Yay, I found an issue ! (grin)
OK I am not sure if this is BOINC or either Rosetta or Einstein. I have allocated the same crunching time to both - 1 hour and then a change back to the other etc. However it would appear that Rosetta is getting most of the time - granted Rosetta is a much smaller WU, but the deadlines for Rosetta were around mid November, while Einstein was a week or so earlier, so by my understanding of BOINCs ability, surely the earliest deadline has priority. Yet poor old Einstein seems to be left with just a cpu cycle or two to scrabble together (grin).
Do you need to go through all that deleteing just to install BOINC 5.2.6,? Is it o.k. to just install over the old version as in updating the system.
I've installed the new BOINC over the old one for about 5 upgrades now - starting with 4.45 over 4.19 and most recently 5.2.6 over 5.2.5. I've never had a problem.
Hello
Nope you don't need to do that - the only and sole reason for me doing that is to avoid BOINC picking up any previous info from the previous install - I want to be able to evaluate the BOINC version on it's own merits, without having to factor in the various timings of WUs etc that I had done before.
Yay, I found an issue ! (grin)
OK I am not sure if this is BOINC or either Rosetta or Einstein. I have allocated the same crunching time to both - 1 hour and then a change back to the other etc. However it would appear that Rosetta is getting most of the time - granted Rosetta is a much smaller WU, but the deadlines for Rosetta were around mid November, while Einstein was a week or so earlier, so by my understanding of BOINCs ability, surely the earliest deadline has priority. Yet poor old Einstein seems to be left with just a cpu cycle or two to scrabble together (grin).
Any ideas here ?
Gray
I have just adjusted Einsteins resources so that the ratio is 80% Einstein to 20% Rosetta to see if this help
OK I am not sure if this is BOINC or either Rosetta or Einstein.
The only thing responsible for work allocation is BOINC.
Quote:
However it would appear that Rosetta is getting most of the time -
Your results list actually shows quite regular work for EAH over the last week or more. Your perception is a little faulty, I think :).
Quote:
... so by my understanding of BOINCs ability, surely the earliest deadline has priority.
Not usually. Only if there is a risk of missing a deadline, in which case BOINC will go into EDF mode to prevent that. If there is no risk of missing a deadline, BOINC will simply allocate time to each project in accordance with the resource shares you have set. If one project has had more than its fair share, BOINC will make sure that the other project runs until the balance is restored. That must be why Rosetta is getting extra time at the moment. If you learn patience and stop trying to micromanage BOINC based on faulty perceptions, BOINC will do an excellent job of "getting it right" :).
Quote:
I have just adjusted Einsteins resources so that the ratio is 80% Einstein to 20% Rosetta to see if this help
BOINC will now do an excellent job of giving 80% of the time to EAH and 20% of the time to Rosetta - until you next fiddle with the settings and confuse poor old BOINC!! :).
It may not do it immediately. It depends on how much debt is owed to Rosetta, but it wont be too long before that is repaid and EAH starts becoming the dominant project.
One other comment that needs to be publicised in case others think that your upgrade procedure was a good idea. Your action of completely cleansing all previous config and state information when you made the upgrade has just forced BOINC to gather all that useful information all over again. This is fine if you are really into micromanaging and studying in intimate detail the inner workings of BOINC. However it's just a waste of valuable data as far as the average user is concerned.
For the average user, a Windows uninstall is very easy and helps keep the registry clean, and is really needed if coming from a very early 4.xx (eg 4.19). Please don't manually delete anything. If coming from a previous 5.2.X version, installing "over the top" is fine. There is a "sticky" containing full upgrade instructions on the "Problems" message board.
OK I am not sure if this is BOINC or either Rosetta or Einstein.
The only thing responsible for work allocation is BOINC.
Quote:
However it would appear that Rosetta is getting most of the time -
Your results list actually shows quite regular work for EAH over the last week or more. Your perception is a little faulty, I think :).
Quote:
... so by my understanding of BOINCs ability, surely the earliest deadline has priority.
