Originally Posted by cstring
tried the aopen bios too but the memory still is too hot.
I had bought some bfg heatsink kits and used the little ones.
I cut them to fit each single memory.
As bfg didnt put the thermal monitor, and I have no temp monitor
I just physically,carefully, touch the mem sinks and thier hot enough
to burn fingers if constant contact is made, way too hot.
I didnt change gpu core , is at 390/430 memory,even at bfg stock speed
736 will roast my flanges (fingers)
I know I'm running the ragged edge but obtain no problems, weather it be burning a full dvd , running a game , or just surfing the web.
As a test to see how high I could go, I ran at 390/430/930 way o/c'ing
(actually hopeing to burn it up) I already ordered a leadtech, lol
It ran w/o lockups or artifacts!!! but gpu core reset to 399.
Checking temps I first just tried to touch back of the pcb of video card
and it was extremely roasting, needless to say, I didnt get near the memory sinks but gpu core was cool !!!
I've reflashed back to stock, these are well built for taking extreme heat, I'll give bfg/chaintech that credit, although I wouldnt run like that too long,
I have this power supply
Anyway I'm looking at this cooling solution
but am not sure if it'll work on this card.
Your PSU is probably ok for now if it is really operating at these levels, but as I said theres not a lot of safety margin available for extra loading ?
to give an example, for a very long time I used a Qtech 550W, which although not considered amongst the good PSU's, I thought would be good enough for my use ? Well, unfortunately, my recent activities with overclocking my mobo/CPU/dimm and my 5900nu, seamed out of the blue to be causing random lockups/reboots/artifacting ? I finally got it down to be problems with my PSU 12v rail's continuous capacity ?
The Qtech 550 12v rail is rated at 14 amps continuous and 20 amps peak. Unfortunately for me it was the continuous rating that mattered ? My mobo/CPU o/c activities coupled with high overclocking of my 5900nu flashed as a 5950u was occasionally too much by overstressing my PSU's 12v rail just enough to cause me problems !! Resolved all problems by replacing my Qtech with a Vantec 520W stealth which has a 28 amp continuous rating on 12v rail ! So if ever getting random issues similar to my experiences, don't automatically ignore your PSU just because the manufacturers spec looks good on paper ? Any performance enhancing overstressing and/or overclocking we do increases PSU loading to unknown and sometimes unforgiving levels !
back to your temp problems; without proper memory heatsing/cooling you can't run at memory speeds of 930 ! Keep to no more than 850 ! Also, any kind of ramsink will give huge benefits over no heatsinking ? If you place a fan to blow over even small ramsinks, you will be amazed just how much heat can be removed ?
your GPU resetting to 399 still confuses me ? unless as I said, it is just your driver losing control of the overclocked 430 level ? and if so can be fixed by modifying your bios to 430 and it will stay fixed !
High overclocking with good temperature control with a XT card is always going to be a major problem ! the card has no accessable temp monitors and just about every XT type card has a diff kind of cooling/heatsinking and too many have inadequate or no direct cooling of memory chips ??? Because the memory stress/heat problem will be greater in proportion to the chip NS rating, it is better if considering buying one that you choose a card where the poorly cooled chips have a 2.4 or better NS rating ? preferably 2.2 if you can get it ?