Welcome to MilkyWay@home

Posts by Conan

21) Message boards : Number crunching : Ati 4200 iGPU (Message 34819)
Posted 24 Dec 2009 by ProfileConan
Post:
Hey there guys!
Can any one tell me if the integrated 4200 is able to process MW work nits?



See the thread Titled ATI GPUs
One of the postings has a list of compatible devices.
22) Message boards : Number crunching : Unexpected App of Unknown Origin (Message 34524)
Posted 14 Dec 2009 by ProfileConan
Post:
I shutdown BOINC and edited my app_info.xml file. When I restarted BOINC the client deleted the WU I had been working on, but when it downloaded a new WU BOINC Manager showed the app to be milkyway 0.20. Strange bug, that. One would think the client would identify the fastest app and go with that.

In reading not just the above but also some of your earlier statements, there are quite a few things you are getting confused about.

Firstly in the above quote, you say that showing the app in BOINC manager as "milkyway 0.20" is a bug. In fact there is no bug and it is perfectly correct because that is exactly what your app_info.xml says the description in BOINC manager should be. I have explained this in response to another question so please feel free to read the explanation of why this is so here. Warning: You may need to start at the beginning of the thread to get the full picture.

Secondly, you say that the client should identify the fastest app and go with that. How is it supposed to do that? The whole point of AP is to make sure that both the client and the server have ZERO choice in the matter. The precise app to be used is specified in app_info.xml. It is NEVER downloaded by the client. It MUST exist in the project folder before the client starts up. You will always get an error if the app specified in app_info.xml doesn't exist.

You said in an earlier message that an app was downloaded with a task. When that happened, you could not have had a valid app_info.xml in place and what was downloaded would have been the normal stock app and not an optimised one. You were correct to get rid of that app and start again.

The app_info.xml file you published implies the following facts/conditions

  • You must supply an app called 'astronomy_0.20_SSE.exe'.
  • The app will be called 'milkyway 0.20' when displayed in BOINC manager.
  • You must place the app called 'astronomy_0.20_SSE.exe' in the MW project folder.
  • This program must already exist before the client is started.
  • This program will never be downloaded by a client request.
  • If this program doesn't exist, it will be a fatal error.
  • This program can be used to crunch any existing tasks already 'branded' as 0.19 or 0.20
  • Any new tasks downloaded will be branded 0.20


If you need further details about this, please read the message I linked in the 2nd paragraph above.

Also, you were given incorrect information that version 20 has to be listed before version 19. The natural way is to list 19 first but it will work either way. BOINC always assumes that the bigger number numerically is the latest version.



G'Day Gary,
Yes I know that it should make no difference and on other projects where I have used an app_info.xml file I have placed the various apps listed from oldest to youngest (some of the Einstein app_info files from a while back were very long so as to include all the various apps running at the time with the Beta releases).
However I have not been the only one to have to had modified my app_info file both here and I think Collatz was the other place, in order to get the latest app to download.
As for incorrect information, well if the syntax is wrong that is unfortunate, BOINC is an amazing thing that is often hard to fathom, but as you have seen it worked and it works for me also.

I currently get round the problem by running out all the old work and then just adding the new app_info.xml file with only the new app listed in it.

Conan.
23) Message boards : Number crunching : Unexpected App of Unknown Origin (Message 34474)
Posted 12 Dec 2009 by ProfileConan
Post:
Because of a couple of issues it was necessary for me to detach from and reattach to MW. I had been using the SSE opti app although, after some time, the name changed from astronomy 0.20 to milkyway 0.20. So after I reattached I again downloaded the SSE version of the opti app. What follows is the app_info.xml file:

<app_info>
<app>
<name>milkyway</name>
</app>
<file_info>
<name>astronomy_0.20_SSE.exe</name>
<executable/>
</file_info>

<app_version>
<app_name>milkyway</app_name>
<version_num>19</version_num>
<file_ref>
<file_name>astronomy_0.20_SSE.exe</file_name>
<main_program/>
</file_ref>
</app_version>

<app_version>
<app_name>milkyway</app_name>
<version_num>20</version_num>
<file_ref>
<file_name>astronomy_0.20_SSE.exe</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