Not usually. Only if there is a risk of missing a deadline, in which case BOINC will go into EDF mode to prevent that. If there is no risk of missing a deadline, BOINC will simply allocate time to each project in accordance with the resource shares you have set. If one project has had more than its fair share, BOINC will make sure that the other project runs until the balance is restored. That must be why Rosetta is getting extra time at the moment. If you learn patience and stop trying to micromanage BOINC based on faulty perceptions, BOINC will do an excellent job of "getting it right" :).
Quote:
I have just adjusted Einsteins resources so that the ratio is 80% Einstein to 20% Rosetta to see if this help
BOINC will now do an excellent job of giving 80% of the time to EAH and 20% of the time to Rosetta - until you next fiddle with the settings and confuse poor old BOINC!! :).
It may not do it immediately. It depends on how much debt is owed to Rosetta, but it wont be too long before that is repaid and EAH starts becoming the dominant project.
One other comment that needs to be publicised in case others think that your upgrade procedure was a good idea. Your action of completely cleansing all previous config and state information when you made the upgrade has just forced BOINC to gather all that useful information all over again. This is fine if you are really into micromanaging and studying in intimate detail the inner workings of BOINC. However it's just a waste of valuable data as far as the average user is concerned.
For the average user, a Windows uninstall is very easy and helps keep the registry clean, and is really needed if coming from a very early 4.xx (eg 4.19). Please don't manually delete anything. If coming from a previous 5.2.X version, installing "over the top" is fine. There is a "sticky" containing full upgrade instructions on the "Problems" message board.
Hello Gary
Thank you for that (very) comprehensive reply. As I mentioned below, I AM deliberately looking for BOINC problems in the hopes of getting them sorted - LOVE fiddling with software - gives me an excuse to not do any actual work (grin). The frustrating part is that I can't really find many problems now... (big grin).
So, in order to be fair to the latest version of BOINC I do the abort unfinished WUs, uninstall, folder delete, registry clean and re-install, re-attach routine. I admit to expecting an immediate balance between the clients tho' so at least my knowledge has been improved here.
Thank You again. I'll just leave things as they are for now (well...maybe - grin) and - BTW - do you know why, when I signed up with Rosetta I got this error in the messages: Missing account key. The attaching went through fine however.
I should have mentioned that the join up with Rosetta happened over Saturday and I was concerned that just after joining, Einstein seemed to be put to one side while Rosetta took over.
Funnily enough, I remember the exact opposite happening, when I was running SETI initially, and then signed up with Einstein, using version 5.2.2 (I think it was). I found Einstein all loaded up, but no crunching being done.
I should have mentioned that the join up with Rosetta happened over Saturday and I was concerned that just after joining, Einstein seemed to be put to one side while Rosetta took over.
Now things are a lot clearer. You joined Rosetta less than 24 hours ago? At that time BOINC would have known that it had been working on EAH so it probably started with an initial bias towards Rosetta just to let it "catch up" a bit. As Rosetta work is a lot shorter, you could easily have finished several before finishing your next EAH result, thereby giving the perception of doing more on Rosetta than EAH. BOINC would need at least a day or two to settle into a comfortable pattern and during that time may well spend what seems to be too much time on a given project.
Quote:
Funnily enough, I remember the exact opposite happening, when I was running SETI initially, and then signed up with Einstein, using version 5.2.2 (I think it was). I found Einstein all loaded up, but no crunching being done.
When you first started posting here you were trying 5.2.1 but at one stage you mentioned reverting to 4.25. At that time, about three weeks ago, you were having some problem with the EAH password as I recall.
Quote:
BTW - do you know why, when I signed up with Rosetta I got this error in the messages: Missing account key. The attaching went through fine however.
You actually reported a similar thing at the time you added EAH. Here is a quote from your message at that time (16 October):-
Quote:
OK now I am confused - the new client 5.2.1, which accessed SETI@home with no issues (still downloading WUs as I type this - around 35 of the lil blighters) does something a bit weird. I tried the "attach to project" wizard for einstein@home, and right at the end I was informed that the attaching had failed. I was staring in frustration at the screen, when suddenly einstein@home shows up in the project list.
By the sound of all this, you have attached to projects multiple times, both under the old (25 char random key) and new (user settable password) systems and perhaps you have done things that are causing some sort of bad interaction between the two. Apart from that I've no idea why you are having problems.
