Welcome to MilkyWay@home

Separation updated to 0.80

Message boards : News : Separation updated to 0.80
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · Next

AuthorMessage
Len LE/GE

Send message
Joined: 8 Feb 08
Posts: 261
Credit: 104,050,322
RAC: 0
Message 49229 - Posted: 13 Jun 2011, 2:14:55 UTC - in response to Message 49226.  

The initial sleep time is always 0. Hypothetically it should slightly reduce the CPU usage, but if I tried anything more than 0 it always ended up noticeably slower. The default is still to do 1ms polling. It shouldn't be doing busy waiting (unless you tell it to); It isn't for me, and I just checked out your tasks and the reported used CPU time isn't consistent with busy waiting. There still is a burst of 100% CPU at the end about 2 seconds long, so you might have seen it then.


The busy wait shows as system time used, not showing up at process cpu time; so you will not find it in the log but see it in task manager.
For the logs I checked, the estimated time is ~5% less than the average time per iteration. That looks like a good value. Only the last 5% of an iteration the cpu would actively wait for a reply from the gpu.

Did you see my last comment about the command line param leading to errors?
I edited it in while you replied.
ID: 49229 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49230 - Posted: 13 Jun 2011, 2:43:16 UTC - in response to Message 49227.  

When I put something in the every WUs directly ends up with an error ::

<core_client_version>6.12.28</core_client_version>
<![CDATA[
<message>
Fonction incorrecte. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4' 
Error reading astronomy parameters from file 'astronomy_parameters.txt'
  Trying old parameters file
Number of parameters doesn't make sense
<search_application> milkywayathome_client separation 0.80 Windows x86_64 double CAL++ </search_application>
03:47:06 (4040): called boinc_finish

</stderr_txt>
]]>
Yep. That looks broken. I suppose I should get around to actually implementing preferences correctly... I really wish BOINC would prepend arguments where they belong...
ID: 49230 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49231 - Posted: 13 Jun 2011, 2:44:22 UTC - in response to Message 49229.  

The busy wait shows as system time used, not showing up at process cpu time; so you will not find it in the log but see it in task manager.
For the logs I checked, the estimated time is ~5% less than the average time per iteration. That looks like a good value. Only the last 5% of an iteration the cpu would actively wait for a reply from the gpu.

The accuracy varies greatly between different GPUs it seems
ID: 49231 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49232 - Posted: 13 Jun 2011, 2:45:17 UTC - in response to Message 49228.  

I'm running a Mac mini core duo (NOT core 2 duo) with OS X Tiger (10.4.11). Everything errors out immediately with the following stderr text:

6.10.56

process got signal 5


dyld: Symbol not found: ___stack_chk_guard
Referenced from: /Library/Application Support/BOINC Data/slots/4/../../projects/milkyway.cs.rpi.edu_milkyway/milkyway_separation_0.80_i686-apple-darwin
Expected in: /usr/lib/libSystem.B.dylib

I tried both resetting and detaching - no luck. Thanks for your work on the 32 bit update, I hope a minor tweak will fix everything. I would hate to go back to the version that I had to run under Rosetta - it took a looong time to complete a unit that way.
That's probably because I think I built it against the Leopard/10.5 SDK. I'll try again after I install the 10.4 SDK.
ID: 49232 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
paris
Avatar

Send message
Joined: 26 Apr 08
Posts: 87
Credit: 64,801,496
RAC: 0
Message 49233 - Posted: 13 Jun 2011, 2:54:34 UTC - in response to Message 49232.  

That's probably because I think I built it against the Leopard/10.5 SDK. I'll try again after I install the 10.4 SDK.


Wow, thanks for the quick reply. You are probably right - I read something a while ago about needing to use the 10.4 SDK because the 10.5 did not compile properly for the Tiger applications. No rush. I'll be here when it is ready to go.


Plus SETI Classic = 21,082 WUs
ID: 49233 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49234 - Posted: 13 Jun 2011, 3:13:28 UTC - in response to Message 49206.  

I'm also seeing an issue on Linux with the GPU applications getting permission denied when trying to set the process priority, even though it works fine outside of the manager.

I'll try to get everything that's broken fixed by tomorrow...
ID: 49234 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
hoarfrost

Send message
Joined: 4 Sep 09
Posts: 20
Credit: 187,688,252
RAC: 0
Message 49235 - Posted: 13 Jun 2011, 5:05:44 UTC

Hello!

