anyone know which bits there have contact with the stock heatsink?Quote:
Originally Posted by largon
Printable View
anyone know which bits there have contact with the stock heatsink?Quote:
Originally Posted by largon
Cozwin, might be possible. Waiting on bios atm.
I can squeeze a bit more out of this card. One of the things I do is re-torque each screw making sure the core is as tight as possible without breaking it. Make sure you have the right screw driver to prevent damage or dont do it at all. I have not taken the heat sink off.
SLI on intel chipset? i want one :DQuote:
Originally Posted by FUGGER
skulltrail me thinks :D
7900GS vs 8800GT
http://img86.imageshack.us/img86/5483/dsc0073xs2.jpg
the GPU is sooo big compared to 7900GS, so i need a vmod anyone has done it?
metro.cl,
vGPU and vMEM mods should be quite similar to those on 8800GTS/GTX, macro shots of the areas highlighted in the jpg attached to this post would help to pinpoint the exact soldering locations. :)
The rectangular chip should be labeled:
Primarion
PX3540BDSG
DO NOT USE THIS IMAGE
will update proper one soon
Thanks, dinos22. :)
Need clarification on the existence of certain traces and few resistance measurements, see attachements.
First pic:
There should be traces between points 1, 2 and 3. It certainly does look like there is a trace between 2 and 3, but is there one between 1 and 2? = if there are traces between 1, 2 and 3 then resistance between each is ~4-6Ω.
Also,
Resistance between points 2 and 4? (might be <10Ω)
Second pic:
Resistance between points 5 and 6? (might be around 500Ω)
All those screws on the back need to come off for the heat sync removal? Anything else?
What cooler are you putting on it, I have the Fatal1ty sitting here for my card, should be here today and I am putting the cooler on it right away.
largon that image was from nvidia FTP
i just looked at my card and it looks a bit different so i took it DOWN for now
and will upload my macros
also can you measure resistance with card not turned on (pretty sure yes) and also how the hell do you measure it :D hahaha
with your ddm just measure
black on one side of resister and red to other side
that i knowQuote:
Originally Posted by cowie
which nob is for resistance though :ROTF: :D sorry noob with this kinda stuff
I posted a thread about G92GT vmods in Xtreme Graphx Vmods -section:
http://www.xtremesystems.org/forums/...d.php?t=163929
Let's continue there. :)
"the GPU is sooo big compared to 7900GS"Quote:
Originally Posted by metro.cl
the ihs u mean is bigger?the 7900gs is naked and has the bigger die.
"Thanks to a die shrink to 90nm, the G71 GPU on the 7900 GT was able to deliver the same basic technology and performance as the 7800 GTX while running cooler and costing less to produce."
http://64.233.169.104/search?q=cache...lnk&cd=4&gl=us
"Well, the 8800GT is 65nm instead of 90nm. Smaller die size might have dropped the cost quite a bit. And who knows what ATI has coming. ..."
post #3
http://www.halflife2.net/forums/show....php?p=2426376
so its a bigger ihs with a SMALLER die 90nm(g71-g80) shrunk down to 65nm(g92)..i honestly belive naked gpus run way cooler.wish my gtx was naked.
sorry fugger,was telling him die shrunk..
Both are naked.Quote:
Originally Posted by railmeat
might be possible crossfire is working on different mobos now inc nvidia chips so maybe, but im sure 2 of these and a 45 nm cpu and a nice sli mobo could probs take world record for 3dmark06Quote:
Originally Posted by metro.cl
I ran the Crysis bench, 1680x1050 custom settings, I will re-run with high and very high detail.
Crysis for me run good with the video card at stock.
GPU/CPU
All stock settings with E6850, at resolution 1024X768.
More than what a 8800GTS scores
You know you could always take off the heatsink and see what happens.Quote:
Originally Posted by railmeat
Honestly i don't see how a naked GPU would be cooler than one with a good heatsink and thermal compound.
