Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
World Community Grid Forums
Category: Beta Testing Forum: Beta Test Support Forum Thread: Computing for Clean Water 64 bit |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 81
|
Author |
|
pramo
Veteran Cruncher USA Joined: Dec 14, 2005 Post Count: 703 Status: Offline Project Badges: |
Go to the messages tab -on the top line when boinc starts you'll see one of these:
----------------------------------------1/27/2011 11:58:58 AM Starting BOINC client version 6.10.58 for windows_intelx86 2/18/2011 10:56:26 AM Starting BOINC client version 6.10.58 for windows_x86_64 ps- This might need updating? http://www.worldcommunitygrid.org/help/viewTopic.do?shortName=bnc#411 Does World Community Grid offer a 64 bit version of the BOINC software? No. The science applications are only available in 32 bit so everything we do, including validation, is offered in 32 bit. Going forward, we will re-evaluate the feasibility, cost and timing of offering a 64 bit version of the BOINC software to the member community. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Gee pramo, you're way ahead of the curve... give the techs/webmaster some space to bring this to production pretty plz, and of course, it will be a while to compile ** all and avoid performance degradation which could happen too... loads of Beta testing ahead ;P
----------------------------------------As for the performance comments... well as I understand it the first drive is that WCG was observing "pure" 64 bit devices connecting, that could not possibly have the 32 bit libs. Then, once we get it in production we can finally put that ''performance'' story to bed. Floating Points is 32 bit, hence why a 32 bit or 64 bit don't see any much of a difference when doing the Whetstone benchmark. We'll see... HCC might start truly blazing, obviating GPU computing and necessitation of a much bigger upload pipe :O Crunch On. Edit: Maybe the System Requirements page gets an extra column or an additional footnote to identify which on what platform will be on offer in 64 bits. [Edit 1 times, last edit by Former Member at Feb 24, 2011 12:50:00 PM] |
||
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2977 Status: Offline Project Badges: |
As to the actual Beta test itself, the only thing I've seen (and this may be due to my computer not having done any C4CW unit's recently), is that the initial estimate and what it's looking at coming out at, are totally off - by at least a factor of 4...
----------------------------------------Initial estimate, around 3 1/2 Hrs. So far, it's looking to complete (after ~65% done), at 13 hrs... |
||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: |
All three of my betas have completed, uploaded, and validated. Both machines are running under Ubuntu 10.04. Neither one has hyperthreading.
----------------------------------------AMD Phenom II X4 610e, 2.6 GHz BETA_ c4cw_ target03_ 061006324_ 0-- kate-amd64 Valid 2/24/11 03:33:45 2/24/11 12:31:18 4.28 108.4 / 89.6 BETA_ c4cw_ target03_ 061000474_ 0-- kate-amd64 Valid 2/24/11 03:34:05 2/24/11 12:31:18 4.22 107.0 / 89.7 Core 2 Duo P8700, 2.53 GHz BETA_ c4cw_ target03_ 061005404_ 0-- system76-pc Valid 2/24/11 03:44:22 2/24/11 12:30:53 4.14 108.6 / 89.6 I haven't run C4CW on these machines in a long time, so I don't know exactly how these times compare with the 32-bit version. I seem to recall being perplexed that the AMD took 20 mins or so longer than the System76. Here with the 64-bit, the timings are much closer (as they should be). I will see whether I can tear myself away briefly from the full court press on HFCC to time a couple of C4CW WUs. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Kate, don't know if you recently did 32 bit C4CW on your Linux box, but presuming an identical result cut (same results already run through in production), it will certainly be of interest to learn of time differentials.
