i'd say either one will be good... looks like opb got some insane stuff with his gskill at low volts... and the board has the high voltage for the VX...
so if you want 2-2-2-5, get the VX, if you wanna show off 300mhz... get the gskill.
i'd say either one will be good... looks like opb got some insane stuff with his gskill at low volts... and the board has the high voltage for the VX...
so if you want 2-2-2-5, get the VX, if you wanna show off 300mhz... get the gskill.
Yep - but do not foerget the "good old" BH-5, always good to have some reserved.Originally Posted by pik-ard v1.1
On my NEO2 they are maxed out at 275MHz 2-2-2-5 with 3,9 Vcore, but who knows, maybe we will see a better performance on DFI NF4 - even with 4,0 Vdimm.
I'm 99% sure we will - tested 2 kits BH-5 (CORSAIR & MUSHKIN) and both maxed out at exactly the same point: 275 - so I think it's the NEO2 and not the BH-5 which is holding me back.
For higher HTT I use the G.Skill 4800LA.
Previous system:
DFI NF4 ULTRA 0453A3 KOREA CHIPSET / BIOS 510-2FIX / FX-57 0516WPMW@3.62GHZ / 2x256 CORSAIR 3200LLPT BH-5@13x278MHZ 2-2-2-5@3.69VDIMM / MACH II GT@MOD BY BERKUT / ACTIVE COOLING FOR RAM - MOSFETS - GPU RAM / CHIPSET & GPU CORE WATERCOOLED / OCZ POWERSTEAM 600W / BUILT BY ATI X850XT@660/651 - VGPU@1.73-VDD@2.26-VDDQ@2.21 PENCIL MOD / WIN XP 2x80GB SAMSUNG SPINPOINT SP80 SATA - RAID 0 & WIN 2K 40GB SAMSUNG SPINPOINT SP40 IDE BENCH DRIVE / PIC
----------------><------------------
I've got a very noobish question here... Intel person attempting AMD platform for the 1st time and trying to absorb as much as possible in the last weeks... please bear with me...
OPB, I see above that you have the following settings:
292 HTT X 4x LDT = 1168Mhz HTT Data Rate
300 HTT X 4x LDT = 1200Mhz HTT Data Rate
230 HTT X 5x LDT = 1150Mhz HTT Data Rate
I was under the impression that HTT x LDT cannot exceed 1000Mhz HTT Data Rate as instability would arise. I've checked DFI's site (you've guessed what m/b I'll be attempting ) and I can see the NF4 supports up to 2000MT/s or 1000Mhz HTT Data Rate (am I right in concluding this?)...
My question would be how you managed to run your settings as above when you've clearly exceeded the max HTT of the NF4 platform?
What is the max HTT of the NF4 platform?
I appreciate the input and clarification from all
Officially, 1000.Originally Posted by Xcalibur
Unofficially, whatever you can push it to
Some boards work under higher HTT settings than others, and the DFI would certainly be at the top of the list
yeah some boards handle >1000 better, i.e. DFI NF4 i think Oskar said 1200-1250 should be fine
---
This is Errata 94:
"94 Sequential Prefetch Feature May Cause Incorrect Processor
Operation
Description
On an instruction cache miss, the sequential prefetch mechanism may enable the early prefetch of the
next sequential cache line. Under a highly specific set of internal pipeline conditions this mechanism
may cause the processor to hang or execute incorrect code in 64-bit systems running 32-bit
compatibility mode applications.
Potential Effect on System
Processor may deadlock or execute incorrect code.
Suggested Workaround
BIOS should disable IC sequential prefetch by setting IC_CFG.DIS_SEQ_PREFETCH (bit 11 of
MSR C001_1021).
Fix Planned
Yes"
In short, it applies to all 3F,C0 and CG Revisions (Winchesters are D0 revision and should be fine).
It also only is an issue if you run a 64 bit OS with some 32bit applications in compatibility mode.
So you can safely keep it disabled if you don't run a 64-bit OS, otherwise you might want to leave it on for 3F,C0 and CG (e.g. Clawhammer and Newcastle).
Killeroy.
GA-K8NF-9, A64 3200+ Winnie Week 44, 2*512MB Samsung TCCC, Sparkle 6600GT
Thanks for the heads up matt and eva
THX - now we all know.Originally Posted by Killeroy
Previous system:
DFI NF4 ULTRA 0453A3 KOREA CHIPSET / BIOS 510-2FIX / FX-57 0516WPMW@3.62GHZ / 2x256 CORSAIR 3200LLPT BH-5@13x278MHZ 2-2-2-5@3.69VDIMM / MACH II GT@MOD BY BERKUT / ACTIVE COOLING FOR RAM - MOSFETS - GPU RAM / CHIPSET & GPU CORE WATERCOOLED / OCZ POWERSTEAM 600W / BUILT BY ATI X850XT@660/651 - VGPU@1.73-VDD@2.26-VDDQ@2.21 PENCIL MOD / WIN XP 2x80GB SAMSUNG SPINPOINT SP80 SATA - RAID 0 & WIN 2K 40GB SAMSUNG SPINPOINT SP40 IDE BENCH DRIVE / PIC
----------------><------------------
Are those GSkill TCCD LE 2x512 better than OCZ Plat PC3200 Rev 2? Or similar?
