Im waiting for a 3845A935 w3520. Whats the best bios for xeons?
Printable View
Im waiting for a 3845A935 w3520. Whats the best bios for xeons?
Hey guys. How do I add QPI voltage on this board?
I cant get this board stable with my I7 920 D0 for the life of me.
1.21 vcore (Proven stable at 1.2)
1.45 cpu vtt
1.08 nb voltage
1.65 dram voltage.
What else can I do?
Could my 32M superpi instabillity problem be related to my SSD or only SP1 ( since SP2 is'nt on danish yet ) Could it be a software issue?
I'm using 5/08 offisial BIOS, and it is working fine with w3520.
I was thinking about SW/driver-problems the other day (after your system passed the memetest), because it had already passed prime95 too. I have seen many BSOD while OCing on a unstable i7-system, but not many errors. An instable i7 would BSOD on you.
If you can pass prime95 and memtest stability tests, and your daily application runs without problems, then you shouldn't care much about Superpi-errors, would i suggest.
[CPU Feature
Set VR Current Limit Max: Enabled
Thermal Management Control: Disabled
EIST Function: Enabled
CxE Function: Disabled
Execute Disable Bit: Disabled
Virtualization Technology: Disabled
***** Logical Processor Setting *****
Intel HT Technology: Enabled
Active Processor Cores: All
DRAM Timing
Channel Interleave Setting: AUTO
Rank Interleave Setting: AUTO
Memory LowGap: 3072M ( Running quad SLi )
DRAM Command Rate: 1N
CAS Latency Time (tCL): 7
RAS# to CAS# Delay (tRCD): 7
RAS# Precharge (tRP): 7
Precharge Delay (tRAS): 20
REF to ACT Delay (tRFC): AUTO
Write to PRE Delay (tWR): AUTO
Rank Write to Read (tWTR): AUTO
ACT to ACT Delay (tRRD): AUTO
Row Cycle Time (tRC): AUTO
Read CAS# Precharge (tRTP): AUTO
Four ACT WIN Time (tFAW): AUTO
Ch1 Round trip latency: AUTO
Ch2 Round trip latency: AUTO
Ch3 Round trip latency: AUTO
tREFI: AUTO
Voltage Setting
O.C. Shut Down Free: Enable O.C.S.D.F
CPU VID Control: 1.39375V
CPU VID Special Add Limit: Enabled
CPU VID Special Add: Auto
Vcore Droop Control: Enabled
DRAM Bus Voltage: 1.665v
DRAM PWM Switch Frequency: Nominal
DRAM PWM Phase Control: 2 Phase Operation
CPU VTT Special Add: AUTO
CPU VTT Voltage: 1.39V
VTT PWM Switch Frequency: Nominal
VTT PWM Phase Control: 2 Phase Operation
CPU PLL Voltage: 1.80v
IOH/ICH 1.1V Voltage: 1.23V
IOH Analog Voltage: 1.20v
ICH 1.5 Voltage: 1.5v
ICH 1.05V Voltage: 1.05v
DIMM 1/2 DQ/DQSTB Bus VREF: -00.0%
DIMM 3/4 DQ/DQSTB Bus VREF: -00.0%
DIMM 5/6 DQ/DQSTB Bus VREF: -00.0%
ADDR/CMD VREF Control: Disabled
x DIMM 1/2 ADDR/CMD Bus VREF: 110
x DIMM 3/4 ADDR/CMD Bus VREF: 110
x DIMM 5/6 ADDR/CMD Bus VREF: 110
CPU QPI Drive Strength: Normal
IOH QPI Drive Strength: Normal
Exit Setup Shut down: Mode 2
O.C. Fail Retry Counter: Enabled
O.C. Fail CMOS Reload: Disabled
PPM Function: Enabled
Turbo Mode Function: Enabled
1 core Max Turbo Ratio: 22x
2 core Max Turbo Ratio: 21x
3 core Max Turbo Ratio: 21x
4 core Max Turbo Ratio: 21x
CPU Non-Turbo Clock Ratio: 20x
* BCLK/UCLK/QPI Controller Settings *
QPI Control Settings: Enabled
QPI Link Fast Mode: Enabled
QPI Frequency: AUTO
CPU Base Clock (BCLK): 200 Mhz
Boot Up CPU Base Clock: AUTO
PCIE Clock: 100 Mhz
DRAM Frequency: 8x
UnCore Frequency: 16x
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
Looking good - yes you might need more vcore in hyper pi for 8x 32m, happened on both my 2x i7 920 3851A368 D0 need 0.032v more vcore than prime95/linx stable vcore. Much different to my 2x i7 920 3836A756 C0 or W3520 3845A935 D0 cpus.
