5-5-15 at 2.23v

Straight boot from bios managed to push the 4GB Corsair PC3-9136C5DF to 595Mhz 5-5-5-15 at 2.23v for single Super Pi 32M pass. Have yet to try quad super Pi or higher vdimm

click image for full screenshot

[zsamz_]

i think this will be usefull for 45mn cpus

[eva2000]

i think this will be usefull for 45mn cpus

that table for DFI LT X48 right ? is it same as LT X38 ?

[zsamz_]

that table for DFI LT X48 right ? is it same as LT X38 ?

should be its the same chip
always good to measure gtl volts its more needed for quads duals are more forgiving
eva2000 did you test max fsb yet for the board?

[Dimek]

Which % do you find better - 61 or 63 for 45nm DualCore CPU ?

[zsamz_]

Which % do you find better - 61 or 63 for 45nm DualCore CPU ?

you gotta test n find out every board n every cpu reacts different

[zfactor]

so anyone using the newest x48 on the x38 or has it shown issues?? has anyone shown the x48 dfi is a better oc'er than the x38? im going to start another system here soon and im wondering if i should wait for p45 or grab a x48. ill wait if x48 is no better oc wise than x38 is.... thanks