MB OC Report :: EVGA nForce 790i SLI FTW Digital PWM

[Glow9]

Anyone have a 8 phase evga board to run benchmarks against this thing?

[3oh6]

sweet photos and info there mate

those photos must of taken ages in post-processing !

thanks George. i did have to adjust my work flow a pinch, but for the most part, it was the same script ran in PS. the biggest adjustment is the white balance which i struggled with so i am not sure how many more black background reviews i will be doing.

Anyone have a 8 phase evga board to run benchmarks against this thing?

if you mean the 790i Ultra, the two boards are going to be the same with identical performance give or take a bit of changes to bandwidth that wouldn't show up in anything but Everest and maybe SPi 32M. i have a bunch of results about the clocking differences that you can compare in my 790i Ultra OC Report...

http://www.xtremesystems.org/forums/...d.php?t=195257

[3oh6]

i have finally gotten serious with this board...enough said



plenty more to come tonight...GTX 260 SLI action at the same clocks

[dinos22]

faints

[Johnny Bravo]

Looking sharp Jody

I have to say you really have got the forum review look down to a fine art - I've seem website reviews looking less professional; great stuff

[3oh6]

faints

yup, another 15K 790i board. both Ballistix and G.Skill have no problem on this board as well. i don't quite have the FSB as that i do with the 790i Ultra but am making up for it with the extra CPU speed of the 10X multi and LN2

Looking sharp Jody

I have to say you really have got the forum review look down to a fine art - I've seem website reviews looking less professional; great stuff

thanks buddy. i am getting closer to where i want things to be and have started finding some acceptable workarounds to HTML being disabled. i appreciate the comments, much appreciated guys.

...and a bit of the encore from the session last night. it ran short because i bugged the CPU and the pot just wasn't the same after. so no 06/vantage but lord knows i spend more time on 01/AM3 anyway

i will try to get down to temps again this weekend and do some 32M action as well then get the cards prepped for LN2.

I finally got a hold of a pair of GTX 260s, EVGA Superclocked + GTX 260s to boot. These two are from a retail store and appear to clock about as good as others with stock cooling in normal ambients. I only have them for a short period of time but will do my best to make use of the them as best I can. First up will be some clocking with the cards on air and the CPU under LN2 to get a feel for what they can do on this board and how the CPU will clock FSB under cold on this guy. Here is a complete list of hardware used for these results and then a photo of the setup as always:

• CPU: Intel C2D E8600 (L814A526)
• CPU Cooling: Rev3 CU Mousepot w/LN2
• RAM: Crucial Ballistix 2x1GB PC3-16000 CL9 // G.Skill HZ 2x1GB PC3-12800 CL7
• NB Cooling: Swiftech MCW30 / MCP655 / MCRES-Micro / PA120.2
• SB Cooling: Thermalright HR-05-SLI
• PWM Cooling: Swiftech MC14 x 4
• GPU: 2 x EVGA GTX 260 Superclocked+ (177.92)
• PSU: Ultra XPro 750W
• HD: Seagate SATAII 80GB 8MB NCQ
• OS: Windows XP Pro SP3 (fully updated)

The setup as tested...
click for full size...


I missed any Vantage or 06 benchmarks because I bugged the CPU and then it was all downhill from there. I didn't want to stick to the same clocks for each benchmark as 05/06 are limited by FSB so I basically maxed out each benchmark. The goal was to test FSB clocks for each benchmark as this was the first time running LN2 on this board. Consider these the shakedown runs, with most tweaks applied...be sure to note the bandwidth in a few of them, 15K on yet another 790i board for this guy

click for full size...

Things went well and getting down to temperature went quite smooth. I was able to boot at the clocks you see above no problem and the board responded as it should throughout the session, until I bugged the CPU and warmed it up with a blow dryer. That was the end of that as the pot was just too unstable to run anymore. Temp swings were nightmarishly quick and all over the place. The pot needed to be warmed up to -80C before the bug went away, with the RE, this CPU would bug and 20 seconds later it was back in action without any outside heat applied which made benching a breeze. With this board, I will just have to make sure not to cold bug the bloody thing.

The other thing I want to mention is that my Ballistix will absolutely not boot at higher than 1050MHz. I thought it was a board issue and threw the G.Skills in to find they were booting just fine at 1096MHz so clearly not a board issue and related just to the Ballistix.

I might have one more session with the cards on air and CPU on LN2 before three potting it, but after getting these scores, there is no doubt I will be getting the GTX 260s cold at some point, even if I am only benching the CPU at 5850MHz. Might even bother with some 32M action at 5850MHz/1096MHz 8-7-6, should be some blinding fast times for the clocks.