Sorry about the delay in my reply - just finished a reformat and re-install along with updates etc (what a PAIN on a dialup). Done tweaking system etc. BOINC back on song. I was trying to get the windows folder size as low as possible to speed up processing (got as far as 887 megs, and ran out of ideas...). The benchmarks seem OK I guess at
1376 double precision MIPS (Whetstone) per CPU
2798 integer MIPS (Dhrystone) per CPU
for a lil 2.6 cpu - would LOVE one of those multiple-cpu server boards - sigh.
That error I mentioned (Missing account key) seems to occur every time I join a project, and I believe I am not the only one: I'm sure I saw it mentioned elsewhere on the message boards - anyhow, it's not serious and I'm not gonna format again just to test this or similar. I suspect there is some error in the codebase where it is expecting the old account key as per the 4.xx series, but as the attaching works who am I to complain - happy crunchin'
You mention just one PC in your specs at the top - are you shoving RAM in and out and doing multiple re-installs, or are you the lucky possessor of lots of PCs (if this seems a daft statement - sorry I just woke up and the coffee hasn't kicked in yet..... Where I come from PCs are amazingly expensive).
Did you do any tweaking to the system/s at all, or are we talking about a stock "as recomended by Mr Bill Gates" install in all cases. While I don't suggest ripping files out of the Windows folder without any forethought, certainly the smaller that folder, the better (hence mine-own re-installs, much to the amusement of my friends) - there are certain files I remove as a matter of course tho'.
Also dropping the amount of running services in XP makes a difference, obviously, tho' again caution (and research) is required. There are one or two registry tweaks too, and things like no wall-paper, bare minimum desktop stuff (shove large folders elsewhere and make a shortcut rather) - hey I'm sure you know all this stuff, if you run all those boxes.
Yay, I found an issue !
)
Yay, I found an issue ! (grin)
OK I am not sure if this is BOINC or either Rosetta or Einstein. I have allocated the same crunching time to both - 1 hour and then a change back to the other etc. However it would appear that Rosetta is getting most of the time - granted Rosetta is a much smaller WU, but the deadlines for Rosetta were around mid November, while Einstein was a week or so earlier, so by my understanding of BOINCs ability, surely the earliest deadline has priority. Yet poor old Einstein seems to be left with just a cpu cycle or two to scrabble together (grin).
Any ideas here ?
Gray
RE: RE: Do you need to go
)
Hello
Nope you don't need to do that - the only and sole reason for me doing that is to avoid BOINC picking up any previous info from the previous install - I want to be able to evaluate the BOINC version on it's own merits, without having to factor in the various timings of WUs etc that I had done before.
Gray
RE: Yay, I found an issue !
)
I have just adjusted Einsteins resources so that the ratio is 80% Einstein to 20% Rosetta to see if this help
Gray
RE: Yay, I found an issue !
)
Actually, no, you haven't :).
The only thing responsible for work allocation is BOINC.
Your results list actually shows quite regular work for EAH over the last week or more. Your perception is a little faulty, I think :).
Not usually. Only if there is a risk of missing a deadline, in which case BOINC will go into EDF mode to prevent that. If there is no risk of missing a deadline, BOINC will simply allocate time to each project in accordance with the resource shares you have set. If one project has had more than its fair share, BOINC will make sure that the other project runs until the balance is restored. That must be why Rosetta is getting extra time at the moment. If you learn patience and stop trying to micromanage BOINC based on faulty perceptions, BOINC will do an excellent job of "getting it right" :).
BOINC will now do an excellent job of giving 80% of the time to EAH and 20% of the time to Rosetta - until you next fiddle with the settings and confuse poor old BOINC!! :).
It may not do it immediately. It depends on how much debt is owed to Rosetta, but it wont be too long before that is repaid and EAH starts becoming the dominant project.
One other comment that needs to be publicised in case others think that your upgrade procedure was a good idea. Your action of completely cleansing all previous config and state information when you made the upgrade has just forced BOINC to gather all that useful information all over again. This is fine if you are really into micromanaging and studying in intimate detail the inner workings of BOINC. However it's just a waste of valuable data as far as the average user is concerned.
