Well I'm back on my soapbox for a moment and then I'll get off....

From the very beginning of starting work on the MW project, there's been nothing but headaches. Granted it's only been for several months and during that time I've managed to collect quite a pile of credit for my computer efforts. I thank the project for that and it shows I'm willing to do a part to help out.

During this period, I've seen credit lowering, longer work units, compatibility problems with other projects and the list goes on.....

Everyone on this forum knows what I say is true here, yet still, I have MW scheduled to run WU's but when running ANY other project, the MW WU's stop flowing and the ONLY way to get the system to run them is to have to continually babysit the project and futz with it and the other WU's to get MW to run.

Well...I've had enough of that. I'm not going to futz with it any longer. I still have MW set to run, but guess what? I'm not suspending any other project or altering my system configuration any more in any attempt to get MW to run on MY system. I'm tired of all the BS and crap I've gone through to try to reach a "happy medium" with MW and ANY other project that I've finally come to the point where I just don't care any longer if MW runs on my system or not.

I'm currently running Aquas and Collatz WITHOUT ANY conflicts or need to change/alter/modify any of the configuration files to get these projects to run harmoniously together.

Yet, AT ANY TIME I try run run MW-WU's, I always have to either have only one project running (guess which) or, I HAVE to suspend EVERYTHING else to get MW to run, or even request new WU's.

There's something wrong with that and as far as I can see, it's within the MW project itself and no other. Why this is, I have my idea that the MW project wants EXCLUSIVE cpu time. Given the time frame to complete WU's, yeah, I think maybe it needs to be that way.

Guess what, I have other projects I'd like to run also and they aren't ALWAYS MW!

I have NO problem whatsoever running WU's for MW, but if you want exclusivity, it's not going to happen with me any longer.

I am still attached to the project as I believe the work to be done is useful. The project is not suspended. The project is allowed to download new tasks. The project has equal resource sharing with any other project I run.

The door is open yet MW fails to step through it and be a part of the bigger picture which is to share our computing power with ALL other projects equally and faithfully.

I want to help, but the project continually keeps me tied down to itself if I want to run the WU's and I just can't have that any longer, there's more than one project that needs work too.

My CPU power is here to work with you, not for you.

Yo-

Welcome to the land of BOINC, where science is done with few resources. Overall I think the project has done extremely well considering it was at the forefront of GPU crunching. The project has recently done a software upgrade and they are working through the issues slowly since BOINC comes with very little documentation.

Patience grasshopper. But please do crunch for more than one project.

Live long and BOINC.

Well I'm sorry you're having so many problems, but now that we finally have the server code updated (and i finally have access to do it myself), I'm hoping we'll be able to get things working a bit more smoothly.

This week we should have the ATI applications (if you were using those) sent out from the server, which may help some of the conflicts you're having.

If you're using CUDA applications (from the server), I'd like to know if you're having problems with those so we can fix them.

Looks like Travis logged in with the other account this time.

---

Project unfriendly.....

Only if they have an anonymous admin named "Fred".

---

me@rescam.org

Project unfriendly.....

Only if they have an anonymous admin named "Fred".

And what's wrong with "Fred" :-((

F.

---

Ah the Flintstones :-)

---

Don't drink water, that's the stuff that rusts pipes

no, right said...

---

no, right said...

And now who's too sexy for this project? *LOL*

---

Lovely greetings, Cori

... which is why there is more than one project. Just choose one that suits you.

---

Cheers,

PeterV

.

Sexy or not, but i getting wu less than 40 sec long, everything ok?

... which is why there is more than one project. Just choose one that suits you.

Just the one? I'm doing three right now. How many are you doing right now?

Oh, and yes, I'm definitely too sexy for my T shirts. Blah blah blah.

BTW I am complely stopping all crunching for BOINC. But like others I will just blah blah blah and carry on.

---

Sexy or not, but i getting wu less than 40 sec long, everything ok?

If you are still getting work, and the MW website is loading ok, don't worry about it. You just have a fast (and/or sexy) system, that's all. :)

Let us know (in a new thread, if possible) if you can't get any work at all, or if the website slows to a crawl, or if anything else seems out of place.

Thanks for the feedback!

Sexy or not, but i getting wu less than 40 sec long, everything ok?

If you are still getting work, and the MW website is loading ok, don't worry about it. You just have a fast (and/or sexy) system, that's all. :)

Let us know (in a new thread, if possible) if you can't get any work at all, or if the website slows to a crawl, or if anything else seems out of place.

