GPU client help

benjo · 30 Jul 2008 at 00:59

Thanks, i'll give it a try

benjo · 31 Jul 2008 at 02:18

Just finished an SMP WU, but it's having trouble sending

Code:

Finished Work Unit:
[01:06:36] - Reading up to 3726768 from "work/wudata_02.arc": Read 3726768
[01:06:36] - Reading up to 1942532 from "work/wudata_02.xtc": Read 1942532
[01:06:36] goefile size: 0
[01:06:36] logfile size: 46113
[01:06:36] Leaving Run
[01:06:37] - Writing 5721061 bytes of core data to disk...
[01:06:37]   ... Done.
[01:06:37] - Failed to delete work/wudata_02.sas
[01:06:37] - Failed to delete work/wudata_02.goe
[01:06:37] Warning:  check for stray files
[01:06:37] - Shutting down core
[01:08:37] 
[01:08:37] Folding@home Core Shutdown: FINISHED_UNIT
[01:08:37] 
[01:08:37] Folding@home Core Shutdown: FINISHED_UNIT
[01:08:41] CoreStatus = 64 (100)
[01:08:41] Sending work to server
[01:08:41] Project: 2653 (Run 2, Clone 25, Gen 81)


[01:08:41] + Attempting to send results [July 31 01:08:41 UTC]
[01:08:42] - Couldn't send HTTP request to server
[01:08:42] + Could not connect to Work Server (results)
[01:08:42]     (171.64.65.64:8080)
[01:08:42] + Retrying using alternative port
[01:08:44] - Couldn't send HTTP request to server
[01:08:44] + Could not connect to Work Server (results)
[01:08:44]     (171.64.65.64:80)
[01:08:44] - Error: Could not transmit unit 02 (completed July 31) to work server.
[01:08:44]   Keeping unit 02 in queue.
[01:08:44] Project: 2653 (Run 2, Clone 25, Gen 81)


[01:08:44] + Attempting to send results [July 31 01:08:44 UTC]
[01:08:46] - Couldn't send HTTP request to server
[01:08:46] + Could not connect to Work Server (results)
[01:08:46]     (171.64.65.64:8080)
[01:08:46] + Retrying using alternative port
[01:08:47] - Couldn't send HTTP request to server
[01:08:47] + Could not connect to Work Server (results)
[01:08:47]     (171.64.65.64:80)
[01:08:47] - Error: Could not transmit unit 02 (completed July 31) to work server.


[01:08:47] + Attempting to send results [July 31 01:08:47 UTC]
[01:12:57] - Server does not have record of this unit. Will try again later.
[01:12:57]   Could not transmit unit 02 to Collection server; keeping in queue.
[01:12:57] - Preparing to get new work unit...
[01:12:57] + Attempting to get work packet
[01:12:57] - Connecting to assignment server
[01:12:58] - Successful: assigned to (171.64.65.63).
[01:12:58] + News From Folding@Home: Welcome to Folding@Home
[01:12:58] Loaded queue successfully.
[01:13:03] Project: 2653 (Run 2, Clone 25, Gen 81)


[01:13:03] + Attempting to send results [July 31 01:13:03 UTC]
[01:13:04] - Couldn't send HTTP request to server
[01:13:04] + Could not connect to Work Server (results)
[01:13:04]     (171.64.65.64:8080)
[01:13:04] + Retrying using alternative port
[01:13:06] - Couldn't send HTTP request to server
[01:13:06] + Could not connect to Work Server (results)
[01:13:06]     (171.64.65.64:80)
[01:13:06] - Error: Could not transmit unit 02 (completed July 31) to work server.

I guess the server is down or something? Will i lose this WU?

TheKnat · 31 Jul 2008 at 06:36

Probably just the work server being down, wouldn't worry about it really.

Yeggstry · 31 Jul 2008 at 08:17

The one thing I find frustrating at times is the whole sending/receiving works units system. You would imagine that its one of the most important parts of this, as without it there would be nothing to fold. However, there have been quite a few instances recently (last couple of months) where my clients, both GPU and CPU, have been idle for hours because I can't connect to the server

Biffa · 31 Jul 2008 at 08:30

Thats sort of the beauty of the BOINC system, you can have multiple projects and if one isn't crunching for some reason or another then the other project can kick off.

TheKnat · 31 Jul 2008 at 09:24

Yeggstry said:
The one thing I find frustrating at times is the whole sending/receiving works units system. You would imagine that its one of the most important parts of this, as without it there would be nothing to fold. However, there have been quite a few instances recently (last couple of months) where my clients, both GPU and CPU, have been idle for hours because I can't connect to the server

Yeah that annoys me as well. The other day I had 4 different clients all waiting to connect to the server, with the GPU's being the worst culprits for it.

Sometimes I find restarting the client seems to force it to look at a different IP address and it will then pick up some new work to do, even if the completed work is still in a queue to be sent.