For the average user, a Windows uninstall is very easy and helps keep the registry clean, and is really needed if coming from a very early 4.xx (eg 4.19). Please don't manually delete anything. If coming from a previous 5.2.X version, installing "over the top" is fine. There is a "sticky" containing full upgrade instructions on the "Problems" message board.
Cheers,
Gary.
RE: RE: Yay, I found an
)
Hello Gary
Thank you for that (very) comprehensive reply. As I mentioned below, I AM deliberately looking for BOINC problems in the hopes of getting them sorted - LOVE fiddling with software - gives me an excuse to not do any actual work (grin). The frustrating part is that I can't really find many problems now... (big grin).
So, in order to be fair to the latest version of BOINC I do the abort unfinished WUs, uninstall, folder delete, registry clean and re-install, re-attach routine. I admit to expecting an immediate balance between the clients tho' so at least my knowledge has been improved here.
Thank You again. I'll just leave things as they are for now (well...maybe - grin) and - BTW - do you know why, when I signed up with Rosetta I got this error in the messages: Missing account key. The attaching went through fine however.
Gray
Hello Gary I should have
)
Hello Gary
I should have mentioned that the join up with Rosetta happened over Saturday and I was concerned that just after joining, Einstein seemed to be put to one side while Rosetta took over.
Funnily enough, I remember the exact opposite happening, when I was running SETI initially, and then signed up with Einstein, using version 5.2.2 (I think it was). I found Einstein all loaded up, but no crunching being done.
Anyhow, will leave all as is for now
Gray
RE: I should have mentioned
)
Now things are a lot clearer. You joined Rosetta less than 24 hours ago? At that time BOINC would have known that it had been working on EAH so it probably started with an initial bias towards Rosetta just to let it "catch up" a bit. As Rosetta work is a lot shorter, you could easily have finished several before finishing your next EAH result, thereby giving the perception of doing more on Rosetta than EAH. BOINC would need at least a day or two to settle into a comfortable pattern and during that time may well spend what seems to be too much time on a given project.
When you first started posting here you were trying 5.2.1 but at one stage you mentioned reverting to 4.25. At that time, about three weeks ago, you were having some problem with the EAH password as I recall.
You actually reported a similar thing at the time you added EAH. Here is a quote from your message at that time (16 October):-
By the sound of all this, you have attached to projects multiple times, both under the old (25 char random key) and new (user settable password) systems and perhaps you have done things that are causing some sort of bad interaction between the two. Apart from that I've no idea why you are having problems.
Cheers,
Gary.
Hello Gary Sorry about the
)
Hello Gary
Sorry about the delay in my reply - just finished a reformat and re-install along with updates etc (what a PAIN on a dialup). Done tweaking system etc. BOINC back on song. I was trying to get the windows folder size as low as possible to speed up processing (got as far as 887 megs, and ran out of ideas...). The benchmarks seem OK I guess at
1376 double precision MIPS (Whetstone) per CPU
2798 integer MIPS (Dhrystone) per CPU
for a lil 2.6 cpu - would LOVE one of those multiple-cpu server boards - sigh.
That error I mentioned (Missing account key) seems to occur every time I join a project, and I believe I am not the only one: I'm sure I saw it mentioned elsewhere on the message boards - anyhow, it's not serious and I'm not gonna format again just to test this or similar. I suspect there is some error in the codebase where it is expecting the old account key as per the 4.xx series, but as the attaching works who am I to complain - happy crunchin'
Gray
Well here is the result of
)
Well here is the result of the installation...