probably the ioh/ich and ioh analogue combined with higher vcore helped :)
200x21 HT off / 1000mhz / 7-7-6-20 / tRFC 60
+12h Prime Blend 5.4GB mem / LinX x 100 5.5GB
set vCore --------------- 1.3375V + vdroop control set 'enabled' (droops)
set vTT ----------------- 1.39V + 0.0375V spec add
set vDIMM -------------- 1.680V
DMM (idle) -------------- vCore 1.344 / vTT 1.421 / vDIMM 1.686
DMM (16K FFT load) ----- vCore 1.310 / vTT 1.435 / vDIMM 1.685
DMM (2048k FFT load) --- vCore 1.312 / vTT 1.442 / vDIMM 1.685
DMM (LinX load) --------- vCore 1.311 / vTT 1.440-4 / vDIMM 1.685
http://i447.photobucket.com/albums/q...77620_PB12.gif
http://i447.photobucket.com/albums/q..._L100_B2_1.gif
Code:CPU Feature
Set VR Current Limit Max: Enabled
Thermal Management Control: Disabled
EIST Function: Enabled
CxE Function: Disabled
Execute Disable Bit: Disabled
Virtualization Technology: Disabled
***** Logical Processor Setting *****
Intel HT Technology: Disabled
Active Processor Cores: All
DRAM Timing
Channel Interleave Setting: 6 Way
Rank Interleave Setting: 4 Way
Memory LowGap: 1536M
DRAM Command Rate: 1N
CAS Latency Time (tCL): 7
RAS# to CAS# Delay (tRCD)7
RAS# Precharge (tRP): 6
Precharge Delay (tRAS): 20
REF to ACT Delay (tRFC): 60
Write to PRE Delay (tWR): AUTO
Rank Write to Read (tWTR): AUTO
ACT to ACT Delay (tRRD): AUTO
Row Cycle Time (tRC): AUTO
Read CAS# Precharge (tRTP): AUTO
Four ACT WIN Time (tFAW): AUTO
Voltage Setting
O.C. Shut Down Free: Enable O.C.S.D.F
CPU VID Control: 1.33750
CPU VID Special Add Limit: Enabled (Disabled = no droop)
CPU VID Special Add: Auto
Vcore Droop Control: Enabled
DRAM Bus Voltage: 1.680v
DRAM PWM Switch Frequency: Nominal
DRAM PWM Phase Control: 2 Phase Operation
CPU VTT Special Add: +0.0375
CPU VTT Voltage: 1.39v
VTT PWM Switch Frequency: Nominal
VTT PWM Phase Control: 2 Phase Operation
CPU PLL Voltage: 1.80v
IOH/ICH 1.1V Voltage: 1.11v
IOH Analog Voltage: 1.10v
ICH 1.5 Voltage: 1.5v
ICH 1.05V Voltage: 1.05v
DIMM 1/2 DQ/DQSTB Bus VREF: -00.0%
DIMM 3/4 DQ/DQSTB Bus VREF: -00.0%
DIMM 5/6 DQ/DQSTB Bus VREF: -00.0%
ADDR/CMD VREF Control: Disabled
x DIMM 1/2 ADDR/CMD Bus VREF: 110
x DIMM 3/4 ADDR/CMD Bus VREF: 110
x DIMM 5/6 ADDR/CMD Bus VREF: 110
CPU QPI Drive Strength: Normal
IOH QPI Drive Strength: Normal
Exit Setup Shut down: Mode 2
O.C. Fail Retry Counter: Enabled
O.C. Fail CMOS Reload: Disabled
PPM Function: Enabled
Turbo Mode Function: Enabled
1 core Max Turbo Ratio: 22x
2 core Max Turbo Ratio: 21x
3 core Max Turbo Ratio: 21x
4 core Max Turbo Ratio: 21x
CPU Non-Turbo Clock Ratio: 20x
* BCLK/UCLK/QPI Controller Settings *
QPI Control Settings: Enabled
QPI Link Fast Mode: Enabled
QPI Frequency: 4.8GT/s
CPU Base Clock (BCLK): 200 Mhz
Boot Up CPU Base Clock: 185 Mhz
PCIE Clock: 100 Mhz
DRAM Frequency: 10x
UnCore Frequency:20x
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
nice work cheapseats... another tweaking master joins the thread :D
100 pass linx @DDR3-2000Mhz 7-7-6-x with trfc 60 :cool:
Isn't that crazy that only around 70C @ 1.32v on AIR :D
I have one of these mobos. I have not tweaked the bios yet. Running it at 175x91 with a 920 C0 chip. I need to speed it up or see about dropping the voltage on the cpu.