Just like a naked CPU runs hotter than one with a heatsink and good thermal compound. Metal just conducts heat better than air thus you can move heat way from the source faster.
Now back to the topic, I can't wait to get my hands on one of these. I plan on seeing how far i can push it before it breaks. Can anyone tell me how hot does this card usually get at stock?
silly boyQuote:
Honestly i don't see how a naked GPU would be cooler than one with a good heatsink and thermal compound.
he is refering to a naked DIE like the 7800/7900/8800GT use vs. the 8800 GTS/GTX which uses an IHS to cover the die like your cpu does.
an IHS leaves too much room for heat gaps between the die and the IHS.
thus that is why you see people taking the IHS off their cpu's. in general your only going to get 5 deg better temps but there are extreme cases where i have seen twice that in a temp drop.
Oh, thank you for the explanation. That could indeed be better for tempsQuote:
Originally Posted by Lestat
Nicely done Charles :hammer:
Last night I was playing around trying to get some sort of result out of my cards after two previous nights of wasted benching time and motherboard horrors with beta BIOSes trying to get something stable with the Yorkfield + high FSB.
Everything came crashing apart night before last when the GA-X38T-DQ6 went out with a pop and it's soul puffed out into the surrounding atmosphere.
So last night, pissed off and just wanting results, I swapped in the GA-X38-DQ6 with the F5h BIOS loaded and cranked the multiplier just to get something happening with this setup. Here's what happened:
System:
Intel Core 2 Extreme QX9650 @ 4.66GHz cooled by Frozen SS phase cooler
GIGABYTE GA-X38-DQ6 with BIOS F5h
Corsair Dominator PC2-10000 @ 533MHz (DDR2-1066) 4-4-4-4
NVIDIA reference 8800GT @ 760/1701/1025 (2050 effective) with stock cooling and 169.01 drivers
Silverstone OP1000 PSU
Western Digital Raptor-X HDD
3DMark06 result: 16,357
http://img20.imageshack.us/img20/4465/cropped1xb2.jpg
HWBot compare link: http://www.hwbot.org/result.do?resultId=661269
HWBot profile: http://www.hwbot.org/user.do?userId=11510
Casualties:
1x GA-X38T-DQ6 replaced with GA-X38-DQ6
After getting angry with having no results the platform was moved to the GA-X38-DQ6. Further wanting to simply get results, I avoided the annoyance of FSB overclocking and instability and just cranked the multiplier for some quick results. The memory and general bandwidth of this sytem is quite low for this reason.
There is a lot of efficiency and general extra speed to be gained from this setup. Next step is to get some volt mods happening and get some dry ice on these cards - they're great fun :D
Due to currently being on deadline at work (AtomicMPC magazine) I hadn't a chance until just now to see these results and had thought I'd scored myself a WR but you certainly showed me up mate :p: :ROTF: :clap:
Hopefully I'll get another chance with the gear over the weekend and can get some decent FSB and memory frequencies happening.
BTW, what BIOS did you run on the P5E3 Deluxe? I got the 0508 BIOS earlier in the week but am yet to test it. If it's a winner I'll make a swap to that board for weekend benching with a set of Patriot PC3-15000 modules or the Corsair PC3-14400 that just arrived an hour ago :up:
SM3 seems very low. I'm getting 100point more even on stock ultra frequence with 1 year old GTX mod to ultra. It's strange 150mhz more on GPU and 200 more on Shaders with 600mhz more on CPU and 50% more cache to give you such a low SM3 result:rolleyes:Quote:
Originally Posted by moloko
http://img91.imageshack.us/img91/269...tocksm3du1.jpg
Hey fugger, how is the noise of the HSF?
I'm a little disappointed it ain't a dual slot HSF, but it probably doesn't need it because it runs so cool.
My main concern is that the little fan supposedly causes an annoying high pitch whine when it spins up - do you notice this?
I cant hear it while gaming. My Vapo is louder.
Not bad Josh, 0503 bios still.