GPU from Olx - Best shipping service within India

Status
Not open for further replies.
with artifacts and crashes at there place there is a strange problem with this card the HDMI is showing waves on monitor and yes I tied replacing the vga and power cable checked for any interference also tried only plugging the monitor in a single board
But as soon as the dvi is connected the waves on the HDMI monitor reduces to very very low so not a big problem I can deal with that but on the dvi monitor it is showing waves with very high intensity more visible on white color and less noticeable on black color It is very hard to even look at the dvi monitor.


Big Update: I disassembled the card repasted the thermal paste/pads and also replaced the fan screwed it again tightly and guess what the card is working without any external support with the corrected bracket (not bent) so one problem is fixed but the card still seems to artifact and crash in the same manner.
I tried downclocking it, debug mode and even power limit to 63%.
I also tried with fan speeds at 90%. directed more fans to the card.
Air conditioned the room at 25 degrees Celsius.
rolled back to driver 512.95
still same.

Also games like tekken 7 and loz botw haven't crashed also take in account hose games have a frame rate cap tekken 60 and zelda 40.
rottr ,ac unity , re4 remake ,unreal engine crashes same error in event viewer
Already wasted a lotta time with the card not gonna try anything more soon .
requesting refund.

unsurprisingly the crashing problem is fixed when downclocked by -502 on both memory and clock , but I did not buy the card for that.
BullettuPaandi said:
There was a lot of confusion as well - you mentioned PC freezes; then it was some cuda error; then this
Click to expand...
blender doesn't crash on a gpu fail event it just simply shows there is something wrong with the graphic processor and stops the 3d viewport rendering until restart that is why.

UPDATE: downclocking by 502 on both core and memory seems to have fixed the artifacting and crashes completely.
but I am not using it like that
 
Last edited:
-180 on mem and -50 on core also seems to work fine for now.
Pritambb said:
-180 on mem and -50 on core also seems to work fine for now.
Click to expand...
No crashes/artifacts so far in re4 remake.
Gonna say something about that bro? @BullettuPaandi
BullettuPaandi said:
out first by yourself, before proceeding with asking help from others.
Click to expand...
Update : Card downclocks like crazy during occt test pushing the clock to 1265 and hence runs stable unlike games . With temperatures reaching nearly 90C. All settings for the card are default. Games doesn't rise the gpu temp as much and hence crahes after a while because of not throttling.
Power draw around 220 watts during the test in games it is around 175-180 at most or less as low as around 130 watts

crahed on -180 and -50
;-(
nothing crahes except gpu intensive applications cemu works fine doesn't use much gpu it is cpu bound.
 
Last edited:
1686290732176.png

is this normal? occt identifies the card wrong.

@BullettuPaandi I tried with different clock speeds, and it looks like the less the gpu clocks the less are the crashes and artifacts.
-190 mem and -55 core is working very well for now. When I try to use it on default settings the card always crashes.
Somehow occt does not want to run on higher clocks it just downclocks to 1265mhz in test.
 
Last edited:
Pritambb said:
is this normal? occt identifies the card wrong.
Click to expand...
Does it? Giving you the benefit of doubt, you uploaded the wrong picture perhaps? Because this one is clearly from the hardware's info tab. It wouldn't show defects here even if there is any.

Pritambb said:
it just downclocks to 1265mhz in test.
Click to expand...
1266MHz is the boost clock. You can look that up.
 
This video is worth checking out
It relates to the problems I am having with the card and faulty vram.
and he shows how downclocking the vram fixed it.
Exactly same issues the heating and stuff and artifacts crahes everything.
It also makes some sense with the rtx lineup blender cuda illegal address errors which the seller mentioned.
Cannot risk files by using the card for work purposes because of the crashes, and nobody knows it would fail completely tomorrow, seller should take it back.
And while downclocking I am also keeping the fans at 70-80% because of the smell and stuff
 
Last edited:
Crazy_Eddy said:
add standoffs
Click to expand...
Thanks for pointing that out.
Already there. The photo is old. Sure, looks messy: P
Sorry for the spam I nearly forgot.

After a mid-long session of gaming it crashed even on -190 on memory this time I turned the fans to auto mode and did +0 on the core and -190 on the memory. I can smell iron like smell from at least a meter away while gaming.
It crashes much early if not underclocked.
https://imgur.com/a/tF9CU3W
 
Last edited by a moderator:
Mod note : Please contact Olx support.

looks like this thing likes to crash even if not on 100% utllization the only solution is to downclock the thing.
 
Last edited by a moderator:
Status
Not open for further replies.
Back
Top