PRJ
Nice to see you able to get HyperPi to run. It's normally my first sop for stability testing once in Windows on the i7. It's quick and gets you at least 80% towards full stability.
@ eva2000
Well, thanks for your tips on where to go
Now I can start optimizing, now that i finally have a stable setup :)
Would you be satisfied with these Voltages for 24/7 use?
@ Cheapseats
Set VR Current Limit Max: Enabled Does this not need to be disabled when OC..?
yes very good for 24/7 - i use 19x211bclk 844mhz 6-7-6-x 24/7 hehe
Set VR Current limit max = enabled for best clocks - tip #4 http://i4memory.com/102537-post3.html to disable TDC/TDP protection :)
PRJ, what kind of block are you using on your 'Northbridge'? Looks clean! :)
I finally got to building the w3520 system yesterday but couldn't even get win7 to install....it kept on freezing during the percentage part...sometimes it would make it to 30% and othertimes not even 1 percent....I had everything in the BIOS set to defaults and made no changes...I put my Areca 1231ML on the 2nd PCI-E 16x slot (where your 2nd SLI or crossfire videocard would normally go) and my PCI-E X-Fi on the x16 yellow slot which is supposed to run at 8x....I didn't update the BIOS yet and am thinking that is the problem, but could anyone help me figure this out?
Almost forgot - I'm running a single 4870x2 videocard.
Ooooh, shiny! :p:
But no I wouldn't buy a Mosfet block, I particularly like the fact that one can cleanly cool the whole board with just that one block. Low pressure drop and temps are also fine, plus it looks clean! :up:
Edit: It would be also interesting to know what mounting kit exactly it was that you used. The one supplied with the blocks or self-brewed?
Hey guys,
seems I "fixed" my issue from earlier. After days of testing I finally decided that the mobo doesn't like the ram I was using (2 kits of Gskill F3-12800CL8T-6GBPI)
The DFI seems really picky about ICs, so just a heads-up for anyone looking to buy that ram for the DFI.. don't. It really only likes HCF0 and MNH-E.
memory tips http://i4memory.com/102537-post3.html - different ic types have varying cpu vtt to vdimm optimal differential voltages as well as dram / pwm switch values. Just a matter of finding which values are for which memory ics and cpus. Micron ics like decrease 15% while samsung hcf0 and elpida like nominal or increase 15%. Of course my experiences might not translate the same for you.
Thanks eva, but I have tried that... VTT from 1,21 to 1,45V, Vdimm from 1,55 to 1,71V, with all possible relations. I couldn't run LinX with 11GB of ram for longer than a few minutes no matter what. Not at close to 1600Mhz anyways. It does work at 190 BCLK with 6x dram.. 1150Mhz ram, no thanks.
Besides one shouldn't have to start tweaking switch values for a meagre 1600Mhz ram - I mean come on, that's even officially supported.
What do your timings look like? I could be way off base here but I've noticed that a couple of my mem kits are very sensitive to voltage in this board. One in particular, some gskill pi 1600, gets weird with 1.65 but if I go to 1.635 they are solid.
If that's the 920 I sent you, while it will overclock pretty nicely, it really isn't going to use low volts like some. Iir, temps should be reasonable even using 1.45v or more though. I had pretty much decided that all my C0's were 3.8-3.9 crunchers and no cherries. I didn't really do a hell of a lot better with the batch of D0 xeons I bought to replace them. I'm traditionally very unlucky when it comes to buying cpu. :shakes:
Finished installing my UT T3H8, did some quick runs to get the hang of this board. It is my first DFI board, and my results have not been very good yet, but once I get the hang of it, should be decent.
