Welcome to MilkyWay@home

MilkyWay_GPU - Almost There!

Message boards : Number crunching : MilkyWay_GPU - Almost There!
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 4 · 5 · 6 · 7

AuthorMessage
Profile Bymark
Avatar

Send message
Joined: 6 Mar 09
Posts: 51
Credit: 492,109,133
RAC: 0
Message 22896 - Posted: 21 May 2009, 14:11:54 UTC - in response to Message 22892.  
Last modified: 21 May 2009, 14:15:22 UTC

My personal internet access is down for a few days so I didn't tried, so I ask the question(s) directly : is the CUDA project operationnal or not ? Has anyone tried with a small graphic card like 8600M GS, 8600 GT, etc. ?


I tried on a 8600GT.
"5/21/2009 6:31:51 AM Milkyway@home Message from server: No work available"


Confused, should not the right error message be:

21.5.2009 16:29:12 Milkyway@Home_for_GPU Message from server: Server error: feeder not running

Or where is the first cuda wu appearing?
ID: 22896 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Cluster Physik

Send message
Joined: 26 Jul 08
Posts: 627
Credit: 94,940,203
RAC: 0
Message 22897 - Posted: 21 May 2009, 14:23:39 UTC - in response to Message 22874.  

It's good to hear that modifying the CUDA app for use by an ATI card is not as big a deal as I thought it was.

At least I hope so. Don't know yet what Travis has exactly done to the app. But based on the previous versions I assume it is nothing complicating the porting.
ID: 22897 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 22900 - Posted: 21 May 2009, 14:34:29 UTC - in response to Message 22896.  

My personal internet access is down for a few days so I didn't tried, so I ask the question(s) directly : is the CUDA project operationnal or not ? Has anyone tried with a small graphic card like 8600M GS, 8600 GT, etc. ?


I tried on a 8600GT.
"5/21/2009 6:31:51 AM Milkyway@home Message from server: No work available"


Confused, should not the right error message be:

21.5.2009 16:29:12 Milkyway@Home_for_GPU Message from server: Server error: feeder not running

Or where is the first cuda wu appearing?


As far as I know the GPU site isn't 'working' yet. The server hasn't been turned on yet, and the app still hasn't been released.

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 22900 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
BarryAZ

Send message
Joined: 1 Sep 08
Posts: 520
Credit: 302,524,931
RAC: 15
Message 22909 - Posted: 21 May 2009, 16:44:07 UTC - in response to Message 22817.  

Perhaps some folks here haven't been paying attention. The deal with ATI GPU support -- it currently has to be an 'offline' effort -- 3rd party folks. The reason for this is that the Native BOINC client (for various reasons which some might characterize in less charitable terms) is focused only on providing CUDA support. The only reason there is ATI GPU support here is thru the dedicated efforts of the folks generating the optimized clients here. (The only reason they were able to do that is that the MilkyWay project folks were forward looking enough to make their application open source for outside development.

What the project folks have been working on *within the constraints* of the BOINC client control folks, is a CUDA GPU application, elsewise they would not only incur the Wrath of DA, but also be spending a lot of time figuring out a lot of new code. Instead, the plan is to release the CUDA GPU client from the project, and then open it up as they have done with the CPU client. Once that is done, the same angels who have optimized the CPU client (and end-run supported the ATI GPU piece) will likely optimize the CUDA application AND develop an optimized ATI GPU client for the Milky_GPU project.

The folks you need to lobby largely or the DA Consortium which has been influenced to support CUDA *to the exclusion* of ATI. Might be the same reason that many standard BOINC projects appear more efficient with Intel CPU's -- not sure about that.




wow...
Your the only project that runs ATI and instead you come out with a CUDA GPU app instead of finishing the ATI WU app...
Sorry if this sounds harsh but it is just the way I feel.

I feel the same way! Focus should have been on ATI!
Longer workunits to start with... more workunit availability... enough to get done there first....


ID: 22909 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jord
Avatar

Send message
Joined: 30 Aug 07
Posts: 125
Credit: 207,206
RAC: 0
Message 22911 - Posted: 21 May 2009, 17:16:02 UTC - in response to Message 22909.  

The folks you need to lobby largely or the DA Consortium which has been influenced to support CUDA *to the exclusion* of ATI.

God, talk about conspiracy theories. The faked moon landings, the giant face on Mars, the start of the war on terrorism. They're all fault of the DA Consortium! Wouldn't the world be a better place without this beast?
Jord.

The BOINC FAQ Service.
ID: 22911 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Crunch3r
Volunteer developer
Avatar

Send message
Joined: 17 Feb 08
Posts: 363
Credit: 258,227,990
RAC: 0
Message 22914 - Posted: 21 May 2009, 17:45:07 UTC - in response to Message 22911.  
Last modified: 21 May 2009, 17:45:42 UTC

The folks you need to lobby largely or the DA Consortium which has been influenced to support CUDA *to the exclusion* of ATI.