When I downloaded a WU the app was named milkyway 0.19. The first time this happened the app had downloaded with the WU, since milkyway 0.19 had appeared in the ...projects\milkyway.cs.rpi.edu_milkyway subdirectory. So I deleted milkyway 0.19 from the subdirectory, aborted the WU, and downloaded another WU. This time the app named in BOINC Manager again was milkyway 0.19 (the complete name was milkyway_0.19_windows_intelx86.exe), but the app didn't show up in ...projects\milkyway.cs.rpi.edu_milkyway. Interestingly, when I downloaded the most recent WU, the file astronomy_0.20_SSE.exe asked permission from my AV software to connect to the internet. Given the contents of app_info.xml it looks to me that astronomy_0.20_SSE.exe will play nice with either v0.19 or v0.20 WUs, and it looks to me like the stock app is crunching the WU, although I can't explain why the app is not listed in ...projects\milkyway.cs.rpi.edu_milkyway. Can anybody explain what's goin on?

Steve


Edit your app_info.xml file to put app 20 first and app 19 second in the list.

It has been found that the first one is loaded rather than the second even if the second is the one you want and is available.
24) Message boards : Number crunching : Milestones II (Message 32347)
Posted 13 Oct 2009 by ProfileConan
Post:
EUREKA! Have now passed 10,000,000 cobblestones on MilkyWay.
Also have passed (or soon will)1,000,000 cobblestones on 3 Projects; Collatz Conjecture plus AQUA as well Docking.

Team Cobar Spidershas also passed 10,000,000cobblestones on MilkyWay.
Plus also passed 1,000,000cobblestones on Collatz Conjecture plus AQUA as well Docking.

Team Cobar Spidershas also passed 20,000,000 cobblestones in BOINC Projects Overall Total.

Congratulations to you ICE on 50,000,000 and to TGG for 20,000,000 and all others on your milestones.
25) Message boards : Number crunching : Problem with ATI GPU crunching (Message 30055)
Posted 31 Aug 2009 by ProfileConan
Post:
I'm still suffering from the same problem. I'm really surprised the M@H team are refusing to update these drivers to support post 9.1/9.2 drivers. What are M@H going to do when XP and Vista have been replaced by Win 7? Simply give up and demand everyone start calculating on their CPU or assume that nobody is going to update their drivers because they've got nothing better to do than just run M@H?


Has nothing to do with Milkyway@home.
The drivers are produced by ATI (now owned by AMD) and not every driver version works with all card/OS combinations.
It is up to you to find the correct driver version (of which there are a number available) to suit your particular ATI card and operating system combination, not the Milkyway@home team.

So while you might be surprised at the team from Milkyway for 'refusing' to do what you think they should be doing, unfortunately they have no control over the situation.
So all they can do is produce the Application, this was adapted by some clever users (information shared with the project) so that it can work on certain ATI cards, then it is up to us users do a bit of work to get it work on our different computer setups.
26) Message boards : Number crunching : Small WUs is doing it again! (Message 29147)
Posted 10 Aug 2009 by ProfileConan
Post:
Out of work again.

If these short WUs are going to be the norm then we will need the GPU project and a different server to keep demand and supply up to scratch.
27) Message boards : Number crunching : Conflict MW (ATI) & Aqua CPU (Message 29029)
Posted 8 Aug 2009 by ProfileConan
Post:
Well both Milkyway and Aqua both run in High Priority on my machines.
They still both seem to work switching when required.

The Linux machine I have left 'as is' and it runs CPU Milkyway WUs OK as well as AQUA, Climate Prediction, Docking and QMC.

The Windows machine runs Milkyway on an ATI HD4870 with AQUA, Docking and Spinhenge on the CPUs.

I modified the cc_config.xml file on the Windows machine to show 5 CPUs (I have 4) and use 6.4.7 Boinc Client with 8.12 ATI drivers on an Win XP SP3 install.
Milkyway runs 3 jobs at once as per default settings for the card.

With AQUA using up to 4 CPUs it still leaves room for Milkyway to run on the ATI GPU. Has been working a treat so far.

The High Priority appears to have begun with the last couple of batches of AQUA work units, but for me is not causing a problem.

