[First Look] DFI 680i LT NF-680i LT SLI (56K Warning~~~)

[xTr33m sammich!]

old news? IDK, just thought I'd share this. Just used winflash184 to update from beta504 to beta521 bios without error.

posted by virtualrain on his blog
At this time, Winflash does not work. It throws an error saying "Onboard BIOS not Award BIOS".

[virtualrain]

old news? IDK, just thought I'd share this. Just used winflash184 to update from beta504 to beta521 bios without error.

Yeah, it should have been qualified to say it doesn't work on Vista x64.

[C-N]

Right after a full days worth of messing about I have found out what was causing the shutdowns it was the SB voltage.

I initially upped the NB & SB voltage almost together with the CPU drive set on strong as its the only way the board would boot. I have now found by switching the CPU drive to weak I can significantly reduce the SB voltage needed yet it will still allow me to up the NB voltage.

I have run 12 attempts at 462.5fsb x 8 = 3.7GHz today all ending between 20 minutes & an hour & a half with an Orthos fatal error rather than a shutdown.

From this I come to the conclusion that 3.7GHz isnt going to happen with this E6400 on air despite it being 5 hours+ Orthos stable on small fft's.

I reduced the fsb to 1825 456.25 x 8 = 3650MHz & it ran Orthos for 8 1/2 hours on blend until I stopped it.

Regards

Craig.

[xTr33m sammich!]

thats really odd..
I had the same problem with the same settings and orthos going BSOD on my arse at 462.5x8 3.7ghz....
what are your bios settings, got mine stable for 8hrs with a significant Vcore increase.. it will still bsod at that speed if i leave it going for 10-20hrs.. I havent played with the gtl ettings at all, maybe I should.

[Borgod]

Nice find C-N, cheers

[C-N]

thats really odd..
I had the same problem with the same settings and orthos going BSOD on my arse at 462.5x8 3.7ghz....
what are your bios settings, got mine stable for 8hrs with a significant Vcore increase.. it will still bsod at that speed if i leave it going for 10-20hrs.. I havent played with the gtl ettings at all, maybe I should.

My settings are nothing special at all but here they are for 3650MHz & I think I could get 3.7GHz Orthos stable with even more vcore but as you can see I am already piling it on.

I have had a little play with vtt & gtl but didn't find anything that helped me but I haven't finished playing with it yet.

One more thing even with the ldt vtt & sb dual voltage all on default minimum the board will still run PCI-E 130mhz & SPP 300MHz quite happily.

Regards

Craig.

Genie BIOS Settings:

Voltage Control:

CPU VID Control................ - 1.60000V
CPU VID Special Add............ - 101.55%
CPU Drive Strength............. - Weak
DRAM Drive Strength............ - Weak
DRAM Voltage Control........... - 2.18V
NB Voltage Control............. - 1.57V
SB Voltage Control............. - 1.62V
SB Dual Voltage................ - 1.52V
LDT Voltage Control............ - 1.22V
CPU VTT 1.2V Voltage........... - 1.21V
GTL REF Voltage Control........ - Disable


FSB & Frequency Seting:

FSB-Memory Clock Mode.......... - Linked
FSB-Memory Ratio............... - Sync Mode
FSB (QDR), MHz................. - 1825

=====

PCI Bus, Slot 1 MHz............ - 100
PCI Bus, Slot 2 MHz............ - 100
SPP <-> MCP Ref Clock, MHz..... - 200
LDT Frequency.................. - 5x
CPU BSelect.................... - BSel_7
CPU Clock Ratio Unlock......... - Enabled
CPU Clock Ratio................ - 8x


CPU Feature:

PPM Mode....................... - Native Mode
Limit CPUID MaxVal............. - Disabled
C1E Function................... - Disabled
Execute Disable Bit............ - Enabled
Virtualization Technology...... - Enabled
Core Multi-Processing.......... - Enabled


Memory Timing Setting:

Memory Timing Setting.......... - Expert
CAS Latency Control (TCL)...... - 5
RAS# to CAS# Delay (TRCD)...... - 5
Row Precharge Time (TRP)....... - 5
Min RAS# Active Time (TRAS).... - 15
Command Per Clock (CMD)........ - 2T

=====

Bank To Bank CMD (TRRD)........ - 4
Row Cycle Time (TRC)........... - 28
Write Recovery (TWR)........... - 6
Write -Read Commad............. - 10
DRAM Refresh Rate.............. - 7.8uS

[Spazilton]

Ive finally gave up with the throttling issue on this board with Quads. No matter how much I cool the PWM, anything over 1.3875 vcore will start going back and forth between 6 and 9 multi on my Q6600 after around 5 minutes of Quad Prime Small FFT.

Core temps dont exceed 65c fully loaded at 1.425. It happens much quicker if I remove my PWM fan.