First I tried for max Ghz, 4.9.. nearly.
http://valid.canardpc.com/show_oc.php?id=583098
Then tried 32M, no windows tweaks, ect.
Saved this 32M run, and took off for work, lol I guess you will have to guess the Ghz. ROFL.. http://i18.photobucket.com/albums/b140/hoss3430/1-2.jpg
Trying to get Prime 95 stable now, has not gone well so far... hopefully I will get the hang of the board soon.
Whats the max tolereted temp on System in SG, I'm at 66 under Prime95, is that ok for testing 4 hours...?
Maximum stable PWM temp is around 110-115C. Maximum I'd recommend is 85-90C, so you're fine ;)
These suckers do get hot.
w3520 has been playing along so smoothly with the MB and RAM that i had to give a shot on "all default" setup.
w3520 @4.01GHZ (211x19) 1.15v :) 1 hour primestable
Everything else on stock/default, including VVT and NB core :D
http://img269.imageshack.us/img269/1...me1hour.th.jpg
VVT is set to 1.21v in BIOS but the board set it to 1.23v under load, still cool.
Same setup (with "one notch" higher VVT and NB core) has passed 11 hours prime before.
This setup is running so smooth and :cool: that it is very tempting to discard my @4.2GHZ (and attempts for higher), and use this for 24/7, but I haven't decided yet.
Yes it is the one that I got from you. Still on the bios that it came with. Have not played with it much. Got it crunching stable and let it go.
http://i683.photobucket.com/albums/v...3/settings.jpg
Ioonym: I feel with you, buyed a Xeon 3520 becouse everyone i know got great oc, but do you think i did..plæ:S
Some of us are doomed:P
anyone got tips for this board with 12gbdominator GT 1866 + 920 d0? mine is having very bad stability issues and I have used ever combination of high-low/low-high/low-low Vcore and Vtt I can think of... I have had to back off to 3.6ghz no turbo :shakes:
Ok, this is ridiculous...
Has anyone had issues with the board "Hissing"?
I'm on my second HD 4890. Any time the GPU is put under load, the system 'hisses'. I was convinced that it was the GPU, as when I put an old 3870 into the system, under GPU load, no hissing.
Is there any chance that it is the PWM on the motherboard hissing?
EDIT: Nevermind, it appears to be my PSU that is hissing... yay, second RMA! :(
Ahhhhh I finally got it stable at 4.0 ghz (Crosses fingers) this board has been amazing. Very good bios and no problems as of yet.
I got 4.0ghz stable at 1.21vcore, 1.4 htt, 1.23v nb and 1.65 dram. So far so good. Temps are great all around and no signs of any problems.
I have problems with GTX295 quad SLi on this board..
The funny thing is that quad SLi is working fine in Vantage, but not in 03, 05 and 06 -running on one or two GPUs, i have tried vista 32 and 64bit and win7 32bit, no change.
As u can see here, quad sli is working fine in Vantage http://service.futuremark.com/compare?3dmv=1126636 if i try to run one of the other benches afterwards the score is... boring..
I dont understand why it works fine in Vantage but not in the others ? Is it a know bug ? (Its my first time running quad sli)
What exactly does not work? You only get a low score in older 3Dmark apps?
Well, that's because 06 and older don't use more than 1,5 cores in all GPU tests, hence you get beaten there by a 6Ghz Dualcore. If you want high scores in older 3dmark get your CPU to 5Ghz :D
By the way, what memory lowgap setting are you using? I found that 2560MB works best, along with 2048MB.
eva, i need your help
with this mobo and an i7 920 when i set in the bios CPU VTT Voltage at 1.61 i see error code 88 on the display and the mobo doesn't boot
if i set CPU VTT Voltage at all others value the mobo boot.
if i set with the special add a vtt voltage higher than 1.61 the mobo boot...
i need more than 1.61volt on vtt because i want use mi D9-GTR at 2.2/2.3volt
any suggestion???
thanks
edit: i've 428 pre-beta bios
Never saved the scores, but in 05 canyon started at ~100fps with SLi, and ~200fps with a single gtx295, so that must be a 1 out of 4 gpu's in use when quad sli is enabled. With SLi it should start around 300fps
I cant test again because i only got one card with stock cooler, the other two is mounted with Tek 9's :) and im out of ln2 atm.