With app_info.xml placed below, my computer became a computation errors generator:
<app_info>
<app>
<name>milkyway</name>
</app>
<file_info>
<name>milkyway_separation_0.80_windows_intelx86__ati14.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>milkyway</app_name>
<version_num>80</version_num>
<plan_class>ati14</plan_class>
<flops>1.0e11</flops>
<avg_ncpus>0.05</avg_ncpus>
<max_ncpus>1</max_ncpus>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<cmdline>--gpu-target-frequency 120</cmdline>
<file_ref>
<file_name>milkyway_separation_0.80_windows_intelx86__ati14.exe</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>


When I remove <cmdline> tags computations run normally, but GUI works slow...
ID: 49235 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Simplex0
Avatar

Send message
Joined: 11 Nov 07
Posts: 232
Credit: 178,229,009
RAC: 0
Message 49236 - Posted: 13 Jun 2011, 5:46:47 UTC

Seams to work fine now, only 2 marked as 'invalid' on HD 5870.
Using Vista 64 and HD4870, HD5870 & HD6970.
ID: 49236 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
[boinc.at] Nowi

Send message
Joined: 22 Mar 09
Posts: 99
Credit: 503,422,495
RAC: 0
Message 49237 - Posted: 13 Jun 2011, 6:05:31 UTC

No problems so far with the 0.80 on two 5850 on crossfire. (Win7-64, Boinc 6.12.28)

Calculation-time seems to be 3 to 4 seconds shorter than with 0.62. For me it looks like that depends on the final cpu-calculation time, which is now closer.

Thank you, Matt!
ID: 49237 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>Occitania] Meteore31

Send message
Joined: 11 Dec 07
Posts: 11
Credit: 95,037,645
RAC: 0
Message 49238 - Posted: 13 Jun 2011, 8:08:35 UTC
Last modified: 13 Jun 2011, 8:16:49 UTC

How can we use the flag --process-priority (-b) ?
I tried to add it on my app_info like this :
<cmdline>--process-priority=0</cmdline>
or this <cmdline>--process-priority (-0)</cmdline>
or this <cmdline>-b0</cmdline>

Is it correct ? Apparently not because wu goes in error immediately

If I remove this flag, it's ok but I only have 1 wu, only 1 by 1, in priority high
ID: 49238 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile BladeD
Avatar

Send message
Joined: 2 Nov 10
Posts: 731
Credit: 131,536,342
RAC: 0
Message 49239 - Posted: 13 Jun 2011, 8:21:31 UTC
Last modified: 13 Jun 2011, 8:24:52 UTC

Finished WUs, but I'm not getting more now! No error messages! How can it say not requiring tasks, when I have no WUs for two 5870s?

6/13/2011 4:22:24 AM Milkyway@home Sending scheduler request: Requested by user.
6/13/2011 4:22:24 AM Milkyway@home Not reporting or requesting tasks
6/13/2011 4:22:26 AM Milkyway@home Scheduler request completed
ID: 49239 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Chris S
Avatar

Send message
Joined: 20 Sep 08
Posts: 1391
Credit: 203,563,566
RAC: 0
Message 49241 - Posted: 13 Jun 2011, 8:50:42 UTC

o.8 workunits validating OK on HD4770, Win 7 64 bit Ultimate, BM 6.12.28.

However there is an error reported in the output file

Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
ID: 49241 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49242 - Posted: 13 Jun 2011, 9:28:48 UTC - in response to Message 49241.  

o.8 workunits validating OK on HD4770, Win 7 64 bit Ultimate, BM 6.12.28.

However there is an error reported in the output file

Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
That's fine. That's fine. It will go away when the new parameters file is actually used.
ID: 49242 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Matt Arsenault
Volunteer moderator
Project developer
Project tester
Project scientist

Send message
Joined: 8 May 10
Posts: 576
Credit: 15,979,383
RAC: 0
Message 49243 - Posted: 13 Jun 2011, 9:29:44 UTC - in response to Message 49238.  

How can we use the flag --process-priority (-b) ?
I tried to add it on my app_info like this :
--process-priority=0
or this --process-priority (-0)
or this -b0

Is it correct ? Apparently not because wu goes in error immediately

If I remove this flag, it's ok but I only have 1 wu, only 1 by 1, in priority high
Any arguments are broken. I'll have the fix release later today
ID: 49243 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
ConflictingEmotions

Send message
Joined: 18 Feb 09
Posts: 9
Credit: 20,005,162
RAC: 0
Message 49247 - Posted: 13 Jun 2011, 13:17:10 UTC - in response to Message 49234.  

I'm also seeing an issue on Linux with the GPU applications getting permission denied when trying to set the process priority, even though it works fine outside of the manager.

I'll try to get everything that's broken fixed by tomorrow...

I am getting this message for all my Linux AMD WUs (not Windows) so I have stopped requesting work on my Linux host.

Stderr output

<core_client_version>6.10.58</core_client_version>
<![CDATA[
<message>
process exited with code 22 (0x16, -234)
</message>
<stderr_txt>
execv: No such file or directory

</stderr_txt>
]]>


ID: 49247 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
paris
Avatar

Send message
Joined: 26 Apr 08
Posts: 87
Credit: 64,801,496
RAC: 0
Message 49258 - Posted: 13 Jun 2011, 14:32:51 UTC - in response to Message 49233.  

