Advanced search

Message boards : Number crunching : computation error for long and short gpu wus

Author Message
San-Fernando-Valley
Send message
Joined: 16 Jan 17
Posts: 3
Credit: 20,093,900
RAC: 0
Level
Pro
Scientific publications
wat
Message 51556 - Posted: 23 Feb 2019 | 16:19:50 UTC

What am I doing wrong?
Win10
GPU stock
Hyperthreading off ...

Stderr output

<core_client_version>7.14.2</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -59 (0xffffffc5)</message>
<stderr_txt>
# GPU [Quadro GV100] Platform [Windows] Rev [3212] VERSION [80]
# SWAN Device 0 :
# Name : Quadro GV100
# ECC : Enabled
# Global mem : 32768MB
# Capability : 7.0
# PCI ID : 0000:3B:00.0
# Device clock : 1627MHz
# Memory clock : 850MHz
# Memory width : 4096bit
# Driver version : r418_77 : 41881
#SWAN: FATAL: cannot find image for module [.nonbonded.cu.] for device version 700

</stderr_txt>
]]>

Would appreciate any/every help.

biodoc
Send message
Joined: 26 Aug 08
Posts: 160
Credit: 1,405,920,847
RAC: 396
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 51557 - Posted: 23 Feb 2019 | 18:32:39 UTC - in response to Message 51556.
Last modified: 23 Feb 2019 | 18:34:55 UTC

I believe a new Cuda 10 app will be required for support of Volta and Turing cards.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 289
Credit: 238,117,713
RAC: 123,455
Level
Leu
Scientific publications
wat
Message 51560 - Posted: 23 Feb 2019 | 19:38:22 UTC

There is no compatible science app yet or Turing or Volta.
https://www.gpugrid.net/forum_thread.php?id=4819

San-Fernando-Valley
Send message
Joined: 16 Jan 17
Posts: 3
Credit: 20,093,900
RAC: 0
Level
Pro
Scientific publications
wat
Message 51565 - Posted: 24 Feb 2019 | 12:30:27 UTC

Oh, OK ...

So I'll look for other projects ...

THANKS for replying

mmonnin
Send message
Joined: 2 Jul 16
Posts: 265
Credit: 647,845,139
RAC: 966
Level
Lys
Scientific publications
wat
Message 51566 - Posted: 24 Feb 2019 | 14:02:51 UTC

A GV card should crunch very well at MilkyWay, Collatz, PrimeGrid GFN and Einstein due to their use of some FP64 compute.

Post to thread

Message boards : Number crunching : computation error for long and short gpu wus