Hm no clue about 05 but in 06 it went over 250 fps where you see the canons of that airship, 4Ghz i7 and stock GTX 295's with 8x AA...
I'm running the X58 JR, and can get to windows at 4.0, but, on load the SYSTEM temps are thru the roof; just watching smart guardian as it slowly creeps to 100 and then reboots. If I can get past the temps, I think I can easily do 4.0 on air.
Running a 285 now, had quad-sli 295's, and didn't have any problems on mild overclocks (166 x 20). Can't do too much testing, cuz my work comp is my oc comp atm heh.
are all you guys running extra fans on the NB and/or mosfet area or what?
Hmm I can get this board to pass like 12+ hours stable prime small ffts. But as soon as I run "blend" it bsod's.
Here are my settings so far.
201 BCLCK
20x ratio
1.216 vcore (Under load)
1.4v CPU-VTT
1.23 NB voltage.
1.64-1.66 VDIMM Voltage.
I dont understand what else I can do to make it more stable.
Could be any number of things... too low Vcore (prime small isn't really working like it used to, blend/large is way more stressful on i7 than small FFT), or wrong VTT/Vdimm relation (try lower VTT unless you run the mems at 2Ghz or something)
I lowered the VTT and got a BSOD almost instantly. What Stress programs do you suggest I run?
Prime blend is best. Up your vcore or down with bclk. As you up the vcore, my guess is you can lower the vtt some. YMMV.
LinX with max mem is almost as good as prime blend, and alot faster to test. I use LinX to dial in approximate oc, and prime blend to verify.
You bet, active cooling over NB should be standard on this board. The system can get pretty hot under stress, but a good fan over NB ( + flame freezer on the back) will keep the system, NB and SB under control. I'm using Intel's stock CPU-fan-mod, LOL, with auto RPM, and it is doing a good job.
Blend would stress your RAM, and you should look at timing/deviders or Vdimm, and try different VVT too, i would suggest.
You guys think it is madness to go for primes stability at BCLK 221? It did BSOD on me after 30 minutes, anyways :(
w3520 @4.2GHZ (221x19) 1.21v, 20 minutes prime
http://img25.imageshack.us/img25/6238/prime10.th.jpg
BSOD after 30 minutes.
Couldn't prime more than 30 minutes, but it was fun trying anyways.
Madness? Not at all. Did this on a C0. Never tried a higher multiplier.
click for larger image
http://www.edgeofstability.com/image...prime_64_s.jpg
I still cant get this board stable. Turbo mode doesn't work also. My patients are wearing thin.:(:(
Yeah, 225 bclk is no problemo on the most of these boards:D
Praz: Realy nice;)
@ Praz >> Awesome running for over 8 hours on Prime, man.
Reason is mentioned 2 posts above.
Thanks guys but there's really nothing I can take credit for. It's all CPU and board.
emertX
Settings and screenshots can be found in posts #24 and #25 of this thread.
http://www.csd.dficlub.org/forum/sho...5034#post75034
Guys read all the links and info in 1st post of this thread - alot of answers and hints in those threads ;)
yeah PWM/SG System reported temps - the pwm is rated to 100+ C IIRC, but for overclocking you want it below 75-85C which can be done easily following the tips outlined in first post link for MOST cpus unless you have a high leakage low vcore/high temp cpu http://i4memory.com/f80/t3eh8-very-h...m-temps-18015/ and tips #31 and #32 at http://i4memory.com/102537-post3.html
nice slightly better than my Xeon W3520 3845A935 you batch is :)
lots to read for memory tips at http://i4memory.com/102537-post3.html and http://i4memory.com/f55/intel-core-i...ide-x58-13051/ and http://i4memory.com/f80/announcement...y-eva2000.html
I never used 12GB as don't have that many matched modules but did mix and match 6x1GB of samsung hcf0 and micron d9gtr (in my sig) and key to mixing or populating slots on any platform is matching the module with the dimm slot. Yes time consuming to test each module in each dimm slot and rate them then match them in the right dimm slots - but they're always some benefit if tiny at best.
