You have said a dirty combination of words, "David Anderson, and cuda".

>made by an outside source and run under the Anonymous Platform mechanism
As we're doing at the moment...
Well its made by Cluster and we're anonymous... or invisible anyway. :)

---

Cheers,

PeterV

.

Certainly Travis and Dave could get the new Milkyway_GPU project moving, with the longer WUs, by giving the source code to Cluster Physic. CP is waiting to work his magic so the ATI HD4xxx and 3xxx GPUs can move on, while the CUDA side is developed.

---

Go away, I was asleep

Ah, the heart of the matter...
Like I said, it's not about scientific progress... ;-P

Like Kevin (campaignforliberty, zeitgeistmovie, et al) said, the more machines you have, the more you are able to get because you are asking for more than others.

Actually I am planning on removing my GPU from project service today. There goes your 'he's only in it for credits theory'.

I will come back when the project is set up to actually give work to all that request it. By my reasoning, that may be sometime next year...

And like many others have said a blurb about any sort of progress would do wonders. You can speak about any other project and say how much better this one is... I don't care. I am a sci-fi nut and came here for two reasons, I like the premise of the project and it used ATI.
So when I talk about this project I am only concerned with this project not any others that I am not a member of.

See ya sometime next year probably.

---

You have said a dirty combination of words, "David Anderson, and cuda".

Well, that's what I think is going on. I don't have "proof", but it would seem to make sense considering nVidia worked with them on getting SETI's CUDA app up and running. We can debate on how much nVidia helped and whether or not there was proper testing, but like it or not, the BOINC platform is aligned with nVidia's stuff. Folding is ATI. It was probably a political / marketing thing by nVidia to try to get their foot in the door in the DC world, thus to some extent BOINC is obligated to nVidia, thus making CUDA the official GPU platform...

Again, all theory, but based on some easily observable facts and fits the general way of the business world...

Ah, the heart of the matter...
Like I said, it's not about scientific progress... ;-P

Like Kevin (campaignforliberty, zeitgeistmovie, et al) said, the more machines you have, the more you are able to get because you are asking for more than others.

Actually I am planning on removing my GPU from project service today. There goes your 'he's only in it for credits theory'.

Not really because you've already indicated that you're mad / jealous because you're not able to post the large numbers and have somewhat implied that you feel that there is some slighting going on, that certain people have some sort of unfair advantage because surely they shouldn't be getting so much while others, like yourself, get so little. This equates to taking your toys and going home. If you really like the project as much as you claim, you'd still participate in the capacity that you're able. The applications are rock solid. They aren't crashing. The only issue with this project at the moment is work availability. People over at LHC used to whine and complain about "fair distribution" of work. This is the same thing going on...

And like many others have said a blurb about any sort of progress would do wonders. You can speak about any other project and say how much better this one is... I don't care. I am a sci-fi nut and came here for two reasons, I like the premise of the project and it used ATI.
So when I talk about this project I am only concerned with this project not any others that I am not a member of.

And, as I said, I agree that an update is due. Where you and I disagree is on the level of dissatisfaction. Also, as I've mentioned, the use of ATI GPUs is not the "official BOINC way of things". CUDA is... I think we're running into politics here...

See ya sometime next year probably.

Hope so...and hope that it is quieter and less stressful then too...

Brian

I think we're running into politics here...

Sadly that is a significant factor in what drives Boinc ......

Ah, the heart of the matter...
Like I said, it's not about scientific progress... ;-P

Like Kevin (campaignforliberty, zeitgeistmovie, et al) said, the more machines you have, the more you are able to get because you are asking for more than others.

Actually I am planning on removing my GPU from project service today. There goes your 'he's only in it for credits theory'.

Not really because you've already indicated that you're mad / jealous because you're not able to post the large numbers and have somewhat implied that you feel that there is some slighting going on, that certain people have some sort of unfair advantage because surely they shouldn't be getting so much while others, like yourself, get so little. This equates to taking your toys and going home. If you really like the project as much as you claim, you'd still participate in the capacity that you're able. The applications are rock solid. They aren't crashing. The only issue with this project at the moment is work availability. People over at LHC used to whine and complain about "fair distribution" of work. This is the same thing going on...

Brian

actually I decided to remove it because work is so infrequent that it interferes with other things that I do with my GPU. If it were a steady thing I would simply turn MW off for these periods of time. But since work is so infrequent and really sparse it just did not make sense to. So yes I am taking my toy and putting it to better use elsewhere.

And again I could care less what other projects that I am not participating in are doing. That is not the subject being discussed, MW is.

