| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Beta Testing
Forum: Beta Test Support Forum
Thread: HCC1 BETA Comment thread |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 95
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I am seeing about a 30-40% decrease in run time on the AMD quad and about a 50% improvement in the Intels. The Celeron it appears may significantly benefit from the new algorithm - about 60% faster than the shortest of the HFCC WUs it has crunched.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Uplinger,
----------------------------------------You will have to formulate another newsletter shortly, so briefly after the one received last night. By esoteric17 calculus 63.6% faster, or by Italian ottimisimo calculation 2.75 times faster. Might win a few members back into active contribution, in the know that jobs can be completed in a matter of a few internet browse hours.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Mar 24, 2010 3:47:45 PM] |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: 
|
Good point Sek, I think many people use their computer infrequently, and dont like the idea of leaving it on too long; even 6h tasks are too much for some.
----------------------------------------If these people could still contribute, it would be beneficial to the project overall, so it is worth doing. This group would still need to have a reasonable computer (an Atom wont finish quickly, but perhaps 7h tasks are as low as you can go to include that group). Of course this would mess up the tasks completed boards; an i7 that could finish a task in 90min could bring 128back per day, about the same as I presently do with many systems. Perhaps the techs could have small, med and large task groups and make 3 small=1large, or something like that. I do hope they dont just double the length of the tasks! [Edit 1 times, last edit by skgiven at Apr 23, 2010 10:41:28 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Various options are being explored... silently a a simplification on the who gets what has been implemented and substantial improvements are being observed by knreed... I'm seeing a very high rate of validation on quorum 2 dependent work even though my caches are 1.0 and 0.8 days on quad and duo. Dynamic sizing to power is high on the list of things to achieve. From past reading, the main worry is the WU's per day for the overall system... Monday we had 514,000 validating, a WCG record and nobody really noticed. Maybe large numbers of DDDT-2 Type C with 20mb a pop and just 45 minutes run time could put a strain on. The main concern is the flow. Think there is 1 DC grid that has very major issue with handling returned work... now what's the point of sending out work if you're not in a position to take in the results?
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges:
|
Selection might explain why I picked up one such task from about a dozen active systems.
----------------------------------------[Edit 1 times, last edit by skgiven at Apr 23, 2010 10:40:07 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Yes the uploads are smallish, it's download sizes I'm talking about, the various project DL/UL sizes printed in the System Req. Matrix
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
With this beta I got my 1 year badge. I unchecked beta on my account since I reached my goal and I want others to have a chance at a badge.
 |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges:
|
brinktastee, congratulations on reaching 1year for Betas
---------------------------------------- Not many in that party; you, uplinger...? System Requirements for Projects, and Work Unit Upload and Download sizes [Edit 1 times, last edit by skgiven at Apr 23, 2010 10:37:50 PM] |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
I think the comments on upload is a combination between DDDT2 and HCC1 in this conversation.
For DDDT2 we are seeing uploads of around 150KBytes (since it's compressed if you're using an agent from 5.4.x i believe). This was for type A and B work units, I don't have the numbers for type C. For HCC1 the upload is around 25KBytes Compressed. We give larger numbers usually for download and upload to give members a general worst case scenario. Back on topic though. The HCC1 beta has gone really well. We are collaborating with the researchers from the results we have received to see if this version is good for release. I'll keep everyone posted :) Hope this helps, And thank you for the help with BETA! -Uplinger |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Se reeson I'm keeping the much more expanded Start Here forum SR Matrix is that A/B/C actually do not produce same size results. An A type good for 24 hrs plus, would be considered light in bandwidth [though not as light as FAAH/HFCC]. At any rate, I'm not seeing 1MB uploads, rather 0.1-0.2 and that's pre-compressing. Mind you I've not seen very many to be able to report the total ranges... that's where members can help.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|