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

[Spazilton]

I have run into a new issue however.


It seems CPU throttling is somehow tied to VRM temps. Even with the CPU under 65c in core temp, the CPU starts to throttle its self. This is on water.

Once I place active cooling on the VRM it stops, take it away and it starts right back up.

Still trying to figure out what would be an effective VRM cooling solution over stock.

[Jobeo]

Still trying to figure out what would be an effective VRM cooling solution over stock.

lol that makes two of us... let me know if ya come up with anything

i don't understand your drawing... where do you have fans and where are they blowing ??

[vincey2kr1]

I have run into a new issue however.


It seems CPU throttling is somehow tied to VRM temps. Even with the CPU under 65c in core temp, the CPU starts to throttle its self. This is on water.

Once I place active cooling on the VRM it stops, take it away and it starts right back up.

Still trying to figure out what would be an effective VRM cooling solution over stock.

Put me in the throttling category, having the same problem. I'm running on air though and have a cooling fan over the VRM. If you look at page 88 of your manual you'll see the options in that screenshot for thermal management but seems like with this bios we dont have that option now?

[virtualrain]

Put me in the throttling category, having the same problem. I'm running on air though and have a cooling fan over the VRM. If you look at page 88 of your manual you'll see the options in that screenshot for thermal management but seems like with this bios we dont have that option now?

What are the symptoms of throttling? How can you tell it's throttling?

[vincey2kr1]

My 3d mark 06 scored do not change much when overclocked (like 500marks) when it should be alot more. Also using Sisoft Sandra CPU benchmarking utility and the scores are changing based on its performance you can tell were its throttling down to. in my situation its throttling from 3.5ghz down around 2.8-3ghz compared to the other cpus

[OldGuy]

Ya'll got EIST or C1E enabled? Not familiar with this bios but make sure you disable all of the cpu features that will throttle down when idle...just throwin ya some food for thought in case you haven't disabled these lil boogers...

[virtualrain]

Ya'll got EIST or C1E enabled? Not familiar with this bios but make sure you disable all of the cpu features that will throttle down when idle...just throwin ya some food for thought in case you haven't disabled these lil boogers...

Also, if you are running Vista, you also need to set your power profile to High Performance to avoid speed-stepping.

[Spazilton]

Ya'll got EIST or C1E enabled? Not familiar with this bios but make sure you disable all of the cpu features that will throttle down when idle...just throwin ya some food for thought in case you haven't disabled these lil boogers...

EIST isn't an option in the BIOS. DFI board always seems to be running at full multi. C1E is disabled.

[Spazilton]

What are the symptoms of throttling? How can you tell it's throttling?

I can tell its throttling multiple ways. The temps will drop 5-7c really quick than go back up in constant cycles.

CPU-Z shows the reduced multi kicking in. Than going back. RM Clock monitoring shows both the clock bumping down to a 6x multi and the VID getting bumped down. Its typical TM2 throttling.

With Coretemp 94 temps never going over 65c ~61-63 on on all 4 cores (@ 1.4 Bios, Strong drive, around 1.32 DMM measured actual load), if I remove my fan from the VRM heatsink for more than 10 seconds it will go into a constant throttle cycle (6x multi for a second 9x for a second etc..)

Even putting the fan back, it will constantly throttle under full load.

Very odd problem, I'm assuming that the PWM temps are being read somewhere and are tied into processor throttling.

I can stop it from happening with RM clock by disabling TM1 and 2 throttling.

My testing is on XP. Btw.

[Jobeo]

CPU-Z shows the reduced multi kicking in. Than going back.

wow ... hrmm... thats not good.

wonder if we can get an official answer from DFI on this ... problem is it doesn't seem like this board has caught on very much. sorta what VR was talking about the other day... without the consumer base backing it i'm afraid they'll just move onto other things.

i guess i'll have to look into cooling this pwm better... so far i've had no luck. =/ not good

[vincey2kr1]

wow ... hrmm... thats not good.

wonder if we can get an official answer from DFI on this ... problem is it doesn't seem like this board has caught on very much. sorta what VR was talking about the other day... without the consumer base backing it i'm afraid they'll just move onto other things.

i guess i'll have to look into cooling this pwm better... so far i've had no luck. =/ not good

Have you tried taking off the pwm and getting rid of the Thermal Clay(yes clay lol), its a tedious job prepare to use a toothpic and qtips though. I added some AS5 and I believe it helped, at least i can touch the heatsink for longer then a 2seconds now

[Raja@ASUS]

I think ceramique is the best to use in this situation,it's non capacitive, make sure your board is not bowed to the extent that some of the devices are not making contact with the sink, or over compensate with some ceramique to ensure contact with the heatsink in all situations.. Be midful of the smt capacitors that also sit under the recessed area of the heatsink, they should not make any form of contact at all..

The circuit handles dual cores great, quads need real sturdy cooling of the pwm area to hold them steady at any meaningful overclock..

regards
Raja

[Roger_D25]

Have you tried taking off the pwm and getting rid of the Thermal Clay(yes clay lol), its a tedious job prepare to use a toothpic and qtips though. I added some AS5 and I believe it helped, at least i can touch the heatsink for longer then a 2seconds now

I could be wrong here but if your heatsink is now cooler to the touch (since you can now touch it for longer than 2 seconds) than the AS5 is most likely not conducting heat as well as the factory installed stuff (if the heatsink was hotter before than it would be safe to assume it was getting more heat transfered to it before)? I went through this same thing on my eVGA after replacing the same clay stuff with AS5. It ended up being that my factory installed heatsink just wasn't making contact with all heat producing chips after the reinstall? Anyway just something to think about...

[virtualrain]

I can tell its throttling multiple ways. The temps will drop 5-7c really quick than go back up in constant cycles.

CPU-Z shows the reduced multi kicking in. Than going back. RM Clock monitoring shows both the clock bumping down to a 6x multi and the VID getting bumped down. Its typical TM2 throttling.

With Coretemp 94 temps never going over 65c ~61-63 on on all 4 cores (@ 1.4 Bios, Strong drive, around 1.32 DMM measured actual load), if I remove my fan from the VRM heatsink for more than 10 seconds it will go into a constant throttle cycle (6x multi for a second 9x for a second etc..)

Even putting the fan back, it will constantly throttle under full load.

Very odd problem, I'm assuming that the PWM temps are being read somewhere and are tied into processor throttling.

I can stop it from happening with RM clock by disabling TM1 and 2 throttling.

My testing is on XP. Btw.

Something doesn't seem right here... This is the first this has been mentioned (you and vincey) and lots of people have been overclocking. If this was a design/feature, I'm sure many more people would have been running into this by now. I've personally never witnessed this issue with my board and I was running some high clocks and Vcore before I decided to put a fan on the PWM area. My initial throttling was Vista's power management at work. If you're certain you've disabled all software speed-step type of stuff, you should consider logging this over on the DFI Club Forums to ensure this is expected behavior.