Welcome to MilkyWay@home

On a single host always Validation Error


Advanced search

Message boards : Number crunching : On a single host always Validation Error
Message board moderation

To post messages, you must log in.

AuthorMessage
gambatesa

Send message
Joined: 23 Feb 18
Posts: 7
Credit: 1,429,004,992
RAC: 4,556,285
1 billion credit badge1 year member badge
Message 67677 - Posted: 20 Jul 2018, 16:32:29 UTC

Dear Sirs,
recently i built a new dual GPU rig to crunch milky@home, This is the host:

https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=775931

all work units crunched always returns as validation error: https://milkyway.cs.rpi.edu/milkyway/results.php?hostid=775931&offset=0&show_names=0&state=5&appid=

I tryed almost everything
-detach and reattach
-uninstall and reinstall boinc
-tried an older boinc version
-a different AMD driver (latest and older)
-install latest updates (but same also before updates)

Temperature are ok (72°C on the hottest GPU)

But Server always says validation error. On client everything seems to be ok but when sent back to server status is always validation error.

What's wrong? everything seems to be ok, is there anything i can check on server side?

Thanks for your help
ID: 67677 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profilemikey
Avatar

Send message
Joined: 8 May 09
Posts: 2228
Credit: 256,158,327
RAC: 165,870
200 million credit badge10 year member badgeextraordinary contributions badge
Message 67678 - Posted: 21 Jul 2018, 10:08:00 UTC - in response to Message 67677.  

Dear Sirs,
recently i built a new dual GPU rig to crunch milky@home, This is the host:

https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=775931

all work units crunched always returns as validation error: https://milkyway.cs.rpi.edu/milkyway/results.php?hostid=775931&offset=0&show_names=0&state=5&appid=

I tryed almost everything
-detach and reattach
-uninstall and reinstall boinc
-tried an older boinc version
-a different AMD driver (latest and older)
-install latest updates (but same also before updates)

Temperature are ok (72°C on the hottest GPU)

But Server always says validation error. On client everything seems to be ok but when sent back to server status is always validation error.

What's wrong? everything seems to be ok, is there anything i can check on server side?

Thanks for your help


Are you running with one cpu core free just for the gpu to use? Are you running multiple wu's at the same time?
ID: 67678 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
gambatesa

Send message
Joined: 23 Feb 18
Posts: 7
Credit: 1,429,004,992
RAC: 4,556,285
1 billion credit badge1 year member badge
Message 67679 - Posted: 21 Jul 2018, 21:05:18 UTC - in response to Message 67678.  

Via app_config.xml there are 4wu/gpu running..
0.24 cpu
0.24 gpu
A maximum of 8 task

No cpu project are running, really i cannot understand.. i have other 3 rigs with same gpus and settings but no problems

Last chance is to reinstall windows7 but before i want to try to solve..
ID: 67679 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
gambatesa

Send message
Joined: 23 Feb 18
Posts: 7
Credit: 1,429,004,992
RAC: 4,556,285
1 billion credit badge1 year member badge
Message 67680 - Posted: 21 Jul 2018, 21:05:26 UTC - in response to Message 67678.  
Last modified: 21 Jul 2018, 21:06:09 UTC

Double post
ID: 67680 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profilemikey
Avatar

Send message
Joined: 8 May 09
Posts: 2228
Credit: 256,158,327
RAC: 165,870
200 million credit badge10 year member badgeextraordinary contributions badge
Message 67681 - Posted: 22 Jul 2018, 12:46:26 UTC - in response to Message 67679.  

Via app_config.xml there are 4wu/gpu running..
0.24 cpu
0.24 gpu
A maximum of 8 task

No cpu project are running, really i cannot understand.. i have other 3 rigs with same gpus and settings but no problems

Last chance is to reinstall windows7 but before i want to try to solve..


Try cutting back to a single wu at a time and see if they are all okay, then add in one more until you get to the unacceptable error rate, then back off one wu at a time. I'm guessing you have a gpu that isn't quite up to the same stuff as the others are.
ID: 67681 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
melk

Send message
Joined: 10 Dec 17
Posts: 47
Credit: 629,890,625
RAC: 35
500 million credit badge1 year member badge
Message 67683 - Posted: 23 Jul 2018, 15:35:53 UTC

Are you willing to upgrade to Windows 10? You can use your existing windows license key and create a USB installer using the Microsoft Media Creation Tool

https://www.microsoft.com/en-us/software-download/windows10

You may need to pull one GPU from the system and see which one is causing the errors. What motherboard are you using?
ID: 67683 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
gambatesa

Send message
Joined: 23 Feb 18
Posts: 7
Credit: 1,429,004,992
RAC: 4,556,285
1 billion credit badge1 year member badge
Message 67708 - Posted: 15 Aug 2018, 14:08:02 UTC

I don't like Windows 10, too many times updated components (especially drivers) caused problems.. Windows 7 correctly configured can just do the job without any sort of headache. This is not last generation hardware so fully supported and cheap (DDR3).

This is my rig:
Asus Maximus Formula VI - Z87
Intel i3 4160 (upgrade planned)
8Gb of DDR3 1600
Intel SSD 120Gb
Windows 7 Sp1 Updated
PSU 1000W 80Gold

I proceeded with a clean install of windows from scratch, Just drivers and some basics and seems to work.. maybe some misconfiguration or something went wrong in past attempt.

It's returning valid results with some queue.

For Milky At this time these GPU are installed:
Sapphire HD7970 OC (Blue PCB, 2x8pin 12V, 1st 16x slot)
Asus HD7990 (Dual GPU, 2nd 16x slot)

the 7990 is in the second slot to have better clean air. I suspect with this configuration the Pcie Lanes of Z87 can bottleneck the 7990.. with 2 phisical gpus installed motherboard gives 8x-8x configuration. The 1st 7970 should have the full 8x, but since 7990 is a dual chip, both chips share the 8x bandwidth, a 4x-4x. I think i'll replace the 7990 with a twin Sapphire HD7970

Have 2 distinct Hosts (A Single7990 and a Dual7970) are a better option and easyer to keep temps in a reasonable range


Ps: The Web user interface of Boinc for Milkyway in my language is completely wrong wrong, worst then first editions of google translate :), it's almost impossible to understand the meaning of buttons, if you are't familiar with english version you have to go blind, bad translated and with spelling errors, It's possible to have contact with someone who can corret the translation?
ID: 67708 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : On a single host always Validation Error

©2019 Astroinformatics Group