LINUX Beta Test App 4.51 available
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4312
Credit: 250596344
RAC: 34436
7 Apr 2006 14:39:46 UTC
Topic 191044
(moderation:
Hi!
I put a new Linux App for testing on the Beta Test Page. It incorporates ideas from E@H usser "AkosF" and some others (including me) making it much much faster than the current official App. How fast exactly is left to you to find out.
We had some problems (i.e. segfaults) on certain machines which I expect to be caused by the graphics driver there when using SSE. If your App encounters this (segfault, catches signal 11 soon after start), please report here and follow the instructions at the bottom of the Linux setion of the page, either disabling SSE or graphics.
BM
BM
LINUX Beta Test App 4.51 available
)
Any outlook on new FreeBSD app?
RE: Any outlook on new
)
Sooner or later these changes will make it into Apps for FreeBSD and Solaris, too, but we don't have so many machines of this type here yet, so it's not the highest priority. I'll probably even have to build new gccs on that platforms to make the vector code work. May take a while.
BM
BM
RE: RE: Any outlook on
)
Okay, will wait. Feel free to drop FreeBSD 4.X if it's old compiler gives you trouble - boinc-client 5 is not supported on FreeBSD 4.X anymore anyway.
RE: boinc-client 5 is not
)
Thx for the info - this is quite useful.
BM
BM
RE: Sooner or
)
Bernd, im sorry, but the people spend their cpu-time and pay electricity for the project.
I can understand, that FreeBSD hasn't the highest priority, but "sooner or later" sounds very bad.
CU
Martin
I got my first result and it
)
I got my first result and it is valid. :)
The WUs are compareable.
Old time: ~13707 sec.
New time: ~ 7280 sec.
The speedup is ~47%.
t_old / t_new ~= 1.88
The host is an AMD Athlon 64 with SSE.
A good result, almost half the speed of the best turboalbert and hopefully a base for further development.
Well I guess we are allready spoilt by the excellent work of akos, but we are still happy to see this advancement.
cu,
Michael
RE: RE: Sooner or
)
Please, cut him some slack.
According to Boinc Synergy, there is exactly 89 FreeBSD and 155 Solaris machines working on E@H. Compared to 200,000 of Windows, or even to 15,000 Linux boxes, that's a tiny slice.
We FreeBSD users are well used to waiting up for big beasts to be served first. No problem.
Hi all, I tried albert
)
Hi all,
I tried albert 4.51 and I have a problem. The client finishes with an error code 22 (0x16). You will find all the details below. I do not know what this code 22 means. You can find the result_id concerned at this address :
http://einsteinathome.org/task/24367525
[Edit]
CPU : Intel P-4 1,7GHz
Kernel : 2.4.25-P4-SMP (quite old, I know)
X Server :
XFree86 Version 4.3.0
Release Date: 27 February 2003
X Protocol Version 11, Revision 0, Release 6.6
Build Operating System: Linux 2.4.18-24.8.0smp i686 [ELF]
Build Date: 06 March 2003
Video Driver : Radeon module delivered with the X Server
BOINC Log :
2006-04-07 18:36:13 [---] Starting BOINC client version 5.2.4 for i686-pc-linux-gnu
2006-04-07 18:36:13 [---] libcurl/7.14.0 OpenSSL/0.9.8 zlib/1.2.3
2006-04-07 18:36:13 [---] Data directory: /home/bernard/BOINC
2006-04-07 18:36:13 [Einstein@Home] Found app_info.xml; using anonymous platform
2006-04-07 18:36:13 [---] Processor: 1 GenuineIntel Intel(R) Pentium(R) 4 CPU 1.70GHz
2006-04-07 18:36:13 [---] Memory: 249.14 MB physical, 517.71 MB virtual
2006-04-07 18:36:13 [---] Disk: 11.16 GB total, 3.95 GB free
2006-04-07 18:36:13 [LHC@home] Computer ID: 103698; location: home; project prefs: default
2006-04-07 18:36:13 [Einstein@Home] Computer ID: 24864; location: home; project prefs: default
2006-04-07 18:36:13 [---] General prefs: from unknown project http://setiathome.berkeley.edu/ (last modified 2005-11-18 16:10:22)
2006-04-07 18:36:13 [---] General prefs: no separate prefs for home; using yourdefaults
2006-04-07 18:36:13 [---] Remote control not allowed; using loopback address
2006-04-07 18:36:41 [---] request_reschedule_cpus: project op
2006-04-07 18:36:46 [Einstein@Home] Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
2006-04-07 18:36:46 [Einstein@Home] Reason: To fetch work
2006-04-07 18:36:46 [Einstein@Home] Requesting 8640 seconds of new work
2006-04-07 18:36:52 [Einstein@Home] Scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi succeeded
2006-04-07 18:36:53 [Einstein@Home] Started download of skygrid_1050_z_T04.dat
2006-04-07 18:36:57 [Einstein@Home] Deferring communication with project for 54 seconds
2006-04-07 18:36:57 [Einstein@Home] Finished download of skygrid_1050_z_T04.dat
2006-04-07 18:36:57 [Einstein@Home] Throughput 358122 bytes/sec
2006-04-07 18:36:58 [---] request_reschedule_cpus: files downloaded
2006-04-07 18:36:58 [Einstein@Home] Starting result z1_1046.0__699_S4R2a_2 using albert version 451
Archive: ../../projects/einstein.phys.uwm.edu/skygrid_1050_z_T04.dat
inflating: ./skygrid_1050_z_T04.dat
2006-04-07 18:37:03 [Einstein@Home] Unrecoverable error for result z1_1046.0__699_S4R2a_2 (process exited with code 22 (0x16))
2006-04-07 18:37:03 [---] request_reschedule_cpus: process exited
2006-04-07 18:37:03 [Einstein@Home] Computation for result z1_1046.0__699_S4R2a_2 finished
2006-04-07 18:37:06 [Einstein@Home] Started upload of z1_1046.0__699_S4R2a_2_0
2006-04-07 18:37:07 [Einstein@Home] Deferring communication with project for 55 seconds
2006-04-07 18:37:09 [Einstein@Home] Finished upload of z1_1046.0__699_S4R2a_2_0
2006-04-07 18:37:09 [Einstein@Home] Throughput 584 bytes/sec
2006-04-07 18:37:43 [---] request_reschedule_cpus: project op
2006-04-07 20:20:21 [---] Exit requested by user
2006-04-07 20:20:21 [---] request_reschedule_cpus: exit_tasks
This looks like the problem
)
This looks like the problem we had with graphics&SSE. Try disabling either of these as described on the Beta Test page.
BM
BM
Bernd, Ok, I will do it. I
)
Bernd,
Ok, I will do it. I just started a WU with the 4.40 version in order to check that everything was all right. As soon as the WU is finished, I'll try each solution in order to check which one solve the problem.
I think I will able to post my results tomorrow in the evening (UTC+2).
Cu,
Bernard.