I'd test each module individually as it can reveal alot about a so-called dual or triple channel matched kit i.e. http://i4memory.com/f83/3x2gb-kingst...3-tests-18365/
weird issues unfortunately never used gtx295 so can't help - didn't have any problems with 2x HD4870x2 CF though
Win7 ultimate 64bit RC
http://fileshosts.com/intel/DFI/DFI_...03_aero_tn.png
not all cpus are created equal - 88 seems to happen if you push too far on either memory or cpu i.e. too high bclk or unstable settings for high clock bclk etc. Could be cpu IMC doesn't like high CPU VTT ... check the first post in this thread for links related to prime95/superpi/linx stability and read that one and play with IOH/ICH voltages too as you may not need such high cpu vtt or you may need high cpu vtt in conjunction with ioh/ich voltage tweaking to get higher depending on mem clock and cpu bclk.
Again nice read for SYSTEM/PWM temps which applies for all DFI boards even P35/X38/X48 and X58. Measures to reduce these temps http://i4memory.com/f80/t3eh8-very-h...m-temps-18015/ - ultimately it comes down to air flow around that area. Some folks just don't see this on other boards as alot of other boards don't report the PWM temps.
first post links to first 4 linked threads will be useful for you ;) particularly http://i4memory.com/f80/i7-920-d0-df...ability-16617/
proper testing for i7/x58 is using prime95 custom blend or large fft with 5400mb mem allocation (for 6GB) as i7 mem controller is on cpu now :) Small fft won't stress i7 systems as much as hyper pi 8x 32m pi, linx all memory test or custom prime95 tests i mentioned.
I don't have 12GB of matched memory either. Running 6GB of OCZ 1800 Reaper and 6GB of 2000 speed HyperX together worked well for me at 1800. Posts #22 and #23.
http://www.csd.dficlub.org/forum/sho...?t=8575&page=3
thanks praz, gonna work a bit off of that template :up:
Some have been reporting problems with Prime Blend, so i had to check it out on my favorite low-volt setup.
VVT: 1.23/1.24v (BIOS/Load)
NB Core: 1.13/1.12v
Vcore: 1.20/1.15v :cool:
RAM: 1266MHZ 6-7-6-22, 1.65v
Everything else on stock/default
W3520 @4.01GHZ(19x211) 1.15v, 8 hours Prime Blend stable
http://img7.imageshack.us/img7/6503/...d8hour2.th.jpg
I have massive problems with this board....omg there's no chance to reach the Vcore I had with my Gigabyte Extreme!!!
My CPU is pissing me off right now. I'm trying really hard not to spend more money on a D0, but this thing needs 1.42 vcore to hit 4GHz but will run at 3.8 at stock volts (didn't minimize voltages at 3.8GHz yet).
Anyways, adding my adaptec 5805 RAID card into the system made me bump my vcore from 1.41 up to 1.42v
I submited a RMA for this board. It would BSOD and come up "Board cannot detect multiple cores, Will switch to UNI mode" or some BS like that then it gets stuck at C1 error and wont boot.
Two complaints, my board reports all voltages .05v over and I wish this board could run 3DMark apps higher than 229 BLCK. I hate to jump the Evga Classified bandwagon, but this board is really holding me back. Still though.......I LOVE THIS BOARD!!!
http://img.techpowerup.org/090612/3D...82 Tri sli.jpg
I got my batch 3845A935 look like can't have too much expectations on it~
Yeah, that's right, but what is with the VTT??? For 21x200 I must take 1,32 in BIOS and 1,36 in Win!
Hey foks
Having trouble clocking my i7 920 any higher, and not sure if it's the board or the chip. Current BCLK limite seems to be about 182, and nothing I've tried has managed to get anything stable past that, or to even boot at higher than 187. I've tried lowering CPU multi to 12x, disabling HT and Turbo, vCore up to 1.5, vTT up to 1.45, GMT and RAM at auto. Any ideas? Temps seem to be OK, around 50 for system temps and CPU temps not an issue at 182*12. If this was LGA775 or AMD I'd say that I've got the worst-clocking T3eH8 ever, but I'm still relatively new to the whole i7 thing, so I'm not sure if somehow the CPU is limiting mobo BCLK even when all the multis are turned right down.
Any ideas?