God, talk about conspiracy theories. The faked moon landings, the giant face on Mars, the start of the war on terrorism. They're all fault of the DA Consortium! Wouldn't the world be a better place without this beast?


OMG, just imagine a boinc world without DA...

Devs could focus on the important things like fixing the mess that he made of boinc... no stupid eye candy/social networking on the web pages since that crap is of no importance at all and would get last place on the todo list where it belongs ... and the list goes on and on...

It would be heaven on earth! :)

Join Support science! Joinc Team BOINC United now!
ID: 22914 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>DoJ] supersonic

Send message
Joined: 5 Mar 09
Posts: 19
Credit: 102,651,985
RAC: 0
Message 22917 - Posted: 21 May 2009, 17:57:08 UTC - in response to Message 22873.  

much respect to your involvement, CP.

and may the code be cristal clear ;)
ID: 22917 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jord
Avatar

Send message
Joined: 30 Aug 07
Posts: 125
Credit: 207,206
RAC: 0
Message 22919 - Posted: 21 May 2009, 18:21:12 UTC - in response to Message 22914.  
Last modified: 21 May 2009, 18:43:15 UTC

Devs could focus on the important things like fixing the mess that he made of boinc...

The little thing people forget is this: Without David there would have been no Seti@Home and no BOINC, so also no project developers all huddled together under one umbrella and nothing for the participants to worry, conspire or theorize about. You'd be installing 50 different programs to do your 50 different programs, all working against each other.
Jord.

The BOINC FAQ Service.
ID: 22919 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Paul D. Buck

Send message
Joined: 12 Apr 08
Posts: 621
Credit: 161,934,067
RAC: 0
Message 22920 - Posted: 21 May 2009, 19:07:56 UTC - in response to Message 22919.  

Devs could focus on the important things like fixing the mess that he made of boinc...

Without him there would have been no Seti@Home and no BOINC, so also no project developers all huddled together under one umbrella.

Um, not exactly. Just as Bell was not the only guy that invented the telephone, DA is not the only one that invented distributed computing. Yes he made a successful tool and was an advocate for it, but, I recall almost 10 DC projects underway at the time I started doing SaH Classic ... had he not been the first on the street I think that some other project would have become the big cahuna ...

As to BOINC, in the sense that it is a generic client, that too would have come about... would it have been later, earlier, better, worse, no one can debate the alternative history because we are stuck with the history we have ... but accidents of timing make some look a lot smarter and more capable than they are ...

As to project developers huddled together under one umbrella? Who? We talk of them as though there are a lot ... there are three ... and none of them are all that active in adding changes on a daily basis ... or in correcting Trac tickets ... or any other of a number of things ...

The reason people get the impression that he acts like a dictator is that, well, he does ... being the founder and having a couple of good ideas does not make one infallible. Nor omniscient. Being in charge as a dictator is fun and it is very good at inflating one's ego ... but it is a bad way to conduct business. I could point to a couple historical parallels where events went along on the side of some dictators as they called the shots ... but there is an Internet law about that that everyone can figure out to whom I refer, oops wait, Stalin isn't in that law...

I was told quite recently that one of the reasons the developers do not like my submissions is that I don't compromise ... well, the first problem with that is that to have a compromise you first have to have a discussion ... and to have a compromise, both parties have to give some ... yet history shows that in exchanges with DA it is his way or no way ... his final "compromise" offer on making the UBW the official wiki was that I should discard the totality of 4 years effort and start from scratch ... oh, and never, and I mean never, mention anything specific about a project ... from the participant's perspective, how can you document BOINC and never discuss projects ... how CPDN's memory load may cause issues with Rosetta, how Discovering Drugs may cause tasks to fail with no heartbeat errors ... and so on and so on and so forth ...

I know of developers that have written code and submitted it, to have it ignored, or gutted, or altered so that it no longer works ...

Having one good idea does not make one into a god ... or make all your other ideas better than everyone else's ... teams are stronger because they build on the strengths of all its members ... yes you need leadership, and goals, and direction ... but real leaders cultivate input even from people they don't like and / or don't agree with ... because real leaders know that no one has the corner on ideas, or being right ...
ID: 22920 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jord
Avatar

Send message
Joined: 30 Aug 07
Posts: 125
Credit: 207,206
RAC: 0
Message 22922 - Posted: 21 May 2009, 19:28:27 UTC - in response to Message 22920.  
Last modified: 21 May 2009, 19:56:31 UTC

Um, not exactly. Just as Bell was not the only guy that invented the telephone, DA is not the only one that invented distributed computing. Yes he made a successful tool and was an advocate for it, but, I recall almost 10 DC projects underway at the time I started doing SaH Classic ... had he not been the first on the street I think that some other project would have become the big cahuna ...

Look, I wasn't talking about any of the other distributed computing projects that were out there, some may still be. You read too much into it. I said that without David there would have been no Seti@Home (in the form it was available to all of us) and there certainly would be no BOINC. What would've otherwise happened is one for the parallel universe theorists.

