Message boards :
Number crunching :
4.45 Broke
Message board moderation
Author | Message |
---|---|
keputnam Send message Joined: 2 Jul 99 Posts: 242 Credit: 2,736,564 RAC: 3 |
The work around for the runaway CPU timer on Win98. I double checked and my settings still say to remove from memory for the Locale that my win98 boxes are in, but CC4.45 keeps incrementing the timer even though the log says the previous WU was removed from memory. So all my Win98 boxes are back on 4.44, and my W2k boxes are running fine on 4.45 |
Steve Cressman Send message Joined: 6 Jun 02 Posts: 583 Credit: 65,644 RAC: 0 |
It's not broke as you call it on my 98se. In fact it is running very well. 98SE XP2500+ @ 2.1 GHz Boinc v5.8.8 And God said"Let there be light."But then the program crashed because he was trying to access the 'light' property of a NULL universe pointer. |
RDC Send message Joined: 17 May 99 Posts: 544 Credit: 1,215,728 RAC: 0 |
It's not broke as you call it on my 98se. I can't even get 4.25 or 4.19 to work right on my second PC that has Win98se on it yet alone any of the 4.4x versions (I've tried each version). The timer still increments the CPU time on all projects at the same time unless I suspend all but the one I want to work on, exit out completely and then reload BOINC (repeating the proceedure daily as I manually switch projects). I use 4.43 on my PC with XP on it and it works fine there and will upgrade to 4.45 depending on how many and what kind of 4.45 problems are reported on the board. To truly explore, one must keep an open mind... |
mikey Send message Joined: 17 Dec 99 Posts: 4215 Credit: 3,474,603 RAC: 0 |
It's not broke as you call it on my 98se. If you need to download the older versions of the software you can try this: http://boinc.berkeley.edu/dl/boinc_4.19_windows_intelx86.exe then change the .19 to say .05 and you can download version 4.05, change the number as you see fit as some numbers will work while others will not. ONLY version 4.** clients are usable so do not try and change the number 4 because even if it works you will not be able to crunch with those real old versions. And YES all the old version as long as they are 4.00 or later, if there was one, should work until Berkeley comes out with version 5. This will ONLY work for Seti stuff, other projects have their own minimum version settings. |
Raithmir Send message Joined: 3 Apr 99 Posts: 89 Credit: 385,065 RAC: 0 |
Or just go to http://boinc.berkeley.edu/dl/ and choose the one you want, rather than trying to guess a version number. Raithmir's SPARC64/UltraSPARC Linux Builds http://www.kulthea.net/boinc/ |
keputnam Send message Joined: 2 Jul 99 Posts: 242 Credit: 2,736,564 RAC: 3 |
It's not broke as you call it on my 98se. This runaway CPU timer is a known bug on Win9x machines. The work-around that has worked on all CC versions up 'til now, is to create an alternate location/locale (I used School) and select "no" for "Leave applications in memory while preempted?" The problem is tied to the way that the Win9x OS handles multi-tasking. |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
It's not broke as you call it on my 98se. And it is further complicated by a fix to keep WUs in memory where the WU has not had a checkpoint since the last time that projects swapped in to the CPU. BOINC WIKI |
stomp Send message Joined: 4 Sep 99 Posts: 8 Credit: 238,321 RAC: 0 |