It's one of the originals, D0, right? RAM is in the sig - OCZ Reapers, 1866 CAS-9 1.65V. Tried dropping RAM speeds right down, as above, but doesn't seem to make any difference even when massively underclocked. By spectrum, do you mean the various "spread spectrum" stuff? That's all disabled.
Thanks for the help :)
What's the debug LED shows when you get stucked?
For BIOS settings -
CPU Features
- all disabled except
- Execute Disable Bit - Enabled
- Virtualisation - Enabled
DRAM
- all Auto
- lowgap 1536M
Voltage
- all stock except
- vCore 1.4 (tried both setting and with using VID special Add)
- vTT 1.4
All other Genie settings stock except
QPI - 4.8
DRAM - x06
Uncore - x12
BCLCK - 185
This boots fine. For same BIOS settings but BCLCK 190, will not boot - locks up showing CF code on hot reboot, occasionally gets to 26 on cold boot (with power unplugged), but then reboots and locks again at CF.
Edit: If it's relevant, I once saw 79 displayed - but the board immediately rebooted and I haven't seen it again.
CF can mean its the memory, i used a oem-sett before i buyed Kingston and if i tried to oc then i just get CF.
26 did i got to then.
Do yuo have the chance to try another memory-sett?
since i didn't find any info or faq on the first page let me ask about pwm temp one more time :)
what the max pwm temp i have to worry about? now i have waterblock installed on the nb and a little airflow near pwm area.
but running 180 bclk with all default voltages pwm temp goes upto 65C with 30C ambient (idle pwm temp is 49C).
does pwm temp 65-70C sound ok? what the dfi official recommendation about it?
thx for the info..
If those were the first i7's then sure :p: I can never remember which was which. Bought second-hand on these forums, batch on the box 3839A628.
Unfortunately I don't - I barely had enough for my current RAM.
Is it possible it's the RAM, though? If my board's doing what I'm telling it to, at the settings at which my mobo won't boot the RAM should be running at DDR3-14000, 2T, CAS 10-10-10-30-74 at 1.65V, when it's rated (and will run perfectly fine at with 180 BCLK) at DDR3-1800, 2T CAS 9-9-9-26-74 at 1.65V. And looking at the first T3eH8 thread, and over at the OCZ forums, folks seem to have no problems reaching 225 BCLK with essentially the same kit (although that kit is 1800 CAS-8, and mine's 1866 CAS-9).
Does anyone know what BIOS setting are required to alter to get this interleave option(s) working for memory? Everytime I enable it, my PC just reboots about 30 seconds after I get into Windows...I'm almost positive I saw people here use it successfully, but can't remember where. If you know how to play with these settings, please help me out.
@SoulsCollective
CPU batch# seams to be the most important component for reaching high BCLK. Probably because of the integrated memory controller (which is inside the CPU).
Different MB brands seams to play a role too, but DFI has proven to be capable of handling high BCLK on both i7 CO, and the new w3520 (D0) too.
RAM wouldn't be the limiting factor on how high BCLK the system can run stably, because you can always try to loosen up the dividers/timing, unless you are going for 5x (>2000MHZ RAM).
I suspect your CPU is a crappy one, and probably that's why the first owner has sold it. Can you borrow/try another CPU?
I was sort of thinking along the same lines - I'm new to this whole i7 thing, but I thought that might be the case. If this was 775, with no IMC on the CPU, and I saw this behaviour - crappy FSB, unable to reach higher even with ridiculously high volts and low multi on everything - I'd put it down to mobo. But if the IMC in the CPU is capable of restricting the BCLK the mobo can set? I dunno, can it?
I just don't want to fork out for a new CPU if the motherboard is the crap component in the mix. Unfortunately I'm the only one in my group of friends with an i7 - I'll have to have a grab of the next one from the next i7 client build I do and have a quick test.
Can anyone else chime in here with a confirming opinion?
Yeah I've been looking around at various threads and the spread of poor performers and TEMPS [often occur together] is insane.
You can sell it on OCAU as not everyone is interested in oc'ing.
I'm not sure about the mobo being the guy to blame. So far, all the DFI haters I've seen got multiple DOAs. I'm wanting this mobo but the UD5 is 75 bucks cheaper at Umart. . . . oh decisions . . . Lucky I have till the end of July to decide :D
I may end up doing that at this point. Been looking at W3520s. First I'll try and borrow a CPU off of kayl to test with next time I see him just to confirm it is a mobo issue.