That's probably because I think I built it against the Leopard/10.5 SDK. I'll try again after I install the 10.4 SDK.


Wow, thanks for the quick reply. You are probably right - I read something a while ago about needing to use the 10.4 SDK because the 10.5 did not compile properly for the Tiger applications. No rush. I'll be here when it is ready to go.


OK, I tried it with the new update and it seems to be working just fine. I won't know until I actually have a unit complete and return for credit but it looks promising. Thanks again.


Plus SETI Classic = 21,082 WUs
ID: 49258 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile BladeD
Avatar

Send message
Joined: 2 Nov 10
Posts: 731
Credit: 131,536,342
RAC: 0
Message 49265 - Posted: 13 Jun 2011, 21:00:57 UTC - in response to Message 49239.  
Last modified: 13 Jun 2011, 21:12:00 UTC

Finished WUs, but I'm not getting more now! No error messages! How can it say not requiring tasks, when I have no WUs for two 5870s?

6/13/2011 4:22:24 AM Milkyway@home Sending scheduler request: Requested by user.
6/13/2011 4:22:24 AM Milkyway@home Not reporting or requesting tasks
6/13/2011 4:22:26 AM Milkyway@home Scheduler request completed

Resetting the project fix my problem.

Edit: Only got new WUs that one time!
ID: 49265 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Beyond
Avatar

Send message
Joined: 15 Jul 08
Posts: 383
Credit: 729,293,740
RAC: 0
Message 49271 - Posted: 13 Jun 2011, 23:22:56 UTC

Switched one machine over to .82. This is what i get:

<core_client_version>6.12.28</core_client_version>
<![CDATA[
<message>
Maximum elapsed time exceeded
</message>
<stderr_txt>
Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
Using SSE3 path
Found 2 CAL devices
Chose device 1

Device target: CAL_TARGET_770
Revision: 62
CAL Version: 1.4.900
Engine clock: 835 Mhz
Memory clock: 500 Mhz
GPU RAM: 512
Wavefront size: 64
Double precision: CAL_TRUE
Compute shader: CAL_TRUE
Number SIMD: 8
Number shader engines: 1
Pitch alignment: 256
Surface alignment: 4096
Max size 2D: { 8192, 8192 }

Estimated iteration time 371.049775 ms
Target frequency 30.000000 Hz, polling mode 1
Dividing into 11 chunks, initially sleeping for 0 ms
Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Using 11 chunk(s) with sizes: 144 144 144 144 144 144 144 144 144 144 160
Integration time = 221.760294 s, average per iteration = 346.500459 ms
Integral 0 time = 224.180060 s
Likelihood time = 2.649498 s
<background_integral> 0.000921986123313 </background_integral>
<stream_integral> 306.928839958010830 217.682575146146350 1604.083839551776500 </stream_integral>
<background_likelihood> -3.624712177160628 </background_likelihood>
<stream_only_likelihood> -30.318720303955470 -4.078689643335152 -4.493778141271900 </stream_only_likelihood>
<search_likelihood> -3.101747195361835 </search_likelihood>
<search_application> milkywayathome_client separation 0.82 Windows x86_64 double CAL++ </search_application>
18:13:19 (5924): called boinc_finish

</stderr_txt>

All WUs abort at about 70% with "Maximum elapsed time exceeded". Going back to .62.
ID: 49271 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Nigel Garvey

Send message
Joined: 2 Apr 11
Posts: 14
Credit: 4,527,461
RAC: 0
Message 49292 - Posted: 14 Jun 2011, 16:34:18 UTC
Last modified: 14 Jun 2011, 16:35:03 UTC

The PowerPC Mac OS build of 0.80 appeared to be perfectly happy running in Mac OS 10.4.11 and BOINC 6.10.58, completing a work unit previously flunked by two Windows 7 machines running ati14 apps. However, the task's been given an inconclusive validation. Stderr output shows several iterations of the "Error loading Lua script…" message reported by others above.

The intial completion time estimate was about 8.5 hours, but it actually finished in exactly 14 hours 38 minutes. The older app typically took about 25 hours 40 minutes.

I now have a 0.82 task waiting to start which also had an initial time estimate of 8.5 hours until the 0.80 task completed, but has now changed its estimate to exactly 14 hours 38 minutes too!
ID: 49292 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Beyond
Avatar

Send message
Joined: 15 Jul 08
Posts: 383
Credit: 729,293,740
RAC: 0
Message 49295 - Posted: 14 Jun 2011, 17:00:44 UTC - in response to Message 49271.  

Switched one machine over to .82. This is what i get:
Maximum elapsed time exceeded
All WUs abort at about 70% with "Maximum elapsed time exceeded". Going back to .62.

Update, this was caused by the coincidental running of many of the bad test WUs in a row immediately after switching to v.82. It's working OK now.
ID: 49295 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · Next

Message boards : News : Separation updated to 0.80

©2024 Astroinformatics Group