Conan.
28) Message boards : Number crunching : BSoD with 0xF4 stop error, catalyst 8.12 (Message 28703)
Posted 1 Aug 2009 by ProfileConan
Post:
After few mins of work WinXP stops by BSoD.
GPU: HD4870, driver Catalyst 8.12.
Any suggestions what could I try to solve this problem?
GPU works OK if MW stopped.


Not sure what could be causing this as I have the same set up and am having no problems (Win XP, Catalyst 8.12 and an HD4870 graphics card).
I run AQUA, Docking and Spinhenge (projects like Rosetta, Cosmology and a few others use too much memory and stop the card from working properly) on the CPUs with Milkyway on the GPU

What else are you running at the same time as running Milkyway?
29) Message boards : Number crunching : Updating Server Daemons Tonight (Message 28172)
Posted 22 Jul 2009 by ProfileConan
Post:
Started a Genetic Search
July 21, 2009
I'm testing a genetic search right now with the new verification. Let me know how things are working. Work should be flowing now


The above is a quote from the front page

Sorry Travis but there is NO work flowing at the moment and has not been for at least 2 hours.

Running out of work, can you help please and kick the work generator in the bum.

Thanks
Conan.
30) Message boards : Number crunching : GPU RAM and RAC. (Message 28141)
Posted 21 Jul 2009 by ProfileConan
Post:
Hi all!
Does the amount of RAM on the ATI 4870 based card
affect the RAC?
Seeing as how they are sold with 512meg and 1024meg
versions,will the cheaper lower RAM card affect the
RAC production?

Thanks, Mike.


G'Day Mike,

Not sure of the 512 MB version but my 1024 MB 4870 is getting over 60,000 cobblestones by itself (so you can see where most of my RAC is coming from).

I doubt that the 512 MB version would be too much different to this, probably depends on how many threads (WUs) you try to run at the same time, I only run 3 at a time (the default setting).
31) Message boards : Number crunching : Milestones II (Message 27996)
Posted 19 Jul 2009 by ProfileConan
Post:
Amazing, 5,000,000 cobblestones on MilkyWay, love that ATI 4870.

Also 13,000,000 cobblestones will be reached in Boinc Total as of 20/7/09.

Congratulations to all other milestone makers.
32) Message boards : Number crunching : Zero credit (Message 27825)
Posted 15 Jul 2009 by ProfileConan
Post:
This series crunches just fine for me, no errors at all, proper credits issued. No reason not to like them!!



G'Day Beau,
Yes nearly all mine also process OK and I can like them. It is an annoying few that we manage to catch (due to the almost instant database purge) that we are talking about.
As you would most likely not be able to see all your work units that get processed, you may not be aware of how many are not getting credit, just as we are not really sure how many fail as they are purged before we can see them.

But as reported a lot have been seen and we are trying to alert the project about it, seeing as they are all from the same WU series.
33) Message boards : Number crunching : ALERT MILKY DATABASE IS DOWN (Message 27568)
Posted 12 Jul 2009 by ProfileConan
Post:
ALERT !!

Milky servers are down.
Database has a problem of connection.

unable to D/L ou U/L jobs


Just uploaded some work and got another 8 WUs in return.
34) Message boards : Number crunching : Zero credit (Message 27566)
Posted 12 Jul 2009 by ProfileConan
Post:
I am also seeing them on my GPU as well.
So far all have been 'ps_sgr_210F5_s3' type WUs and all that I have seen have been on Windows machines.
I have not seen them on my Linux machines.

Agree with Banditwolf, scrap that series and move on.
35) Message boards : Number crunching : Has anyone gotten CUDA WUs ? (Message 27088)
Posted 5 Jul 2009 by ProfileConan
Post:
I have never receive CUDA WUs since I join Milkyway_GPU in May 24, 2009.

It always say "Server error: feeder not running".


And it will keep saying that. no CUDA capable application has been released yet. It is still being worked on.
There are numerous threads on Milkyway about this.
Travis is even considering not going ahead with the GPU site if the server can handle the current work for Milkyway and the GPU app.

