DFI 790FXB-M2RSH Rocks

[nuts24]

Hit over 3.3 GHz on my 9850 BE CPU with AMD overdrive, Did hit 3.4 but got BSOD after 10 mins so I think I need a bit more voltage. Will try more tests when I get time.

I have had 3.3 GHz with the settings I have posted and also got 3.3 GHz whith just CPU multiplier and bus set to 200 Hz.

On my old motherboard I got between 2.9 to 3.0.

http://valid.x86-secret.com/show_oc.php?id=436303

[Oese]

nice

you seemingly do not have temperature problems with the nb or pwm heatpipe...

[EXxx]

54 degrees in the chipset :O, under stress and load, mine didn`t go over 42 ( i have a asus m3a79-T)

[nuts24]

yes I know the board temps are a little high, but this board works, unlike the last Asus I got that just gave me BSOD all the time. BSOD I can not fix, but board temps I can.

[Mathos]

thats pretty impressive speed considering the voltage you're at. Whats your NB speed running?

[unkn0wnhacka]

I have the same board on my new setup i just built.. i played far cry 2 alot last night but afterwards i was like "hey i forgot to overclock this" because i just built it yesterday. So i got into AOD and started tweaking.. i got to 3.4 without raising any voltages. Just after that i ran a stability test and failed it i quickly added a few volts and w00t!! its running perfect at 3.4 without any ram tweaking and just a basic overclock.

[Brother Esau]

Can anyone Confirm that these Temps are Accurate on these New D.F.I 750 Boards please?

[nuts24]

Uped the voltages a bit and hit 3.4GHz, then got BSOD with 3.5GHz, but still have lots of room with my voltages and not hit 60°C yet on air so still have room to play.

http://valid.x86-secret.com/show_oc.php?id=436889

Also downed clocked my ram to 3:5 set the CPU bus to 240 Hz and got to 3.360 Ghz, got BSOD at 3.480 Ghz after 10 mins.



This has all been done running on Vista 32bit.

[Particle]

Nice results, man. I'm glad you're seeing such a nice jump.

[xPliziT]

Hmm maybe its just me but i havnt found any "AutoXpress" feature in the BIOS.

Its lame always have to start AOD to enable it.
Anyone knows how to enable AutoXpress from BIOS?

[Sylv]

Hi all, recently bought this board and so far it is doing very well. However I am having trouble OCing my 9950BE.

Here is a Screenshot from AOD http://img120.imageshack.us/my.php?i...dscreenxz4.jpg

As you can see I am running the NB at 2.4ghz thats no prob (have not changed any NB volts over stock). However whenever I set the Mutli to 15 for 3ghz it BSOD's. I have ACC on an set to 4% all cores and I am on the 09/19 Bios. I am on water (Dtek Fuzion) so temps are not a problem. Setting CPU Vid Special Add in the Bios to +150mv for high 1.45v has not seen any change in the BSODs. I feel like im missing one setting that's holding me back but cannot think what, any help appreciated. Thanks!

[Chris_redfield]

You might have to play around with the ACC a lot more.

I suggest switching it from auto then turning all cores up to 4 or 6 and then run prime95 and see which cores drop (or if the system hangs completely,) and then make adjustments from there.

I wouldn't recommend turning all cores as high as 8 though, I did that and it caused data corruption.

[kelevra]

mine barely breaks 43C after an hour of Warhead. i should fire up something better to stress the system though.

Can anyone Confirm that these Temps are Accurate on these New D.F.I 750 Boards please?

[felix_w]

Uped the voltages a bit and hit 3.4GHz, then got BSOD with 3.5GHz, but still have lots of room with my voltages and not hit 60°C yet on air so still have room to play.

http://valid.x86-secret.com/show_oc.php?id=436889

Also downed clocked my ram to 3:5 set the CPU bus to 240 Hz and got to 3.360 Ghz, got BSOD at 3.480 Ghz after 10 mins.



This has all been done running on Vista 32bit.

Very nice results...Have you tried to run stable with tighter ram timings?

[nuts24]

Yes my ram is running at 5-5-5-15 now and still no problems between 3.0GHz and 3.4GHz.

I don't think I can push the ram any more as its cheap & I can not find a spec sheet any for my sticks. they have Emartbuy on them.

[cbjaust]

I got one now and I've been testing memory with memtest and I found something way odd with the bandwidth numbers

Memtest86+ v2.01:

1. 13.0 x 254 = 3302, L1:27067, L2:6725, MEM:4299 (MB/s)
2. 13.5 x 246 = 3321, L1:27223, L2:6764, MEM:2821
3. 13.0 x 256 = 3328, L1:27280, L2:6778, MEM:4333

I ran sisoft mem bandwidth test on the number 2. setting and got a bunch of varying numbers but they seem to settle around about roughly 4600MB/s for float and integer with the buffering setting disabled - I think that it is low.

Gonna have to check the 3. setting.

