Welcome to MilkyWay@home

Updating Server Daemons Tonight

Message boards : Number crunching : Updating Server Daemons Tonight
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3

AuthorMessage
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 28223 - Posted: 22 Jul 2009, 17:46:07 UTC - in response to Message 28219.  

Gday Bandit..So I'm not going crazy then...:))

Hey Glenn, Nope it's only you. You must be going crazy. ;p


Yes it has been very slow for the last 4 or 5 hrs at least. Travis made some changes. I think at least one of them isnt working right?

Don't fix what isn't broken. Atleast the site is still up...so far...
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 28223 · 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 28224 - Posted: 22 Jul 2009, 18:28:26 UTC - in response to Message 28223.  

Perhaps DA required that Travis make MW like SETI -- as in less reliable and slow.


Gday Bandit..So I'm not going crazy then...:))

Hey Glenn, Nope it's only you. You must be going crazy. ;p


Yes it has been very slow for the last 4 or 5 hrs at least. Travis made some changes. I think at least one of them isnt working right?

Don't fix what isn't broken. Atleast the site is still up...so far...


ID: 28224 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Travis
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 30 Aug 07
Posts: 2046
Credit: 26,480
RAC: 0
Message 28226 - Posted: 22 Jul 2009, 19:19:51 UTC - in response to Message 28224.  

Perhaps DA required that Travis make MW like SETI -- as in less reliable and slow.


Gday Bandit..So I'm not going crazy then...:))

Hey Glenn, Nope it's only you. You must be going crazy. ;p


Yes it has been very slow for the last 4 or 5 hrs at least. Travis made some changes. I think at least one of them isnt working right?

Don't fix what isn't broken. Atleast the site is still up...so far...



Server seems to be running pretty fine on my end here, and the webpage seems to be loading pretty quickly. I just started some new searches, and they look like they're running ok as well.

Are you guys still having problems with the webpage?
ID: 28226 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 28229 - Posted: 22 Jul 2009, 19:24:20 UTC

Not at my end, Travis. Perhaps it's an area thing?
Go away, I was asleep


ID: 28229 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [KWSN]John Galt 007
Avatar

Send message
Joined: 12 Dec 08
Posts: 56
Credit: 269,889,439
RAC: 0
Message 28230 - Posted: 22 Jul 2009, 19:27:41 UTC

No problems here, mate...

Click to help Seti City.




ID: 28230 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Bruce
Avatar

Send message
Joined: 28 Apr 08
Posts: 1415
Credit: 2,716,428
RAC: 0
Message 28232 - Posted: 22 Jul 2009, 19:59:22 UTC - in response to Message 28226.  
Last modified: 22 Jul 2009, 20:01:31 UTC

It was affecting both my machines.Whatever it was it seems to be ok now. Pages loading quickly etc.Oh well stranger things have happened.
;-p
ID: 28232 · 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 28239 - Posted: 22 Jul 2009, 22:44:58 UTC - in response to Message 28193.  

I REALLY think that before you mess with the credit awards, you should get CUDA off the ground -- if only to have something to use as a comparison -- compare MilkyWay *released* not optimized application credit to other *released* not optimized credit schemes (say GPUGrid for example). After that, then consider the MilkyWay *released* not optimized CPU application to general CPU client performance within BOINC (say compared to Spinhenge or POEM or one of the other small but reliable BOINC projects).


We actually have a working CUDA double precision app running right now, our new researcher Anthony is working on it. I think he's going to be releasing it as a beta application in the next week or so -- the only issue is that we're having an issue with the likelihood calculation that we want to resolve, when that happens we'll have to update all the applications.

Any specific problem with the likelihood stuff? As I mentioned before, the ATI app does it on the CPU (just takes 1 or 2 seconds).

If you change anything, will the current ATI application need an update? Have you compared the results of the ATI verson with the CUDA one? I guess some differences may occur depending on the used reduction method summing up all the values.

Regarding the credits, you know my opinion that one should use exactly twice the credit multiplier of the single precision projects (i.e. 5.44 credits/TFlop). That is completely justified for CPUs (and probably future GPUs) and is still healthy for current GPUs.
ID: 28239 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Travis
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 30 Aug 07
Posts: 2046
Credit: 26,480
RAC: 0
Message 28244 - Posted: 23 Jul 2009, 0:39:41 UTC - in response to Message 28239.  
Last modified: 23 Jul 2009, 0:40:02 UTC

Any specific problem with the likelihood stuff? As I mentioned before, the ATI app does it on the CPU (just takes 1 or 2 seconds).