I don't have many settings saved for the 1866 CAS9 Reaper. But the following link is 6GB of that memory along with 6GB of 1800 CAS8 memory running 200x20, memory at 1600 with a C0 processor. Posts #14 and #15. Might help somewhat.
http://www.csd.dficlub.org/forum/sho...?t=8575&page=2
I saw your posts in the first DFI thread, and over at OCZ (linked to above), and indeed over at DFI Club - helped me work out the proper subtimings, thanks :up: But alas, tried those already, and nothing - board just does not boot with any settings I've tried so far.
Of the newer 4/28 and 5/08 betas i've only tried 4/28 beta haven't tried 05/08 beta or official bioses. 4/28 beta gave new life to memory related cpu overclocking especially with Elpida 50mm hyper ICs. But 4/28 and 05/08 are meant to be totally different from earlier bioses and from testing 2/17 and 2/27 and 04/28 it's true. What settings work for me on 2/17 and 2/27 in terms of voltages differed slightly from what works on 4/28.
@Systemshock2 err SoulsCollective,
Did you try the 18 or 19 multiplier with turbo disabled? Do you know if your ram has samsung hcfo's and what voltage do you have on the ram? These chips seem to like more volts on this board especially 6GB's. My board over reports voltages by .05v so if your adjusting based on what it reads in bios or windows don't. Typically whatever you set in bios is actual voltage when measured. Lastly pull one stick out, run with green slots close to cpu. If she runs then u need to apply more volts to yor ram like 1.710v or so.
Yessss, got my board, a W3520 (3848A365), a set of OCZ3P160LV6GK and a LGA1366 bolt-thru-kit for my TRUE in the mail today. Hopefully I can start playing this weekend as I'm down with bronchitis atm and the the drugs make my head all fuzzy and I'm worried about setting something on fire here.. :)
Yep, no difference.
I've tried the latest and the next latest, and the original shipping BIOS. Sorry, I don't remember the dates off the top of my head - I know for sure I had the same 180 BCLK limitation on each of them, tried OCing when I first got the board, then had lots of assignments due and left it, then updated and tried again, then had exams, and have just finished exams, updated to latest, and tried again.
:p:
Wish this machine of mine would emulate Her Glorious Majesty - I could do with the Hyper-Optimised, although I'd leave the sentient.
I tried dropping to 19, but still couldn't hit over 185 BCLK bootable, and as above 12x is also no go. Is it worth trying 18?Quote:
Did you try the 18 or 19 multiplier with turbo disabled?
Sorry, I have no idea what chips and ICs are being used - heatspreaders refuse to come off. As above, RAM is the 6GB OCZ Reaper 1866 CAS-9 kit, from Newegg USA via a friend, bought late April. MEMTEST at 185*10 for DDR3-1850 at CAS-9 rated timings and 1.65V BIOS set passes seven consecutive runs, so I don't think it's a RAM stability issue - as above, I've even tried dropping to the lowest memory multi and CAS-10, but the board still refuses to go above 185 BCLK.Quote:
Do you know if your ram has samsung hcfo's and what voltage do you have on the ram? These chips seem to like more volts on this board especially 6GB's. My board over reports voltages by .05v so if your adjusting based on what it reads in bios or windows don't. Typically whatever you set in bios is actual voltage when measured.
I've tried each stick of the pair in DRAM #2 (green slot mentioned), as well as in DRAM #1 (yellow slot next to CPU), board will boot at 185 but still refuses to reliably at anything higher than 186/7ish. Never boots at 190+. So no difference from one DIMM to three, and no difference that I can see between various slots.Quote:
Lastly pull one stick out, run with green slots close to cpu. If she runs then u need to apply more volts to yor ram like 1.710v or so.
Thanks for the help guys, appreciate it! Feels like I'm banging my head against a 185-brick wall here - this bloody board just refuses to shift.
INCREDIBLE...
cpu 4ghz
cpu vtt 1.21v crash after 3 hours of linx
cpu vtt 1.34v crash in a second
eva help me and razer69
Soulscollective have you tried your cpu on a different board or a different cpu on your board? I think this will somehow solve the root of the problem.