The only GPU you can use at the moment is an ATI 38, 45, 47, 48 series cards (I think that was all of them) and use the optimised app 0.19f to get normal Milkyway apps.

If you read some of the other threads you can get more information.
36) Message boards : Number crunching : Configuring BOINC (Message 26728)
Posted 30 Jun 2009 by ProfileConan
Post:

You should really set the avg_ncpus value higher. That way a lot less WUs are loaded and also the time-out issue is no longer present. If you loose one or two tasks is not much of an issue with a Core i7 because of the hyperthreading. And if you really want to avoid it use the cc_config.xml to raise the number of CPUs the BOINC client sees. If you set it to 9 and adjust your MW resource share to 11.1% (1/9) you can more or less reserve one of the 9 CPUs for MW and can afford a lot higher avg_ncpus values (you can even set it to 1.0). At least if you reset the debts it works quite well here.

But generally speaking the scheduler of the BOINC client is quite a mess in my opinion. The features of the newer clients (6.x.xx) like fractional CPU allocation are simply not working as they should.


Where is the cc_config.xml located. I'm running windows7, have change to Boinc 6.6.20 and catalyst to 9.5 from 9.6. It is running better but still tweaking.


It's not located anywhere until you create it. You then place it in the same directory as your Application Data.

Read this thread for more detail.
37) Message boards : Number crunching : More Invalid wu's (Message 26599)
Posted 27 Jun 2009 by ProfileConan
Post:
I have only noticed a few of these faulty work units.

So far all the ones I have caught have been on my Pentium 4 Windows machine and each time the Work Unit in question has been a "ps_sgr_210F5_3s_hiw_" type.

Instead of running for 1 1/2 to 2 hours as normal they run for up to 8 hours.
Always finish and don't error out during whilst running, but I aborted one at 5 1/2 hours last night and another ran for over 8 hours after I had gone to bed, giving zero credit as classed as invalid.

Have probably had more but work units don't hang round all that long.


Just killed another one of these 210F5_3s work units at over 5 hours run time.
Interesting that the reported result shows Zero CPU time not the 5 hours I have been running.

My output on this P4 has dropped a lot recently and i now put it down to this same WU type that is taking so long to run and then gives Zero Credit for the effort, so I know I am getting more than the couple that I have been able to catch.
38) Message boards : Number crunching : More Invalid wu's (Message 26424)
Posted 25 Jun 2009 by ProfileConan
Post:
I have only noticed a few of these faulty work units.

So far all the ones I have caught have been on my Pentium 4 Windows machine and each time the Work Unit in question has been a "ps_sgr_210F5_3s_hiw_" type.

Instead of running for 1 1/2 to 2 hours as normal they run for up to 8 hours.
Always finish and don't error out during whilst running, but I aborted one at 5 1/2 hours last night and another ran for over 8 hours after I had gone to bed, giving zero credit as classed as invalid.

Have probably had more but work units don't hang round all that long.
39) Message boards : Number crunching : Hanging Work Units (Message 26007)
Posted 19 Jun 2009 by ProfileConan
Post:
Have also caught a WU that hung on my GPU

WU completed.
CPU time: 7.875 seconds, GPU time: 25267.7 seconds, wall clock time: 25561 seconds,

Mine appears to have been caused by BOINC allocating too many resources to other projects that I run and did not leave enough CPU power to run the GPU, so it did not do anything for about 7 hours.

It had started and said it was running but all that was happening was the time to completion keep ticking over.

As soon as I suspended a couple of WUs Milkyway started and completed all work it had in the queue, including the 7 hour WU (got a mighty 10 cr/h for that one).

Now I have the unfortunate circumstance of the LTD saying Milkyway has to owe other projects and now I can't get any work. I might have to micro manage BOINC for a while, running Einstein, Docking, AQUA, Ralph at the moment.

Conan.
40) Message boards : Number crunching : Milestones II (Message 25417)
Posted 14 Jun 2009 by ProfileConan
Post:
Have passed 2,500,000 cobblestones.

This is the most I have done in any single project.

Have also passed 10,000,000 cobblestones Total in all BOINC Projects.

Milkyway really helped in getting pass that milestone.


Previous 20 · Next 20

©2022 Astroinformatics Group