Official Seti v7 binary vs. Optimized one

Message boards : Number crunching : Official Seti v7 binary vs. Optimized one
Message board moderation

To post messages, you must log in.

AuthorMessage
Arivald Ha'gel

Send message
Joined: 9 May 03
Posts: 14
Credit: 16,623,619
RAC: 2
Poland
Message 1671991 - Posted: 30 Apr 2015, 12:34:37 UTC

Hello,

As I understand official binary is in fact an old optimized binary.

It can be seen in output as:

Windows optimized S@H Enhanced application by Alex Kan
Version info: SSEx (AMD/Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan
SSEx Win32 Build 1831 , Ported by : Raistmer, JDWhale
SETI7 update by Raistmer
OpenCL version by Raistmer, r1831


I can see for some other PCs that they use newer/newest Optimized binary:

Windows optimized S@H v7 application
Based on Intel, Core 2-optimized v8-nographics V5.13 by Alex Kan
SSSE3xj Win32 Build 2760 , Ported by : Raistmer, JDWhale
SETI7 update by Raistmer
OpenCL version by Raistmer, r2760


Which seems to be considerably faster (>20% faster) - however I do not take into account oclFFT parameter optimizations.

Is there any plan to review changes between version r1831 and r2760 and use newer optimized binary as official one?

IMHO 20% is quite a lot, and increase of computational power by 20% when so many PCs work on SETI seems to be an investment with huge benefits.
ID: 1671991 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1671995 - Posted: 30 Apr 2015, 12:44:07 UTC
Last modified: 30 Apr 2015, 12:47:40 UTC

Yes, thats true.

You can always download more recent binaries from my website or running the Lunatics Installer.
You are wasting computing cycles indeed.

http://setiathome.berkeley.edu/forum_thread.php?id=71867

http://mikesworldnet.de/home


With each crime and every kindness we birth our future.
ID: 1671995 · Report as offensive
Arivald Ha'gel

Send message
Joined: 9 May 03
Posts: 14
Credit: 16,623,619
RAC: 2
Poland
Message 1672513 - Posted: 1 May 2015, 9:11:00 UTC
Last modified: 1 May 2015, 9:18:45 UTC

My question is: Am I?
Is this me who's wasting computing cycles?

As I already explained it's up to SETI@Home personnel to review changes and take newer binary as official if they accept those changes.

As I can see Official Astropulse app is one of the newest:
AstroPulse v7 Windows x86 rev 2742, V7 match, by Raistmer with support of Lunatics.kwsn.net team. SSE2
OpenCL version by Raistmer
oclFFT fix for ATI GPUs by Urs Echternacht


Thus I believe it should be possible to review SETI app without that much effort.
ID: 1672513 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 1672532 - Posted: 1 May 2015, 11:19:34 UTC - in response to Message 1672513.  

My question is: Am I?
Is this me who's wasting computing cycles?

As I already explained it's up to SETI@Home personnel to review changes and take newer binary as official if they accept those changes.

As I can see Official Astropulse app is one of the newest:
AstroPulse v7 Windows x86 rev 2742, V7 match, by Raistmer with support of Lunatics.kwsn.net team. SSE2
OpenCL version by Raistmer
oclFFT fix for ATI GPUs by Urs Echternacht


Thus I believe it should be possible to review SETI app without that much effort.


Well, to get new stock is quite lengthy process indeed.
There are many factors. And undermanned nture of Berkeley's crew is only one of them.
We need to be sure that new version works correctly, that it works on almost if not all configs, that if doesn't make PC unusabe... and that all this happens unattended regarding PC's user. Hence inevitable long testing phase when some bigger than trivial changes are done. Usual time scale is month(s).

But if there is some already achieved speed improvement and app looks stable enough I try to make it publicly available soon after alphs testing of binary. That way enthusiasts can manually download and install it and act as testers also that allows to spot any issues even before "official" beta. That's how it works for last years. Currently OpenCL MultiBeam under additional development so I refrain from starting beta until more improvements will be made. But best what we have currently available to everyone who want it.
ID: 1672532 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1672543 - Posted: 1 May 2015, 12:24:26 UTC
Last modified: 1 May 2015, 12:25:55 UTC

As I already explained it's up to SETI@Home personnel to review changes and take newer binary as official if they accept those changes.


Sorry, but no.
Seti provides the binaries so you can calculate data.
If you want to do it more efficient its up to you.
The only thing seti is responsible for in my point of view is that the binaries produce valid results and those binaries are running on most configs.

Here at the forums are plenty of people (me included) helping to install optimised apps.


With each crime and every kindness we birth our future.
ID: 1672543 · Report as offensive
Arivald Ha'gel

Send message
Joined: 9 May 03
Posts: 14
Credit: 16,623,619
RAC: 2
Poland
Message 1673740 - Posted: 4 May 2015, 7:05:34 UTC - in response to Message 1672543.  
Last modified: 4 May 2015, 7:07:00 UTC

As I already explained it's up to SETI@Home personnel to review changes and take newer binary as official if they accept those changes.


If you want to do it more efficient its up to you.
The only thing seti is responsible for in my point of view is that the binaries produce valid results and those binaries are running on most configs.


It's like you'd be saying that they do not want to speed up the process.
It is also up to them to use better app as an official one. And I'm sure they'd agree.

Since the only person/organization/etc. that can review and take changes into official SETI binary is the SETI personnel, they are responsible for official binaries, and their efficiency. I'm more likely to agree that their resources (time, personnel, etc) is so constrained that it's not that often that a new binary is taken.

I understand that.

However we do see that it was already done at least once. I think that after more than 1000 revisions for Optimized Binary it might be a good idea to try to persuade SETI team to test and review Optimized Binary again.
Even more so since I can see that quite recent Astropulse Optimized Binary is taken as an Official one.
ID: 1673740 · Report as offensive

Message boards : Number crunching : Official Seti v7 binary vs. Optimized one


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.