I know this quad has more in it. But I am just sick of not having my main rig up, (First Crappy IN9-32X, Blown BX2, Blown Gskill Memory, etc..) I'm sitting at almost 4 months on this rebuild cycle.

I put my E6600 back in, got it to 3650 very easily. Much easier and at lower voltages than my IN9.

I think I am just going to dump the Quad for now, and rebuy one sometime later this year when the P35 chipsets mature a bit.

[vincey2kr1]

Ya the PWM heatsink is just a really crappy painted aluminum heatsink, its a shame DFI went really cheap quality on it. Mine are starting to make a hissing sound during quadprime 95 and shutting off.. I've got 2 fans trying to blow over them

On the cpu throttling download rightmark cpu clock utility and you can disable the throttling there.

[Jobeo]

anything over 1.3875 vcore will start going back and forth between 6 and 9 multi on my Q6600 after around 5 minutes of Quad Prime Small FFT.

i feel your pain ... i have a quad (qx6700) and the experience can, at times, feel very un-rewarding with this board.

question for ya though what speeds are you trying to reach going above the 1.3875 vid? like, what are you stable at 1.3875 which is forcing you to up the volts...

[Jobeo]

Ya the PWM heatsink is just a really crappy painted aluminum heatsink, its a shame DFI went really cheap quality on it.

i have a theory... remember how the PWM temp was mis-labeled in all the BIOS's except the most recent? i think when they were designing the board they were relying on those reported temps and thinking the PWM was fine...

otherwise i don't see how they would have released the board with such a bad solution. the HS does look quite cheap and they've sandwiched it in-between a bunch of components... there's no way it could be getting decent airflow the way its designed which is why people have to rig fans on it...

[virtualrain]

Right after a full days worth of messing about I have found out what was causing the shutdowns it was the SB voltage.

I initially upped the NB & SB voltage almost together with the CPU drive set on strong as its the only way the board would boot. I have now found by switching the CPU drive to weak I can significantly reduce the SB voltage needed yet it will still allow me to up the NB voltage.

Craig, So are you saying the SB voltage being too high was causing the shutdown? At what point does the SB voltage cause this?

Also, I'm not sure I understand your SB voltage comments... on one hand, you are saying you can significantly reduce the SB voltage, but in your settings in your next post it's at 1.62V. a) that's not significantly reduced (although I'm not sure where you had it before) and b) 1.62V for the SB is what Tony's overclocking guide recommended.

Another thing, you say running weak CPU drive, allows you to up the NB voltage, but according to your setting, you are running 1.57 which after the droop is about 1.45 actual... not that very much for 450+ FSB.

Finally, Tony had this to say about CPU drive strength...

If the CPU needs Strong Drive you will have a low to middle NB voltage setting.
If the CPU needs Weak Drive you will have a middle to higher NB voltage setting

It sounds like you agree with this, although I still maintain your NB voltage is not very high at all for the FSB you are running.

Some additional clarification on what you have found would be helpful... Thanks!

[virtualrain]

BTW, have you guys seen this (DFI X38 board)? Does it appeal to any of you?

[Raja@ASUS]

I'm sure it will appeal to many...

[C-N]

Craig, So are you saying the SB voltage being too high was causing the shutdown? At what point does the SB voltage cause this?

On my rig 1.74 - 1.76v

Also, I'm not sure I understand your SB voltage comments... on one hand, you are saying you can significantly reduce the SB voltage, but in your settings in your next post it's at 1.62V. a) that's not significantly reduced (although I'm not sure where you had it before) and b) 1.62V for the SB is what Tony's overclocking guide recommended.

I have been up to 1.76v in the attempt to find stability so reducing to 1.62v imo is significant. Also 1.62v is one maybe 2 clicks over what I need to run stable.

I found Tonys guide interesting to read but that was it. None of his findings or settings have benefited me & my rig this time.

Another thing, you say running weak CPU drive, allows you to up the NB voltage, but according to your setting, you are running 1.57 which after the droop is about 1.45 actual... not that very much for 450+ FSB.

My rig would not boot over 1.51v NB with strong CPU drive strength unless I had the SB voltage over 1.7v

My 1.57v atm is one or 2 slicks over what I need for 450fsb too.

It sounds like you agree with this (Tony's guide re cpu drive strength/nb voltage) although I still maintain your NB voltage is not very high at all for the FSB you are running.

I do agree with Tony with an E6400 but wonder how this may differ with a quad core. I notice you are running high v NB with your quad @ 400fsb my E6400 rig will run 490fsb (thats as far as I have taken it) with lower NB voltage than you are using.

Regards

Craig.

P.S Sorry to post a slow reply but I don't get email notifications from this forum although I have it enabled it just don't work for me.

[C-N]

BTW, have you guys seen this (DFI X38 board)? Does it appeal to any of you?

I will take one put my name on the list

Proper Caps on this board wonder if they will still be there at retail launch ?

Craig.