I came home today and I noticed three work units working at once 6/10/2005 7:12:54 PM||Resuming computation and network activity 6/10/2005 7:12:54 PM||request_reschedule_cpus: Resuming activities 6/10/2005 7:12:54 PM|climateprediction.net|Resuming result 205c_200114456_0 using hadsm3 version 4.12 6/10/2005 7:12:54 PM|ProteinPredictorAtHome|Resuming result h0012A_1_7909_1 using mfoldB125 version 4.28 6/10/2005 7:12:54 PM|ProteinPredictorAtHome|Resuming result h0012A_1_10617_3 using mfoldB125 version 4.28 6/10/2005 7:12:54 PM|BURP|Deferring communication with project for 41 minutes and 24 seconds 6/10/2005 7:13:03 PM||Suspending computation and network activity - user is active 6/10/2005 7:13:03 PM|climateprediction.net|Pausing result 205c_200114456_0 (left in memory) 6/10/2005 7:13:03 PM|ProteinPredictorAtHome|Pausing result h0012A_1_7909_1 (left in memory) 6/10/2005 7:13:03 PM|ProteinPredictorAtHome|Pausing result h0012A_1_10617_3 (left in memory) 6/10/2005 7:13:46 PM||Resuming computation and network activity 6/10/2005 7:13:46 PM||request_reschedule_cpus: Resuming activities 6/10/2005 7:13:46 PM|climateprediction.net|Resuming result 205c_200114456_0 using hadsm3 version 4.12 6/10/2005 7:13:46 PM|ProteinPredictorAtHome|Resuming result h0012A_1_7909_1 using mfoldB125 version 4.28 6/10/2005 7:13:46 PM|ProteinPredictorAtHome|Resuming result h0012A_1_10617_3 using mfoldB125 version 4.28 Once I exit BOINC and restarted BOINC then all is well... 6/10/2005 7:18:48 PM|climateprediction.net|Resuming computation for result 205c_200114456_0 using hadsm3 version 4.12 6/10/2005 7:18:48 PM|climateprediction.net|Deferring computation for result 3ed0_200180175_1 6/10/2005 7:18:48 PM|ProteinPredictorAtHome|Resuming computation for result h0012A_1_7909_1 using mfoldB125 version 4.28 6/10/2005 7:18:48 PM|ProteinPredictorAtHome|Deferring computation for result h0012A_1_10617_3 6/10/2005 7:18:48 PM|Pirates@Home|Sending scheduler request to http://pirates.vassar.edu/cgi-bin/cgi 6/10/2005 7:18:48 PM|Pirates@Home|Requesting 172800 seconds of work, returning 0 results 6/10/2005 7:18:48 PM|climateprediction.net|Pausing result 205c_200114456_0 (left in memory) 6/10/2005 7:18:48 PM|ProteinPredictorAtHome|Restarting result h0012A_1_10617_3 using mfoldB125 version 4.28 |
RDC Send message Joined: 17 May 99 Posts: 544 Credit: 1,215,728 RAC: 0 |
This runaway CPU timer is a known bug on Win9x machines. The work-around that has worked on all CC versions up 'til now, is to create an alternate location/locale (I used School) and select "no" for Thanks! It seems to switch now and stop the timers like it's supposed to. I recently inherited the Win98se PC and it's been a while since I had to deal with that particular troublesome O/S so couldn't recall how I had my main computer set up (where BOINC worked fine) when it still had Win98se on it. The main has had XP on it for a while now. To truly explore, one must keep an open mind... |
keputnam Send message Joined: 2 Jul 99 Posts: 242 Credit: 2,736,564 RAC: 3 |
Looks like I may have jumped to a conclusion. After seeing comments from other Win98 users that said they had no problems, I did an uninstall/boot/install/boot. (the first time I had just done an upgrade-in-place) I still had the problem, so I did a little more digging. Of the four projects on this box, they were pointed at three different locales. I did some XML editing to point them all back to "school" the way they were supposed to be, and now they switch and the timer only increments on the "Running" WU. I didn't know that the same machine could be in diffent locales. I though the last project updated migrated it's settings to all others. (I was using the general settings, not project specific) |
Steve Cressman Send message Joined: 6 Jun 02 Posts: 583 Credit: 65,644 RAC: 0 |
Also boinc will run better if you take it out of startup and do a manual start of boinc after win98 is completely booted. 98SE XP2500+ @ 2.1 GHz Boinc v5.8.8 And God said"Let there be light."But then the program crashed because he was trying to access the 'light' property of a NULL universe pointer. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.