For the next BT release notes:
note that effective today, the BOINC "recommended version" for Win 32/64 & Mac OS X is now 6.10.58 ;)
Quote from: jjwhalen on July 13, 2010, 08:47:15 PM
For the next BT release notes:
note that effective today, the BOINC "recommended version" for Win 32/64 & Mac OS X is now 6.10.58 ;)
Not too much has changed.
Quote from: fred on July 14, 2010, 05:33:32 PM
Quote from: jjwhalen on July 13, 2010, 08:47:15 PM
For the next BT release notes:
note that effective today, the BOINC "recommended version" for Win 32/64 & Mac OS X is now 6.10.58 ;)
Not too much has changed.
It looks like this "recommended" update is mostly Mac-oriented. The only BOINC releasenote 'bullet' I've really been interested in is
QuoteFix: Elasped time tracking issue
and how it might affect BT's performance ???
I've been watching an ongoing issue with BT's display of ET of
long-duration WUs being significantly different (and highly inaccurate) when the task is "Waiting" compared to when it's "Running." Though I've always suspected that BT is only displaying the information it's receiving from the core client. The bad news is that I've been running BOINC 6.10.58 since last week and I see no change in this behavior :(
Quote from: jjwhalen on July 14, 2010, 06:29:13 PM
I've been watching an ongoing issue with BT's display of ET of long-duration WUs being significantly different (and highly inaccurate) when the task is "Waiting" compared to when it's "Running." Though I've always suspected that BT is only displaying the information it's receiving from the core client. The bad news is that I've been running BOINC 6.10.58 since last week and I see no change in this behavior :(
You might want to report it in the project board, I think it's project related.
Quote from: jjwhalen on July 14, 2010, 06:29:13 PM
The only BOINC releasenote 'bullet' I've really been interested in is QuoteFix: Elasped time tracking issue
and how it might affect BT's performance ???
I don't think that will affect that much since this one line code change seems to only affect situations where a WU gets restarted after a checkpoint. (Restart might be due to preempting it, restart of workstation/BOINC, crash or whatever.) Previously the elapsed time starts again from zero but after the fix the elapsed time until the checkpoint is used. Which is correct, since that's where the calculation should continue from anyway.
You might want to debug by enabling
<app_msg_receive> logging to see if the values project returns to BOINC seems sane. If not, you can report it to project. :)
Quote from: fred on July 15, 2010, 05:57:46 AM
Quote from: jjwhalen on July 14, 2010, 06:29:13 PM
I've been watching an ongoing issue with BT's display of ET of long-duration WUs being significantly different (and highly inaccurate) when the task is "Waiting" compared to when it's "Running." Though I've always suspected that BT is only displaying the information it's receiving from the core client. The bad news is that I've been running BOINC 6.10.58 since last week and I see no change in this behavior :(
You might want to report it in the project board, I think it's project related.
It's not just one project.