[Leeghoofd]

Brrrrrrrrrrreathtaking...nice stuff, looking forward to some more results... what's your opinion when you use the stock mobo heatsink... 500-533 FSB doable daily ? Really wonder why Nvidia still doesn't have a decent IHS on that NB chip...

[dinos22]

jody
very nice tweaking indeed mate
you are on the wrong drivers though
you CH is terrible...you've lost 2000 points on it

interesting about your bugged CPU issues....my bench session consisted of me having to warm up my CPU to -76C from -120 every time i crashed or restarted my rig but it kept going until it started going a bit funny near the end which is when i pulled the plug

try insulating the whole area with plastic spray....should help a bit

[ElEctric_EyE]

3oh6, You say in your second post about this 790i board, that "Another major feature is one that we are all probably quite familiar with, the ability to run SLI. Until the X58 boards are available, the NVIDA chipsets like the 790i are the only ones you can run two or more SLI compatible video cards".
Do other EVGA boards have hardware/driver problems in SLI? I am considering the EVGA 780i FTW to use 2 spare 8800GT's (g92 watercooled) and my current DDR2 Corsair Dominators... TIA

[3oh6]

jody
very nice tweaking indeed mate
you are on the wrong drivers though
you CH is terrible...you've lost 2000 points on it

interesting about your bugged CPU issues....my bench session consisted of me having to warm up my CPU to -76C from -120 every time i crashed or restarted my rig but it kept going until it started going a bit funny near the end which is when i pulled the plug

try insulating the whole area with plastic spray....should help a bit

man, i knew i should have tested the cards more on air. i was literally testing the card clocks with the CPU under LN2 last night when chatting with Jake, i had about 7 minutes of run time on them before the LN2 started flowing. i'm assuming i can grab the drivers used in your recent GTX 280 exploits, just let me know if not. might fire the setup again tonight just because i shouldn't

as for the cold bug, not sure what it was but after the first bug the whole setup went to snot. temps were bouncing around all over and trying to get anything to run was just impossible. i'll figure it out but more importantly, stay away from bugging it better. it's not like i am really pushing the CPU to its limits at 5850MHz, it runs 6150MHz no problem on the RE. maaan, i want more FSB!!!

3oh6, You say in your second post about this 790i board, that "Another major feature is one that we are all probably quite familiar with, the ability to run SLI. Until the X58 boards are available, the NVIDA chipsets like the 790i are the only ones you can run two or more SLI compatible video cards".
Do other EVGA boards have hardware/driver problems in SLI? I am considering the EVGA 780i FTW to use 2 spare 8800GT's (g92 watercooled) and my current DDR2 Corsair Dominators... TIA

i think you misread the comment just a pinch, i meant that all NVIDIA chipset motherboards are all you can run SLI on, not specifically only the 790i.

the 780i should be fine for running cards in SLI, nothing i am aware of that has been reported as an issue.

[dinos22]

maaan, i want more FSB!!!

maaan every board i try i keep saying that lol

i got a new CPU but its a crappy stepping
actually got 2 to test

fingers crossed at least one of them will be ok

[3oh6]

haha, thanks Dino, cracked out a 123K first run...@590FSB (more vSPP). going to mess around some more with 01/am3 then hammer out a quick 06/vantage and see if i can run 32M at 590FSB.

[dinos22]

nice

with 6GHz you should definitely beat my 126K IF you can keep P1 even though you are on those cards
my run wasnt fully tweaked yet...i still have another 1.5K to get out of it
if i could get P1 working OMG......it would be scary fast

[3oh6]

thanks to Dino there was a small update with different drivers for 01 and while i was at it, i stabilized 590FSB...more vSPP

i also fired off a quick 06 and Vantage (sans PhysX)...

click for full size...

then i went bandwidth hunting and boy did i hit the jackpot. all i can say is that my EVGA 790i Ultra never did this

First up is a small collection of memory bandwidth results. I will keep the highest Everest Read bandwidth I manage with this board posted here. Think of this section as an almost "best of" type deal.

click for full size...

unfortunately i couldn't get a 32M run at 590FSB so i gave up and took a nap. 580FSB seemed to be running fine but i never let it finish as the time was just disappointing after running 6100MHz+ 32M on the RE

[Gautam]

The 01 especially is hardcore, nicely done!

And truly, a lot of reviewers out there could do well to learn a thing or two from you.