Thanks for the feedback!

I hope we're not going back to the shorter WUs again, since along with the tiny WU queue that is the cause of the problem reported by the OP. It helpped a LOT when the WU time was quadrupled, but either the WUs have to be made considerably longer than they are or the queue size for GPUs has to be increased if you want BOINC to be able to manage MW properly. For more information, see this thread:

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1171

Hi Trav, no I've been running MW with ATI 4870's in this box but I have another that runs GPUGrid with nvidia 9800gt's, The Cuda app wasn't available at the time I started running MW and I got a really great deal on these ATI's I have now.

Also, please keep in mind folks, it's not the project I have a problem with, it's the way in which the Boinc manager handles the MW WU's while trying to run any other project. I still have MW set to run WU"s anytime it'll go fetch the work. The Boinc manager just won't go get the work if there's any other project running and that includes CPU only WU's from other projects.

I will try the latest release of the ATI apps to see if that helps.

Yo-

---

Hi Trav, no I've been running MW with ATI 4870's in this box but I have another that runs GPUGrid with nvidia 9800gt's, The Cuda app wasn't available at the time I started running MW and I got a really great deal on these ATI's I have now.

Also, please keep in mind folks, it's not the project I have a problem with, it's the way in which the Boinc manager handles the MW WU's while trying to run any other project. I still have MW set to run WU"s anytime it'll go fetch the work. The Boinc manager just won't go get the work if there's any other project running and that includes CPU only WU's from other projects.

I will try the latest release of the ATI apps to see if that helps.

Yo-

This week we should have the ATI apps available server side, which also might help, because this improves the communication between the BOINC client and server about information which may improve getting work and playing nice with other projects.

Weird, looks like I'm back to my old self again.

---

Weird, looks like I'm back to my old self again.

I'll agree that you're usually weird if that helps.

---

Hi Trav, no I've been running MW with ATI 4870's in this box but I have another that runs GPUGrid with nvidia 9800gt's, The Cuda app wasn't available at the time I started running MW and I got a really great deal on these ATI's I have now.

Also, please keep in mind folks, it's not the project I have a problem with, it's the way in which the Boinc manager handles the MW WU's while trying to run any other project. I still have MW set to run WU"s anytime it'll go fetch the work. The Boinc manager just won't go get the work if there's any other project running and that includes CPU only WU's from other projects.

I will try the latest release of the ATI apps to see if that helps.

Yo-

This week we should have the ATI apps available server side, which also might help, because this improves the communication between the BOINC client and server about information which may improve getting work and playing nice with other projects.

It will have no effect on the problem as the cache size is too small for BOINC to properly manage. Paul Buck and Extraterrestrial Apes have good discussions outlining why this is happening (in the thread I referenced above). In short, because of the small WU size and very small cache there is no way that BOINC manager can schedule MW properly with any other GPU projects activated. It's virtually impossible to even have a backup project running as it will hog BOINC's resource handling for the GPU since BOINC manager will always ask for more work than MW will provide under your present rules. Just had one of my GPUs run dry because of this and had to manually poll for work since your server tells it not to retry for an hour anytime there's a problem. With the WUs that you're now sending out that range anywhere from 39 seconds to 3.5 minutes on that machine and a queue limit of 12 WUs (it's an X2), it's a big problem. Why not limit the short WUs to CPU clients? That will at least help a little. Allowing GPU clients to cache more WUs would also solve the problem. Babysitting these machines constantly is a pain for those of us who are doing your work. The answer is so simple and has been expressed so many times...

Why not limit the short WUs to CPU clients? That will at least help a little. Allowing GPU clients to cache more WUs would also solve the problem.

That's what I was trying to say for a very long time, but got a bunch of static from you folks with GPUs. What I was proposing was longer tasks for GPUs only, either by segregating the tasks here in this project or with a separate project. Unfortunately, with what the project did in lengthening the runtime for all users, that helped the situation some at the expense of those on the lower end of the spectrum (re: the rise in complaints about the short deadlines).

I would think / hope that the server would be able to differentiate between a GPU and a CPU, so once all is in place, in theory the 3-stream tasks could go to those of you with GPUs and perhaps increase your caches as well, up to perhaps double what they are now. After that, the 1-stream and 2-stream tasks can go to CPU participants, again with double the cache (from 6 to 12).