As to project developers huddled together under one umbrella?

Not BOINC (code) developers, but project developers. People like Travis Deschell, Eric Myers, Eric Korpela, Jack Shultz, Nicolás Alvarez, Rytis Slatkevicius. The people who came up with the idea of the project and programmed it into usable applications.
Jord.

The BOINC FAQ Service.
ID: 22922 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 21 Aug 08
Posts: 625
Credit: 558,425
RAC: 0
Message 22925 - Posted: 21 May 2009, 19:38:54 UTC - in response to Message 22920.  


Having one good idea does not make one into a god ... or make all your other ideas better than everyone else's ... teams are stronger because they build on the strengths of all its members ... yes you need leadership, and goals, and direction ... but real leaders cultivate input even from people they don't like and / or don't agree with ... because real leaders know that no one has the corner on ideas, or being right ...


ID: 22925 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
BarryAZ

Send message
Joined: 1 Sep 08
Posts: 520
Credit: 302,524,931
RAC: 15
Message 22943 - Posted: 21 May 2009, 21:00:09 UTC - in response to Message 22919.  
Last modified: 21 May 2009, 21:02:42 UTC

Sure -- credit due indeed (though not too much credit as it might be over the DA standard issue). The incandescent light bulb was way cool in its days and all credit to Edison.

The thing is, as Paul has described, there is a time when strict single person control can inhibit progress and stifle creativity. I think that is what has been happening with BOINC over the past couple of years.

I also replied yesterday in this thread to redirect someone's apparent angst directed at the few folks in the Milkyway project for not developing their first project based Milkyway_GPU application for ATI cards. The point being that was a case of being misdirected. The folks here essentially HAVE NO CHOICE but to develop, within the constraints of the core NO-ATI BOINC client constraints. Otherwise they might incur DA sanctions. Besides, I doubt the folks doing the project application design have that sort of expertise -- doing an application which works directly with the BOINC client is new territory enough.

Fortunately, there are 3rd party folks (outside of the project and thus less constrained by the DA control scenarios) who are quite skilled at optimizing the released Milkyway client. Once the MilkyWay_GPU client (CUDA) is public, that good souls will no doubt work on not only optimizing the CUDA client, but developing a solid ATI GPU optimized application which will work with the MilkyWay GPU project.

Folks pressing for ATI support here already have advocates and 3rd party developers. Now if they want to make some serious noise, they should lobby or ATI support over at GPUGrid, or if they REALLY want to engage in a love fest, lobby for ATI GPU support over in SETI.




The little thing people forget is this: Without David there would have been no Seti@Home and no BOINC, so also no project developers all huddled together under one umbrella and nothing for the participants to worry, conspire or theorize about. You'd be installing 50 different programs to do your 50 different programs, all working against each other.

ID: 22943 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Paul D. Buck

Send message
Joined: 12 Apr 08
Posts: 621
Credit: 161,934,067
RAC: 0
Message 22947 - Posted: 21 May 2009, 21:40:52 UTC - in response to Message 22922.  

Um, not exactly. Just as Bell was not the only guy that invented the telephone, DA is not the only one that invented distributed computing. Yes he made a successful tool and was an advocate for it, but, I recall almost 10 DC projects underway at the time I started doing SaH Classic ... had he not been the first on the street I think that some other project would have become the big cahuna ...

Look, I wasn't talking about any of the other distributed computing projects that were out there, some may still be. You read too much into it. I said that without David there would have been no Seti@Home (in the form it was available to all of us) and there certainly would be no BOINC. What would've otherwise happened is one for the parallel universe theorists.

And my point is that SaH is only one of many projects, and probably one of the least deserving of support of them all. I was pretty rabid in the early days because it was one of the few projects that would just run. I tried CPDN Classic but the models took too long and crashed too often. I was simply pointing out that had not he been there the world right now would not be all that different than it is today. Just different names on things and different artifacts. Again, the point being that DA is not all that unique or all that special. Had he not been here, well, so what?

Some of those early projects are still about, some have morphed. IBM changed from UD to BOINC cause it was cheaper. Had BOINC not been around do you think they might not have been able to get something built to replace UD?

As to project developers huddled together under one umbrella?

Not BOINC (code) developers, but project developers. People like Travis Deschell, Eric Myers, Eric Korpela, Jack Shultz, Nicolás Alvarez, Rytis Slatkevicius. The people who came up with the idea of the project and programmed it into usable applications.

Yes, and I can name some more ... and even those on your list aren't all that active ... even adding those to the list of developers in a universe that at one point had 1.6M people pass through, where are they all now? Surely some of them were developers ... When was the last time anyone of those individuals made a substantial contribution to the code base that actually reflects in the user experience ... not some hidden project enhancing feature ... and if Nicolás Alvarez is so happy with the development model in BOINC why does he have his own forked baseline?
ID: 22947 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 . . . 4 · 5 · 6 · 7

Message boards : Number crunching : MilkyWay_GPU - Almost There!

©2024 Astroinformatics Group