Am waiting for the second run... was shut eye when this launched ;>) --//-- |
||
|
pramo
Veteran Cruncher USA Joined: Dec 14, 2005 Post Count: 703 Status: Offline Project Badges: |
Gee pramo, you're way ahead of the curve... give the techs/webmaster some space to bring this to production pretty plz, and of course, it will be a while to compile ** all and avoid performance degradation which could happen too... loads of Beta testing ahead ;P Edit: Maybe the System Requirements page gets an extra column or an additional footnote to identify which on what platform will be on offer in 64 bits. oops and sorry... i didn't mean anything by it, other than "Can't wait to see this get updated" sort of thing. also I forgot my smartaz emoticon...:) my lone beta: First person who got it blew up with the message below but my copy is running happily at 2:16:55, 54.6%, 1:52:30 to go <core_client_version>6.10.56</core_client_version> <![CDATA[ <message> process got signal 11 </message> <stderr_txt> Commandline = ../../projects/www.worldcommunitygrid.org/wcg_beta12_lmps_6.16_x86_64-pc-linux-gnu -screen none -in in.wcg.acc -var wcgsteps1 10000 -var wcgsteps2 10000 -var loop 0 -var restart 0 -var rinterval 100 -var ifile in.wcg.acc -var wcgseed 61003669 [06:33:00] Percent complete = 0.499975 |
||
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1316 Status: Offline Project Badges: |
I have the impression that a lot of these Beta's ended up as 'Inconclusive'.
----------------------------------------The single one I initial got, was out of a quorum of 2 and wasn't inconclusive, but is PV waiting for my wingman. After the initial flow I got 8 resends, all from initial quorum 1, but ending as 'Inconclusive'. Meanwhile I returned 4 and they retuned into valid for the wingman and me. Still 4 resends in progress. Because I was only crunching HFCC lately, I have no reference to the c4cw production tasks. Therefore I loaded 2 c4cw v6.15 tasks on Linux64 and Win64 each to be able to compare the runtimes. Will report later . . . |
||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: |
Kate, don't know if you recently did 32 bit C4CW on your Linux box, but presuming an identical result cut (same results already run through in production), it will certainly be of interest to learn of time differentials. --//-- Sek, unfortunately I haven't run C4CW on either of these machines for ages, and I don't seem to have records of timings from when I did. I see on "Results status" that I just snagged a regular C4CW on the AMD. When I get home from work (in about 8 hrs), I'll push it to the top; will post the comparison as soon as I have it. Kate |
||
|
pirogue
Veteran Cruncher USA Joined: Dec 8, 2008 Post Count: 685 Status: Offline Project Badges: |
These are from a Linux machine:
----------------------------------------BETA_ c4cw_ target03_ 061001062_ 0-- fed1 Valid 2/24/11 03:56:10 2/24/11 07:58:31 3.66 117.0 / 88.8 BETA_ c4cw_ target03_ 061001710_ 0-- fed1 Valid 2/24/11 03:56:10 2/24/11 07:58:53 3.66 115.7 / 88.9 BETA_ c4cw_ target03_ 061003447_ 0-- fed1 Valid 2/24/11 03:56:10 2/24/11 07:58:31 3.66 117.2 / 88.9 c4cw_ target03_ 053355047_ 0-- fed1 Valid 2/22/11 02:51:54 2/23/11 16:37:15 4.08 113.3 / 89.6 c4cw_ target03_ 053353774_ 0-- fed1 Valid 2/22/11 02:51:34 2/23/11 16:16:14 4.06 112.8 / 89.8 c4cw_ target03_ 053345414_ 0-- fed1 Valid 2/22/11 02:51:12 2/23/11 16:14:34 4.06 113.0 / 89.6 Looks like about a .5 hour (12.5%) improvement. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, that's very nice to know... integer heavy maybe... the techs could spill the beans on that token
----------------------------------------Calling Techs, Calling Techs :) And of course, now the Windows galls/guys could be in envy again :P edit: And of course a new credit variable... wonder how any differentials of 32 v 64 will be plausible ''explained'' (devils hat on ;) [Edit 1 times, last edit by Former Member at Feb 24, 2011 3:24:25 PM] |
||
|
|