Problem with 4070
log in

Advanced search

Message boards : Number crunching : Problem with 4070

Author Message
Profile bellialuss
Send message
Joined: 27 Dec 14
Posts: 4
Credit: 7,936,889
RAC: 47
Message 9710 - Posted: 5 Mar 2024, 17:10:50 UTC

HI i buy https://www.techpowerup.com/gpu-specs/asus-tuf-rtx-4070-gaming-oc.b11031of course I did undervolting and that's when the problems with the project started. If you can lookup my host and workunits why they ended with error, not all but like couple in bath 100wu. https://srbase.my-firewall.org/sr5/results.php?hostid=216341&offset=0&show_names=0&state=6&appid=

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7479
Credit: 43,662,617
RAC: 41,077
Message 9711 - Posted: 5 Mar 2024, 17:53:27 UTC - in response to Message 9710.

HI i buy https://www.techpowerup.com/gpu-specs/asus-tuf-rtx-4070-gaming-oc.b11031of course I did undervolting and that's when the problems with the project started. If you can lookup my host and workunits why they ended with error, not all but like couple in bath 100wu. https://srbase.my-firewall.org/sr5/results.php?hostid=216341&offset=0&show_names=0&state=6&appid=


You have some valid results so I bet all was fine before undervolting.

Profile bellialuss
Send message
Joined: 27 Dec 14
Posts: 4
Credit: 7,936,889
RAC: 47
Message 9712 - Posted: 5 Mar 2024, 21:10:04 UTC - in response to Message 9711.

I have 90% valid result after undervolt gpu...I lowered the core clocks and left the memory on stock clocks.In tasks ending with an error, there is this message at the bottom of each one:
ERROR: cudaGetLastError() returned 700: an illegal memory access was encountered
2024-03-05 16:24:11 (6524): mfaktc-win-64.exe exited; CPU time 0.734375
2024-03-05 16:24:11 (6524): app exit status: 0x1
2024-03-05 16:24:11 (6524): called boinc_finish(195)

Profile rebirther
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 2 Jan 13
Posts: 7479
Credit: 43,662,617
RAC: 41,077
Message 9713 - Posted: 5 Mar 2024, 21:28:19 UTC - in response to Message 9712.
Last modified: 5 Mar 2024, 21:28:33 UTC

I have 90% valid result after undervolt gpu...I lowered the core clocks and left the memory on stock clocks.In tasks ending with an error, there is this message at the bottom of each one:
ERROR: cudaGetLastError() returned 700: an illegal memory access was encountered
2024-03-05 16:24:11 (6524): mfaktc-win-64.exe exited; CPU time 0.734375
2024-03-05 16:24:11 (6524): app exit status: 0x1
2024-03-05 16:24:11 (6524): called boinc_finish(195)


Do you have all results valid if you run in stock speed? Maybe too less watt for the GPU

Profile bellialuss
Send message
Joined: 27 Dec 14
Posts: 4
Credit: 7,936,889
RAC: 47
Message 9714 - Posted: 5 Mar 2024, 22:41:14 UTC - in response to Message 9713.

I'll try tommorow ASAP :D

tito
Send message
Joined: 30 Dec 14
Posts: 12
Credit: 1,032,262,271
RAC: 27,031
Message 9715 - Posted: 7 Mar 2024, 11:27:56 UTC

How did You perform UV?
I have 4070 as well and no errors at any project.
What I did is lowered power limit to 66% and rised core clock by 100MHz all done in MSI afterburner.
Try lower power limit to level You like, crunch some WU. If all ok then rise core clock by 50 at once til errors appeares.

Profile bellialuss
Send message
Joined: 27 Dec 14
Posts: 4
Credit: 7,936,889
RAC: 47
Message 9786 - Posted: 20 Mar 2024, 10:05:43 UTC - in response to Message 9715.
Last modified: 20 Mar 2024, 10:09:56 UTC

This is my sweetspot for gaming (925mV)...power consuption drop from 200-210W to 130W.
https://prnt.sc/YzqKtg2yQ_MC
https://prnt.sc/tkDIAV7OCvNZ
https://prnt.sc/MyEWhu1v1k74


Post to thread

Message boards : Number crunching : Problem with 4070


Main page · Your account · Message boards


Copyright © 2014-2024 BOINC Confederation / rebirther