veganfanaticLooking at Krams screenshots v mine, i use a 4K panel which is wide than is this forum template
MasterMinerGPU core temp != GDDR temps ty is correct. The gpu core can run at a different cycle rate than memory and hashing is brutal on ram. take those software scans of temps with a grain of salt. modern GDDR memory has advanced ceramic coatings that dissipate heat well at its surface where the sensors are located. But underneath that coating is chips getting stir-fried. much like the Intel core processor temp confusion a few years back when they used paste instead of solder for their lid-caps. sensors completely misread temps because they weren’t capturing the real picture.
dream3MasterMinerGPU core temp != GDDR temps ty is correct. The gpu core can run at a different cycle rate than memory and hashing is brutal on ram. take those software scans of temps with a grain of salt. modern GDDR memory has advanced ceramic coatings that dissipate heat well at its surface where the sensors are located. But underneath that coating is chips getting stir-fried. much like the Intel core processor temp confusion a few years back when they used paste instead of solder for their lid-caps. sensors completely misread temps because they weren’t capturing the real picture.What is the conclusion then? Is hwinfo right about temps when mining using stock settings? Or do I have to make sure HWI reads 95c or less so I dont blow my card up while mining?
Dabadger84This is going to be the same deal as when a previous mining bubble burst & the market got flooded with R9 290/290Xs etc a few years ago - so many of those cards are going to have VRAM damage that makes them basically useless from being mined on continuously while the memory chips were borderline overheating all the time.Used market is going to be a mine field once this bubble bursts.
dcellOkay, just a thought here, but if that VRAM was really hitting 110 C, and those miners are running their systems 24x7, wouldn't we have seen some fires at this point? That's a crazy high temperature and things might start to become toxic or melt. I'm not an overclocker, or a miner, but I do have CS. (common sense).
dream3My card doesnt throttle when HWI reports 110C memory temp, but that doesnt mean this isnt bad for the card long term. That's the whole point i see people around making. Ill keep it on the safe side and cap my temps at 95c until more info comes out of this. Not worth losing my single 1.5k card for mining.
dream3nicehash and a 3090 tuf
kram36dream3nicehash and a 3090 tufWhat are the clock and voltage settings you are running on the card?
dream3kram36dream3nicehash and a 3090 tufWhat are the clock and voltage settings you are running on the card?-500 core and clock
kram36 Not exactly what I was looking for. What do you have the power limit set to and what is your memory speed set to?
davecj82Kram36, I know this is a way too late reply, but I just happened upon this forum today. You may have already discovered this a long time ago, but in HWinfo, you're looking in the wrong section. Scroll up to the GPU0 section, right above the section you posted, and look for GPU Memory Junction Temperature, next entry right under GPU temp. That's the reading you're looking for, not the ICX readings. Unfortunately, EVGA doesn't have that particular temp in its ICX section, though I wish it did, that's something I'd love to monitor while in-game.
Pro$JC427Removed
TwoEvilOne74.0c 3090 Kingpin Mem 1600MHz T-rex mining Ethereum 127.01MHs @ 293W. Keep it cool and it will perform.
rhettmarceloTwoEvilOne74.0c 3090 Kingpin Mem 1600MHz T-rex mining Ethereum 127.01MHs @ 293W. Keep it cool and it will perform.How did you cool this down to high 70's? Is it only a fan and the backplate? What thermal pads are you using? Stock?
foxminoMost of these 3000 series cards have bad or cheap thermal pads especialyGigabyte i got these to replace the stock pads and they work great...110c down to 76chttps://www.amazon.com/gp..._s00?ie=UTF8&psc=1