Questions and Answers : Windows : BOINC 5.2.2 released
Message board moderation
Author | Message |
---|---|
Send message Joined: 13 Sep 04 Posts: 228 Credit: 354,979 RAC: 0 |
BOINC 5.2.2 (Windows) is now available from <A HREF=\"http://boinc.berkeley.edu/download.php\">http://boinc.berkeley.edu/download.php</A>. As always, make sure to have a full backup of your BOINC folder before upgrading. |
Send message Joined: 13 Sep 04 Posts: 228 Credit: 354,979 RAC: 0 |
So far I have found two problems with this 5.2.2 release: - There is no Hide function on the menu. I can merely minimize it to the taskbar, but not hide it to the systray. I find this a bit annoying. - The \'To completion\' time is incorrectly calculated. I have currently 23.43% progress on a sulphur cycle, with 497 hours crunched. \'To completion\' shows 2,656 hours, which is 125%, according to my calculation. The progress percentage is the same as with 4.45, but the \'To completion\' was a lot less (like 1,950 hours). |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
- There is no Hide function on the menu. I can merely minimize it to the taskbar, but not hide it to the systray. I find this a bit annoying. It\'s not intuitive, but the close option removes it from the task bar and leaves it in the system tray. - The \'To completion\' time is incorrectly calculated. I have currently 23.43% progress on a sulphur cycle, with 497 hours crunched. \'To completion\' shows 2,656 hours, which is 125%, according to my calculation. The progress percentage is the same as with 4.45, but the \'To completion\' was a lot less (like 1,950 hours). The same over-estimate of the \'To completion\' time was seen in pre-release testing of the coupled model spinup application using 5.2.1. "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 5 Aug 04 Posts: 426 Credit: 2,426,069 RAC: 0 |
The \'to completion\' calculations have been modified to generate a more accurate estimate without the radical changes some projects produce. The hitch is BOINC must learn how well your system processes work for a particular project for this code to work correctly. That means after your system processes a few workunits it should be more accurate than ever before, but in CPDN a few workunits could mean a year in real time. BOINC WIKI BOINCing since 2002/12/8 |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
Aren\'t these massive over-estimates of the processing time required for CPDN results likely to cause hosts running multiple projects to go into EDF scheduling mode unnecessarily? "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 30 Aug 04 Posts: 142 Credit: 9,936,132 RAC: 0 |
So far I have found two problems with this 5.2.2 release: I had noticed it in 5.2.1. Annoying is right. If you click on Exit, it just shuts everything down. Workaround: I click on the rightmost cross (close window). That way, I only have the BOINC icon in the systray, as of old. A final quibble though. Hovering over the icon doesn\'t show you the updated percentages (at least in 5.2.1, I\'ll let you know if it\'s fixed in 5.2.2). The stats seem to remain frozen at the last time you opened BOINC. Forum search Site search |
Send message Joined: 7 Aug 04 Posts: 2187 Credit: 64,822,615 RAC: 5,275 |
- The \'To completion\' time is incorrectly calculated. I have currently 23.43% progress on a sulphur cycle, with 497 hours crunched. \'To completion\' shows 2,656 hours, which is 125%, according to my calculation. The progress percentage is the same as with 4.45, but the \'To completion\' was a lot less (like 1,950 hours). I\'m seeing this also, and looking at it as a person who helps on the forums quite a bit, this problem is going to be a rather big annoyance. People new to CPDN already complain about how long the WUs are, now they will see the estimated time to completion and think it takes a several hundred (or more)hours longer than it actually does, and may abort the WU. Others who have been running it awhile, or are more proactive in asking questions, will come to the forums and ask the question of why it takes so long, over and over again. Then there is the potential problem Thyme points out, this may force EDF scheduling, and if so, they may abort CPDN for that reason alone. Edit...How the heck does it come up with the initial estimated time to completion anyway? It is so incredibly inaccurate for my systems that one has to wonder... |
Send message Joined: 8 Jul 05 Posts: 33 Credit: 1,274,211 RAC: 0 |
So far I have found two problems with this 5.2.2 release: I found when changing to 5.2.2 that while the time taken remained the same, the time to completion dropped by 100 hours - must be a better estimating device in there somewhere. |
Send message Joined: 5 Aug 04 Posts: 426 Credit: 2,426,069 RAC: 0 |
...How the heck does it come up with the initial estimated time to completion anyway? It is so incredibly inaccurate for my systems that one has to wonder... The initial estimate is provided by the projects. See this nice short page in the wiki I need to improve at some point. http://boinc-doc.net/boinc-wiki/index.php?title=Initial_Duration_Estimate BOINC WIKI BOINCing since 2002/12/8 |
©2025 cpdn.org