[dinos22]

so you boot at 1.75vSPP but what do you dial it down to then

it is really odd that you have to boot at such high SPP but that is something i should try

i found the culprit that killed my gem E8600
it was my 790i ULTRA
one of the mosfets is ed on it

i think its time i try FTW with older bioses


you took a hit on a few tests in '01 even thought you got another 50MHz on CPU :S.....probably drivers or just a bad run for '01 or unstable

[3oh6]

thanks Gautam, i am just happy to be north of 120K...now to make it 125K.

@Dino...i boot at 1.75vSPP then dial it down to ummm, 1.75v, the board won't last long running too much like that i don't imagine

it is definitely the drivers that gave the minor hits in DL, and N. i did a whole bunch of testing and they always came out lower. i have briefly tried 790i Ultra BIOSs but they resulted in weird voltages and FSB holes all over the place. the board didn't feel right so i am sticking with the FTWD BIOSs for now. next step is to get something colder on the NB but i just don't have any means to fashion a pot. i am thinking hooking up the loop to an old 360 rad i have laying around and dumping that in a bucket of ice, ghetto styles. ahhh, the memories of running Opteron 146s with the bucket of ice

[dinos22]

yeah fair enough

lol 1.75v is nothing

eeek

[3oh6]

lol 1.75v is nothing

that's what i was thinking, and after digging through a box of 'stuff' i pieced together a plan just crazy enough i thought it might work. not sure how this idea started but it did pan out in the end and provided some much needed information on this chipset...cold is good

forgoing the ghetto styles ice water chiller, i took ghetto to a whole new level, photos will tell the story...

click for full size crazyness...

the hardware is the base from an old Asetek GPU waterblock (same as their NB waterblock) i had laying around from one of their kits reviewed years ago. i saw it in the box and was thinking it would be better than the MCW30 for chilled water...then i really started thinking what else it might just happen to work as. so to the kitchen i went, found my funnels, and the rest is history

the CPU-Z validations are on the drive still and i will grab them later, but here is the money shot...600FSB P1/P2 enabled FTW


3D benching is definitely possible at these clocks and higher but that i will save for another day with a better chipset solution. chipset temps were being reported by Everest as low as -39C before things started getting wierd so perhaps cold bug territory around there. needless to say, plans are in motion for legitimate SPP sub-zero cooling but for 1 hour prep, a quick 1/2 hour test, and some crap laying around...i think the night turned out to be well worth it.

[dinos22]

dont get comfortable
i modded the SPP and chasing you down

just testing superpi at 575MHz atm 1.675vSPP @5463MHz on SS

watercooled NB still

[3oh6]

awesome, good to see yours is clocking pretty much the same as mine with volts. i'm here waiting

[dinos22]

580 seems to be far as it will go on single stage for CPU and water on SPP

i guess with some subzero on SPP and colder temps on CPU i'd get some more but still not good enough

i did some of those everest benchmarks....bloody hell man how easy it is to get 14.5K lolzors

[3oh6]

now are you testing with 3D or 32M? i find 3D to go much higher than 32M with this board. could be because of my memory at 1100MHz as i think the G.Skills are finally showing their days after been beaten for so long. i am still going to try chilled water on the NB because i think even taking temps down to ambient helped a little unless i can get a NB pot relatively quickly. running with just that water block base required a pour every 10 seconds or less. 3D was do-able but wouldn't be with the CPU under LN2 as well. temps were relatively easy to control though, might just make a put out of the base with some copper pipe

[JohnMike]

I have been playing around with this board during this weekend.

My setup:

CPU: Q9650
Ram: Gskill DDR3 PC12800 HZ 4GB (2*2) 7-7-7-18
GPU: Powercolor ATI HD 4870 1GB GDDR5 PCI-E

The system runs fine @1333 FSB of course, but from here to 1600 FSB can't get stable, having boot BSOD as Vista64 starts loading.

I was about to quick on this board after dozens of attemps to OC, until I tryed to push FSB above 1600 FSB.

Now i can push above 1600, 1700 and 1800 FSB, but from 1333 to 1600 FSB, no go!

Is this beahviour regular on 790i boards?!

[dinos22]

now are you testing with 3D or 32M? i find 3D to go much higher than 32M with this board. could be because of my memory at 1100MHz as i think the G.Skills are finally showing their days after been beaten for so long. i am still going to try chilled water on the NB because i think even taking temps down to ambient helped a little unless i can get a NB pot relatively quickly. running with just that water block base required a pour every 10 seconds or less. 3D was do-able but wouldn't be with the CPU under LN2 as well. temps were relatively easy to control though, might just make a put out of the base with some copper pipe

i am testing superPi 8M which is ok
32M is spotty
3D is fine as you said
same behaviour here

i always test with superpi first

i was running my ram around 1040 range