We're looking into it. It's more of an astronomy problem. One possible issue is that the way the likelihood calculates the integral divides the integral space linearly in mu and nu, r (the distance) is a logarithmic value, this leads the integral slices to be much larger in r near the outside of the stripe than near the close area of the stripe. We're thinking that streams can move around inside these large areas near the outside of the stripe which leads to some weird results that give higher fitness than what we think a correct answer would be. I think the astronomers could describe it better, but they've been busy working on this the last few weeks.


If you change anything, will the current ATI application need an update? Have you compared the results of the ATI verson with the CUDA one? I guess some differences may occur depending on the used reduction method summing up all the values.


Yeah, all the applications are going to have to be updated, I don't think it will be too huge of a change however (calculate_probabilities might not be changed at all, which i think is mostly what's done on the GPU). It's not a summation/GPU issue, because we're getting it on the bluegene, our clusters and everything else we run the application on.

Regarding the credits, you know my opinion that one should use exactly twice the credit multiplier of the single precision projects (i.e. 5.44 credits/TFlop). That is completely justified for CPUs (and probably future GPUs) and is still healthy for current GPUs.


I think we're using something similar to that. We're going to deal with credit reductions in the upcoming week or so.
ID: 28244 · 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 28250 - Posted: 23 Jul 2009, 3:48:47 UTC - in response to Message 28226.  

No, things are ok now -- and have been during the day today.




Server seems to be running pretty fine on my end here, and the webpage seems to be loading pretty quickly. I just started some new searches, and they look like they're running ok as well.

Are you guys still having problems with the webpage?


ID: 28250 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Glenn Rogers
Avatar

Send message
Joined: 4 Jul 08
Posts: 165
Credit: 364,966
RAC: 0
Message 28251 - Posted: 23 Jul 2009, 3:54:42 UTC - in response to Message 28223.  

I've always been a bit crazy...So far today the server/site is a lot quicker than yesterday dont know what Travis did but it seems to have solved the slooowness..:))
ID: 28251 · 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 28430 - Posted: 27 Jul 2009, 8:58:08 UTC - in response to Message 28155.  

I quit SETI because they refused to fix problems in their own credit system.

<RANT ON>
(Allowing older clients to claim 0 credits denying other hosts their rightful credit just because the result was valid. They grant the lower of the claims. They should have to fix their own %%%% problems before they can say anything about others granting too much. Bets are the average there would go up a bit if they would inforce a minimum version.)
SETI, Dave Anderson and the rest of them can bite me! I was even a donor.
<RANT OFF>


Just got around to reading the forum here again, and noticed this... Funny, I said the same thing over there and got sacked multiple times by a couple of people who continually took the cheerleader approach and stuck to the numbing tactic of saying that "everyone has the same odds of having that happen". Those same people said I'd stick around so I could post stuff over there. Wrong again. Nothing processed in almost a full year now...
ID: 28430 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Glenn Rogers
Avatar

Send message
Joined: 4 Jul 08
Posts: 165
Credit: 364,966
RAC: 0
Message 28439 - Posted: 27 Jul 2009, 12:29:41 UTC - in response to Message 28430.  

Seti is a CROCK nothing but trouble...And as far as Im concerned the powers that be over there should have a good long hard look at them selves. I have one word for them SELFISH..
ID: 28439 · 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 28449 - Posted: 27 Jul 2009, 15:06:53 UTC - in response to Message 28439.  

"Seti is a CROCK ..." "I have one word for them SELFISH.."

I haven't crunched for them in years and I still got an email wanting money early this year.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 28449 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Glenn Rogers
Avatar

Send message
Joined: 4 Jul 08
Posts: 165
Credit: 364,966
RAC: 0
Message 28450 - Posted: 27 Jul 2009, 16:39:00 UTC - in response to Message 28449.  

Well then mate what does that say to ya??
ID: 28450 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 · 2 · 3

Message boards : Number crunching : Updating Server Daemons Tonight

©2024 Astroinformatics Group