Is Everest consistent in the results it gives?

for the the integer results from sandra mem bandwidth test in eight I got:

3382 MB/s
4230 MB/s
4581 MB/s
4528 MB/s
4486 MB/s
4635 MB/s
4710 MB/s
4556 MB/s

all over the place

Cheers

[Mechromancer]

I got one now and I've been testing memory with memtest and I found something way odd with the bandwidth numbers

Memtest86+ v2.01:

1. 13.0 x 254 = 3302, L1:27067, L2:6725, MEM:4299 (MB/s)
2. 13.5 x 246 = 3321, L1:27223, L2:6764, MEM:2821
3. 13.0 x 256 = 3328, L1:27280, L2:6778, MEM:4333

I ran sisoft mem bandwidth test on the number 2. setting and got a bunch of varying numbers but they seem to settle around about roughly 4600MB/s for float and integer with the buffering setting disabled - I think that it is low.

Gonna have to check the 3. setting.

Is Everest consistent in the results it gives?

for the the integer results from sandra mem bandwidth test in eight I got:

3382 MB/s
4230 MB/s
4581 MB/s
4528 MB/s
4486 MB/s
4635 MB/s
4710 MB/s
4556 MB/s

all over the place

Cheers

Sounds like an "odd deviser" thing going on with the memory controller with half-step multipliers (the 13.5 rather than 13.0). Didn't Athlons have an issue with that? Also it really looks like there could be a few "sweet spots" that allow the asynchronous clocks to work in better harmony around the board. Intel deliberately payed attention to this with their new Core i7 architecture. There should be a few combinations of clock speeds that produce higher bandwidth and/or lower latencies.

[gOtVoltage]

Yes it seems,, IMC/NB contoller likes certain dividers...Viewing CPU-Z i bet that @ 13x vs 13,5x multi sets different ram Divisor...Some dividers can ,,will downclock the ram or NB . Thus affect the efficency of CPU,, even though it shows that the MHZ is the same or simalar as another divisor..

Comes down to the way the BIOS is coded and or the IMC just likes different divisor thats all..

Ive seen this while running Higher Divisor vs Lower Divisor and high HTT... go with the higher performance divisor or Try using non Half multis

Its definatly a AMD balancing act. Just takes time,, find the Sweetspot and Walla performance time .Nothin new ,,Girls kinda act the same sometimes

[Glow9]

Good work man, DFI knows AMD thats for sure

[cbjaust]

Yes it seems,, IMC/NB contoller likes certain dividers...Viewing CPU-Z i bet that @ 13x vs 13,5x multi sets different ram Divisor...Some dividers can ,,will downclock the ram or NB . Thus affect the efficency of CPU,, even though it shows that the MHZ is the same or simalar as another divisor..

Comes down to the way the BIOS is coded and or the IMC just likes different divisor thats all..

Ive seen this while running Higher Divisor vs Lower Divisor and high HTT... go with the higher performance divisor or Try using non Half multis

Its definatly a AMD balancing act. Just takes time,, find the Sweetspot and Walla performance time .Nothin new ,,Girls kinda act the same sometimes

nah, it's definately CPU/7 in CPUz for both cases. I'll have a closer look at it tonight and post some pics and some mem benches from Everest which I just installed and I'll piss that other piece of junk from sisoft right off.

Cheers

Here are some pics I made earlier:

13.0 x 256:



13.5 x 246:



The only difference, apart form the cpu multi and HT Frequency is that Max Async Latency (set to Auto in BIOS) is 8ns at 13.5x vs 7ns at 13.0x

I also ran a quick Memtest at 14 x 236 and got CPU: 3304MHz, L1:27084MB/s, L2:6724MB/s, MEM:4302MB/s so those numbers are back on track. It makes me wonder if I am missing a bit of performance on a 4800+ I have here usings its default multiplier of 12.5x

Cheers

[Aussie FX]

Those 6000's always were a problem with ram.

Try sticking with even multis.

Regarding the motherboard, if anyone has an old Crosshair lying around, the PWM fan that came with it, that was supposed to be for watercooled setups fits pefectly over the pwm cooler on this board.
My pwm temps are now ~40°c in 30° ambients.

[cbjaust]

Those 6000's always were a problem with ram.

Try sticking with even multis.

Regarding the motherboard, if anyone has an old Crosshair lying around, the PWM fan that came with it, that was supposed to be for watercooled setups fits pefectly over the pwm cooler on this board.
My pwm temps are now ~40°c in 30° ambients.

more information please re: Windsor and RAM

Cheers

[ah_khoo]

anyone here change the stock cooling on NB & mosfets? i found the temp on NB was rather high. apart from this (NB temp issue) & acc problem (i can't boot whenever i turn it on), this board is superb... rock solid + lotsa tweaking options...

this is what i can manage so far due to temp issue...

[caveman_2]

I placed (balanced atm tbh) a 60mm fan directly in front of the nb heatsink. my normal nb temp was 57.C, on low fan setting speed it is running at 30.C and on high it has dropped as low as 26.C

[ah_khoo]

I placed (balanced atm tbh) a 60mm fan directly in front of the nb heatsink. my normal nb temp was 57.C, on low fan setting speed it is running at 30.C and on high it has dropped as low as 26.C

thanks mate. can i assume that the fan is blowing upwards if you're using a tower case?

from 57'c to 30'c, that's a drastic change...