Asus P4P800-MX Motherboard
Intel Pentium 4 3Ghz
PC3200/400mhz 184 Pin DDR DIMM 512Meg Ram
Maxtor 40Gig Hard Drive
--------------------------------------------------------------------------------
Machine on WindowsXP/Multithreading on, 256Meg Ram Boinc 5.2.6
10/30/2005 02:15:31 PM||Running CPU benchmarks
10/30/2005 02:16:30 PM||Benchmark results:
10/30/2005 02:16:30 PM||Number of CPUs: 2
10/30/2005 02:16:30 PM||1303 double precision MIPS (Whetstone) per CPU
10/30/2005 02:16:30 PM||1599 integer MIPS (Dhrystone) per CPU
10/30/2005 02:16:30 PM||Finished CPU benchmarks
--------------------------------------------------------------------------------
Machine on Windows98/Multithreading on, 256Meg Ram Boinc 4.45.0
10/31/2005 06:09:49 PM||Running CPU benchmarks
10/31/2005 06:10:46 PM||Benchmark results:
10/31/2005 06:10:46 PM||Number of CPUs: 1
10/31/2005 06:10:46 PM||1379 double precision MIPS (Whetstone) per CPU
10/31/2005 06:10:46 PM||2242 integer MIPS (Dhrystone) per CPU
10/31/2005 06:10:46 PM||Finished CPU benchmarks
--------------------------------------------------------------------------------
Machine on Windows98/Multithreading off, 256Meg Ram Boinc 4.45.0
11/01/2005 07:59:36 PM||Running CPU benchmarks
11/01/2005 08:00:33 PM||Benchmark results:
11/01/2005 08:00:33 PM||Number of CPUs: 1
11/01/2005 08:00:33 PM||1478 double precision MIPS (Whetstone) per CPU
11/01/2005 08:00:33 PM||2294 integer MIPS (Dhrystone) per CPU
--------------------------------------------------------------------------------
Machine on Windows98/Multithreading off, 512Meg Ram Boinc 4.45.0
11/05/2005 06:09:29 PM||Running CPU benchmarks
11/05/2005 06:10:26 PM||Benchmark results:
11/05/2005 06:10:26 PM||Number of CPUs: 1
11/05/2005 06:10:26 PM||1468 double precision MIPS (Whetstone) per CPU
11/05/2005 06:10:26 PM||2190 integer MIPS (Dhrystone) per CPU
11/05/2005 06:10:26 PM||Finished CPU benchmarks
--------------------------------------------------------------------------------
Machine on Windows98/Multithreading off, 512Meg Ram Boinc 5.2.6
11/07/2005 07:38:45 AM||Starting BOINC client version 5.2.6 for windows_intelx86
11/07/2005 07:38:45 AM||libcurl/7.14.0 OpenSSL/0.9.8 zlib/1.2.3
11/07/2005 07:38:45 AM||Data directory: C:\\PROGRAM FILES\\BOINC
11/07/2005 07:38:46 AM||Processor: 1 GenuineIntel x86 Family 15 Model 4 Stepping 1
11/07/2005 07:38:46 AM||Memory: 502.76 MB physical, 1.51 GB virtual
11/07/2005 07:38:46 AM||Disk: 37.26 GB total, 36.48 GB free
11/07/2005 07:38:46 AM||Version change detected (4.45.0 -> 5.2.6); running CPU benchmarks
11/07/2005 07:38:52 AM||Running CPU benchmarks
11/07/2005 07:39:51 AM||Benchmark results:
11/07/2005 07:39:51 AM||Number of CPUs: 1
11/07/2005 07:39:51 AM||1401 double precision MIPS (Whetstone) per CPU
11/07/2005 07:39:51 AM||2431 integer MIPS (Dhrystone) per CPU
11/07/2005 07:39:51 AM||Finished CPU benchmarks
All in all not to bad...any insights? Thank for the help.
Greg
Hello You mention just one
)
Hello
You mention just one PC in your specs at the top - are you shoving RAM in and out and doing multiple re-installs, or are you the lucky possessor of lots of PCs (if this seems a daft statement - sorry I just woke up and the coffee hasn't kicked in yet..... Where I come from PCs are amazingly expensive).
Did you do any tweaking to the system/s at all, or are we talking about a stock "as recomended by Mr Bill Gates" install in all cases. While I don't suggest ripping files out of the Windows folder without any forethought, certainly the smaller that folder, the better (hence mine-own re-installs, much to the amusement of my friends) - there are certain files I remove as a matter of course tho'.
Also dropping the amount of running services in XP makes a difference, obviously, tho' again caution (and research) is required. There are one or two registry tweaks too, and things like no wall-paper, bare minimum desktop stuff (shove large folders elsewhere and make a shortcut rather) - hey I'm sure you know all this stuff, if you run all those boxes.
Gray