Discussion in 'Videocards - NVIDIA GeForce' started by dk_lightning, Mar 25, 2021.
That was some time ago. Gigabyte makes most overbuild cards right now especially his master.
If textures in games are getting corrupted due to damaged VRAM, wouldn't it also mean the mining hash rate would drop dramatically?
Try lowering vram speed an see if the problem persists.
I've just lowered my own vram OC, gotta really look after these puppies now.
Not that i mine but what difference would it make anyways.
*Saying that if i did mine i'd remove all OCs, probably underclock tbh.
Try undervolting as it lower's the temperature a lot to just check if you could fix the issue by doing so.
OP paid good money for card, gotta try for rma.
A few things to try;
1. Remove all GPU overclocking software
2. Uninstall display driver and clean your system using DDU
3. Try a different display driver
4. Verify game files in case of corruption (although unlikely if occurring in mutliple games)
5. Clean install Windows on a spare HDD/SSD and check if it still occurs
6. RMA (or enjoy your new paperweight )
you are channelling your anger at a completely wrong person.grow up.
if he had to resort to paying twice the msrp as the only option for a gaming new card it's understandable he wants to make some money back.
I know we heard ya, We all heard ya.
Folks some posts have been reported in this thread, there is some merit, the OP chose to engage and rightly so; it is his card he can do what ever he wants with it.
If you feel someone is being an ass then show them as such, do it with facts and respectfully, if the person continues being an ass things will be dealt with. And try to avoid dogpiling it does no one any good.
GDDR6X in 3080 and 3090 Hits 110C While Mining Ethereum
"When mining on both the RTX 3080 and RTX 3090, we found that the GDDR6X modules would peak at a much higher 110C, and the GPU would downclock itself severely to compensate for the ridiculously high VRAM temperature. This occurred on multiple different boards, from various vendors. And that's before applying any overclocking settings, which some miners like to do in order to chase every last bit of hashing performance."
The Memory Junction Temperature on the board itself gets insanely hot
"The Hot Spot" on the card.
Wow, after all this is over it will be funny trying to find one that wasnt burning 24/7 and on its way out.
Found this article pretty interesting:-
Nvidia GeForce RTX 2080 Ti “Mining” GPU benchmarked in several AAA PC games, shows reduced performance
The Intel core i9 10900K 3.7GHz CPU was used for the testing. If we take the average score, the used mining 2080 Ti card was about 10% slower than the brand new 2080 Ti.
One exception was Forza Horizon 4 game, in which the mining card was actually 20% slower than the new card. So this does show that any GPU used for mining can deteriorate gaming performance, which seems obvious given how mining works. Though, there is still no hard and fast rule to this.
The main culprit for the loss in gaming performance can be attributed to the GPU clock speed and temperature of the Mining card. There was actually a 100MHz drop in boost frequency for the mining card, due to the fact that the mining card was almost 14-16C hotter than the brand new RTX 2080 Ti, which hindered the overall performance."
Something to think about for people looking at the second hand market in the future
HWiNFO added new GPU temp monitoring recently:-
I used to think that my 1080 was sitting at a nice chilled 34C at idle
But it's not , the Hot Spot Temp is actually nearly 50C at idle and my 1080 Vram isn't anywhere near as fast as the new cards.
Hahaha This thread made my day
Thank you karma
That temp increase is just because of dried thermal paste and dust build-up. Easy to fix.
VRAM temperatures in mining will go through the roof on RTX 3090 or 3080 or 3070 on air cooled GPUs...
Better thermal pads can help with lower VRAM temperatures and pointing there 90mm or 120mm fans can help too, on my RTX 3090 I'm using heatsink with 90mm and 120mm which are pointing on backplate and my VRAM temperatures won't break 60°C(second one is in 72°C but that's I expected because of there no fan which is pointing )
I have done few tests with air cooled Palit RTX 3090 GamingPro and VRAM temperatures would reach 88-92°C in quick 5 minute render, adding extra OC like 700MHz on VRAM would break 100°C
Hope this helps
Interesting thought you point out. I did not think about that, but indeed these second-hand cards can become a serious risk, the seller can used it for mining for three years and than sell it in perfect condition. IMO electricity companies should put a cap on the maximum use of electricity, especially in mind of global environment issues we face.
I don't have GDDR6X on my 3070 it's just GDDR6. It has cooling as it's direct contacting the same copper cooling plate the GPU and VRM'S are.
Thanks to all those who are actually trying to help me out here. MVP to you.
After some testing and messing about I actually found the issue. While Nice Hash's quick miner does set its own gpu/mem clocks on start and then setting them back to stock on exit, I found that Gigabytes own Aorus Engine app was OC'ing my vram just a little to much all on its own. I only have it installed as part of the rgb control. It was OC'ing my ram to around 2025Mhz (16200Mhz effective) and my vram does not like anything much past 2000Mhz.
This was causing the glitches I saw. I do not know what caused this app to suddenly OC my ram like that but meh. Glad I found it and the glitches have stopped.
My vram does not have temp monitoring as far as I can tell. Had a good look through HWmonitor and some other progs but nothing is listed for vram or even the vrm temps. But I do have a thermal probe with my multi meter and did some checking on the vram temps and under gaming load it was around 70c or so, with mining it went up a little to around 82c so a little warmer. I might increase the fan speed for mining if I continue.
Now, If only we can all get DLSS in Warzone so I can play at more that 100FPS, that be great.