---

actually I decided to remove it because work is so infrequent that it interferes with other things that I do with my GPU. If it were a steady thing I would simply turn MW off for these periods of time. But since work is so infrequent and really sparse it just did not make sense to. So yes I am taking my toy and putting it to better use elsewhere.

And again I could care less what other projects that I am not participating in are doing. That is not the subject being discussed, MW is.

Well you hit on the main issue I have with GPU coprocessing in general, and it affects both ATi and nVidia.

If all you have is your main graphics adapter to play around with, they both exhibit undesired side effects on the adapters main function, which is to put 'pictures' on my screen. There are other problems with seamlessly integrating them in as a general purpose computing resource, but those are fairly minor compared to messing up or interfering with the display.

However, since they are both relatively new in the GP computing game, I guess it's understandable why it will take some time to smooth out the rough edges, both from the system as well as application POV. ;-)

Alinator

And again I could care less what other projects that I am not participating in are doing. That is not the subject being discussed, MW is.

I am bringing up examples of the nearly the same situation that occurred in other projects. It is relevant to the discussion because it all boils down to some pretty simplistic human behaviors, not to mention the situations are nearly identical.

For example, LHC at the time was generating work very sporadically. You'd get a batch of like 70,000 tasks maybe once every 2-3 weeks (or longer), but they'd all get consumed within an hour or two at the most. People would complain because they never got a chance and felt like it wasn't fair that others got tasks to work on and they didn't and wanted the project to make it to where there was a "fair distribution".

It's not that far of a stretch to go from that situation to what we have here...

Hope it goes well for you, whatever you do. If you do like the project for the science of it, then I'd urge you to stay...but with an elevated capacity to deal with, as it was said, "disappointment issues"...

You have said a dirty combination of words, "David Anderson, and cuda".

Well, that's what I think is going on. I don't have "proof", but it would seem to make sense considering nVidia worked with them on getting SETI's CUDA app up and running. We can debate on how much nVidia helped and whether or not there was proper testing, but like it or not, the BOINC platform is aligned with nVidia's stuff. Folding is ATI. It was probably a political / marketing thing by nVidia to try to get their foot in the door in the DC world, thus to some extent BOINC is obligated to nVidia, thus making CUDA the official GPU platform...

Again, all theory, but based on some easily observable facts and fits the general way of the business world...

There were also some dark alley handshakes; but we wont get into them here...

I guess that since is see no responce neither on forum or the sites that someone is active at the moment, so i guess what I have read somewhere that Travis seems to be ill looks to be true.
So its no use to keep talking about whatever is going on untill someone from MW gives us an update how it is going or what the bottlenecks are.
To my opinion giving out much bigger units would give the project some much needed slowdown on the server.
Today was again a bad day for getting units and seems some people found a trick to suck the server dry of work.
So i guess its crossing fingers for the next day that hopefully some units trickle over to my pc again.

My RAC is dropping in high speed ! :(

To my opinion giving out much bigger units would give the project some much needed slowdown on the server.

Since tasks would be available to anyone and everyone (CPU and GPU), for that to have an impact you'd need to increase the processing time by a factor of at least 50, probably more towards 100. That kind of an increase would mostly rule out the entire Pentium4/Athlon64 and older processors if they weren't running the optimized application(s). For example, my system using the optimized app would go from taking around 1600 seconds to somewhere between 80000 and 160000 seconds, which is 22.22 and 44.44 hours respectively. Someone still using the project-supplied application, like Boosted seems to be on the CPU side of things on their X2 3800+, would go from around 3000 to 150000 - 300000, or 41.67 - 83.33 hours. Deadlines are 72 hours.

I guess that since is see no responce neither on forum or the sites that someone is active at the moment, so i guess what I have read somewhere that Travis seems to be ill looks to be true.

Dave has been the only one who seems to stop by occasionally most of this year. Travis always seems to be on something else keeping him from posting.

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

I read yesterday in a somewhat nebulous web site that ATI is actually getting away from gaming cards and wants to focus more on apps like photoshop and so on...
Or that could eventually just stop manufacturing all together..
Any truth to that?
I just cannot seem to find this site anymore.

Martin Chartrand

Unless they've merged with Adobe I doubt that.

---

me@rescam.org

Hello Misfit.
I meant any high end apps. Intensive graphic software.
I was actually looking for Overclocking my Q9550 through google and I got into an interesting forum about it and there was a link that brought me to this other URL. But now I just cannot find it
Gonna try to find it again and see what I can come up with.

Martin Chartrand

I just cant find this anymore.
But I just read that NERO is now capable of using CUDA for video processing.
Slowly getting out there.

Martin

I was actually looking for Overclocking my Q9550 through google

I do such things through some BIOS settings ;)