Re: [boinc_dev] host punishment mechanism revisited

Seems to me that the really fast methods of catching a broken host aren't very good, and the good methods aren't very fast. Does BOINC need fast? On 5/27/2010 7:19 PM, Josef W. Segur wrote: > I certainly agree, and even were it a perfect method of finding problem > hosts the best method of minim

Re: [boinc_dev] host punishment mechanism revisited

I certainly agree, and even were it a perfect method of finding problem hosts the best method of minimizing their damage to the project is uncertain. Your inputs along that line have seemed very good to me. -- Joe On 27 May 2010 at 20:0

Re: [boinc_dev] Run 1 task and then exit

Grid platforms such as Condor-G use a mechanism known as backfill to integrate BOINC. Does WLCG use Condor as its grid infrastructure? - Rom -Original Message- From: boinc_dev-boun...@ssl.berkeley.edu [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf Of Rod Walker Sent: Thursday,

Re: [boinc_dev] Run 1 task and then exit

Hi, Maybe I am using boinc in a way it was not intended for. I am investigating running boinc on willing remote clusters around the Grid(WLCG). Resource owners get very upset about leaving daemon-like processes, such as boinc, on their WNs after the batch job has left. Therefore boinc must clea

Re: [boinc_dev] [boinc_projects] users circumventing the workunit cache limits

The host record shows 4 cores which means this machine should only be getting 24 total tasks ... The only viable suggestion that came up on the project site is that the user is "spoofing" the number of cores with the CC CONFIG option . Side speculation, the user is setting the value high, getti

Re: [boinc_dev] Run 1 task and then exit

Did you try it with only --exit_when_idle --exit_after_finish? Abort will likely terminate uploads. And, your upload will likely not happen until the next time you run ... because the upload happens after the end of the running task completion ... and if there is only one task, BOINC will be id

[boinc_dev] users circumventing the workunit cache limits

One of our users has appeared to have found a way to get around the workunit cache limit. I was wondering if there's anything we can do about this. milky...@home has an 8 WU per core limit, and for some reason this user: http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=171705

Re: [boinc_dev] Run 1 task and then exit

Hi, The client I started interactively with --exit_when_idle is still running, having processed several tasks. A client started from a batch job with --abort_jobs_on_exit --exit_when_idle --exit_after_finish did finish after one task, but did not upload the result. Do you know of anyone running b

Re: [boinc_dev] host punishment mechanism revisited

The client-side detection of a sequence of tasks completing in suspiciously short times I suggested earlier could probably be done server-side instead. The average and variance of elapsed time/rsc_fpops_est is maintained for each app_version on a host, and could be used to judge whether an elapsed

Re: [boinc_dev] host punishment mechanism revisited

One thing that Bruce Allen had suggested at one point and I think is a good idea as well is to break validation into two stages. The first stage is a validation of everything that can be done looking at a single file. There are a lot of checks that we do during the "init_result" method that chec

Re: [boinc_dev] Compilation Documentation - DevC++

Cross-compilation is something I have not tried yet. Using MSYS to compile was very easy. I updated the documentation to give instructions for MSYS yesterday at David Anderson's request. If you update the doc's to mention the Cygwin implementation and cross-compilation instructions, possibly di

Re: [boinc_dev] host punishment mechanism revisited

Most tasks are validated within days, not weeks. Do you have a suggestion that will actually work better? For all projects? In particular the project that has the problem the worst has no mechanism for telling that there is a problem other than validation. jm7

Re: [boinc_dev] host punishment mechanism revisited

Ah. My memory isn't failing after all. Mo's http://boinc.berkeley.edu/trac/ticket/985 - Original Message - From: "Richard Haselgrove" To: Sent: Tuesday, May 25, 2010 9:20 PM Subject: Re: [boinc_dev] host punishment mechanism revisited > Part of the trouble here is that the BOINC mess

Re: [boinc_dev] host punishment mechanism revisited

"Punishment" for invalid validation will not help at all IMO with broken CUDA GPUs. Tasks reported in seconds scale, validated in weeks scale and time of existance such broken GPU ~day (host reboot will fix it). That is, in most cases this delayed validation accounting will "punish" for long ago

Re: [boinc_dev] Compilation Documentation - DevC++

Shane Reilly wrote, On 26.05.2010 17:07 Uhr: > Compiling it with MSYS worked without problems. > > Is there any way to update the documentation to mention your suggestion > ? > The statement under the MinGW Compilation heading reads, "To build the >