[Wiggy McShades]

BTW, have you guys seen this (DFI X38 board)? Does it appeal to any of you?

that makes me have buyers regret for getting the 680i from dfi, why would you be so cold hearted to post that in here

[owcraftsman]

Source: www.pcgameshardware.de

[virtualrain]

I do agree with Tony with an E6400 but wonder how this may differ with a quad core. I notice you are running high v NB with your quad @ 400fsb my E6400 rig will run 490fsb (thats as far as I have taken it) with lower NB voltage than you are using.

Thanks for your added insights... I'll add the stuff about the SB voltage causing shutdowns to my FAQ blog post. How long does the board take to shutdown with high SB voltage and what stress testing seems to replicate it reliably?

As for my NB voltage, I have to admit, I'm still not sure what to do with it. I've found I need it relatively high to run my memory at 1T but I haven't spent enough time trying to lower it and retest to know where the limit is.

[virtualrain]

that makes me have buyers regret for getting the 680i from dfi, why would you be so cold hearted to post that in here

Sorry! It shouldn't give you buyers remorse... you have a good board!

Keep in mind that we first saw the DFI 680i board back in Dec! Who knows when this board will actually be available. Hopefully sometime before the next generation!

[C-N]

Thanks for your added insights... I'll add the stuff about the SB voltage causing shutdowns to my FAQ blog post. How long does the board take to shutdown with high SB voltage and what stress testing seems to replicate it reliably?

It took between 20 & 45minutes & Orthos blend on priority 8+

As for my NB voltage, I have to admit, I'm still not sure what to do with it. I've found I need it relatively high to run my memory at 1T but I haven't spent enough time trying to lower it and retest to know where the limit is.

It may be the 4GB or even the quad that is demanding the NB voltage but considering the extra load your NB has compared to mine your voltage isn't crazy its just surprising its that high at 400fsb which would suggest to me you couldn't get much more out of your RIG even if the CPU could do it.

I have just had a little play & my board will boot 1970fsb on 1.51v NB although you do need 1.6 + to run 3dmark & keep the bandwidth up in Everest. All other voltages are at minimum.

I haven't found a massive drop in bandwidth from 1t to 2t why don't you drop to 2t & see how low you can drop the NB voltage just out of curiosity.

Craig.

[virtualrain]

It took between 20 & 45minutes & Orthos blend on priority 8+



It may be the 4GB or even the quad that is demanding the NB voltage but considering the extra load your NB has compared to mine your voltage isn't crazy its just surprising its that high at 400fsb which would suggest to me you couldn't get much more out of your RIG even if the CPU could do it.

I have just had a little play & my board will boot 1970fsb on 1.51v NB although you do need 1.6 + to run 3dmark & keep the bandwidth up in Everest. All other voltages are at minimum.

I haven't found a massive drop in bandwidth from 1t to 2t why don't you drop to 2t & see how low you can drop the NB voltage just out of curiosity.

Craig.

I am definitely FSB limited on my rig... as you say it could be the 4G or the quad. I can boot at 430x9 but not 440x8 so 430 is the most I can get out of it even though the memory can run 1T to 900MHz and the quad is clearly stable at 3.6GHz. You would think 450x8 would be feasible but it's not. It's very stable at 400x9 and I'm happy with that.

It's just a matter of tweaking voltages to run it as cool as possible...

Since I'm going to be cooling the NB with water when I go 24/7 with it, I'm not too concerned about the voltage, but when I get around to it, I'll play with it somre more. I know it can go much lower on 2T with relaxed timings.

I'll definitely report anything new that I find.

[Apollonas]

Sorry! It shouldn't give you buyers remorse... you have a good board!

Keep in mind that we first saw the DFI 680i board back in Dec! Who knows when this board will actually be available. Hopefully sometime before the next generation!

Yes, but we have to keep also in mind that this is a ddr2 board. If DFI wants to stay in the game, they should go out really fast with this board just to be in time with ddr3 also.

[Borgod]

I lapped my QX6700 today. Temps have come down a few degrees which is great, but temps between all 4 cores is very even now under load.

I also rotated my NB heatsink 180 so the pipes are over the NB chip, and put an 80mm fan blowing up aswell. It's really made no difference. On load the NB temp is still just over 80degrees

However I put a 70mm fan over the PWM heatsink so that would of made a good different Im sure as that was too hot to touch before.

[C-N]

I also rotated my NB heatsink 180 so the pipes are over the NB chip, and put an 80mm fan blowing up aswell. It's really made no difference. On load the NB temp is still just over 80degrees

I would have said even taking into consideration you have a quad & your in sunny Perth your NB temp is too high. I can only just hit mid 50's with loads of NB voltage & tight ddr900 1t under full load most of the time my rig runs 45-50oC.

silly question but did you remove the extra little foam triangle from the NB before you put the heatsink on & what thermal compound are you using ?

Craig.

[AlphaHeX]

C-N should that triangle be removed before attaching heatsink?