ASUS P5Q Deluxe / Q6600 G0 3.2Ghz (400x8) @ 1.31V / 2x2Gb G.Skill F2-8800 PC2-1100 1.90V / Corsair HX620 / Noctua NH-U12P / 1x7200.11 640 Gb / BenQ G2400W / Asus DRW-2014L1T / Windows XP uE 2008.1
Similar, both have TCCD and a Brainpower PCB.Originally Posted by xavicl
"Strive for perfection in everything you do. Take the best that exists and make it better. When it does not exist, design it." - Sir Henry Royce
Nice, i've heard that I could put my plat rev2 to ddr600 with 3.3V, or maybe it's too much voltage for this tccd?
ASUS P5Q Deluxe / Q6600 G0 3.2Ghz (400x8) @ 1.31V / 2x2Gb G.Skill F2-8800 PC2-1100 1.90V / Corsair HX620 / Noctua NH-U12P / 1x7200.11 640 Gb / BenQ G2400W / Asus DRW-2014L1T / Windows XP uE 2008.1
What kinda voltage does TCCD like on these boards? I know normally for 280-300mhz TCCD takes about 2.7-2.8vdimm. But i've been seeing that it takes 2.9-3.1vdimm for same speeds on DFI nf4? BTW I have 2x256 of Gskill LE.
Try it.......Originally Posted by apu673
Previous system:
DFI NF4 ULTRA 0453A3 KOREA CHIPSET / BIOS 510-2FIX / FX-57 0516WPMW@3.62GHZ / 2x256 CORSAIR 3200LLPT BH-5@13x278MHZ 2-2-2-5@3.69VDIMM / MACH II GT@MOD BY BERKUT / ACTIVE COOLING FOR RAM - MOSFETS - GPU RAM / CHIPSET & GPU CORE WATERCOOLED / OCZ POWERSTEAM 600W / BUILT BY ATI X850XT@660/651 - VGPU@1.73-VDD@2.26-VDDQ@2.21 PENCIL MOD / WIN XP 2x80GB SAMSUNG SPINPOINT SP80 SATA - RAID 0 & WIN 2K 40GB SAMSUNG SPINPOINT SP40 IDE BENCH DRIVE / PIC
----------------><------------------
Nope, higher vdimm didn't help, should I be maxing out at 280mhz?
OPB,
Please could you ask Oskar why it takes more voltage on the DFI NF4 with TCCD's than on Neo2 or Neo4 ????
opb whats max. with your fx-55?
mrki
3000+ Winchester
DFI SLI-D
2x512 OCZ EL Rev.2
OCZ 520Watt
2x6800GT Inno3D SLI
heh?you sure? HTT I think by that 1000 he asked for HT not HTT :pOriginally Posted by matt9669
anyway the default iis 1000(200x5) on nF4 800(200x4) on nF3 but both they can go over 1000 1200/1300 nF4 10xx/11xx on nF3
nice thread OPB!!
one Q : how u managed to get high Sandra's Bandwidth efficiency on 512x2 ? you het like .. 91% if im not mistaken
regards
Originally Posted by bito
I don't think so..same ram I need to use 2.85 to do 290, but under DFI SLI-DR..I got this: do msn, everything..super pi 8m, 2.67v
what'd you think?
I think you got handpicked sticks.
Hummm??? my LE's wont even post in the SLI-DR on 2.65v
anyone tried with ocz pc4200EL .....the hynix, that wasnt great on engineer samples...wanna find out if i should purchase more rams...perhaps some Gskill
how about USB trouble? most of my things are connected via usb...and most of them by usb 2.0
Last edited by [G.N.U.]Fragman; 02-12-2005 at 05:26 PM.
E3110//4,1Ghz @ Asus Rampage x48 cooled by NorthQ Siberian Tiger
Mushkin 2*2GB pc8000| Gf 8800GTS-512
OCZ 600W | samsung 226BW | logitech G15
Speedlink medusa 5.1 | MS Habu@nova killer
weeeeeeeeeeeeeeeeeeee at last i got my modd and equip in place and gets it up and running..at least in the bios...
my problem (and i have read a lot on dfi street and opb and so ons guides...but i cant get the board to run stable at any speeds with 2 sticks of dynamite (RAM)
3200+ winny, dfi sli-dr, ocz pc4200el hynix..a lot of air and no oc so far....only thing i can try tomorrow ould be to them into yellow sockets.
E3110//4,1Ghz @ Asus Rampage x48 cooled by NorthQ Siberian Tiger
Mushkin 2*2GB pc8000| Gf 8800GTS-512
OCZ 600W | samsung 226BW | logitech G15
Speedlink medusa 5.1 | MS Habu@nova killer
bump, screenshots seem to be down, could anyone help me im having huge problems with my lp nf4 ultra
I'm not sure what you want a screenshot of but WindWithMe has also done a review so perhaps something in there might help...
http://www.xtremesystems.org/forums/...ad.php?t=62623
...
Bookmarks