rpi_logo
Separation updated to 0.80
Separation updated to 0.80
log in

Advanced search

Message boards : News : Separation updated to 0.80

1 · 2 · 3 · Next
Author Message
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 49206 - Posted: 12 Jun 2011, 21:03:19 UTC

I've updated 15 of the 16 different versions of the separation for all systems. Post here if you have any problems (particularly if you have a problem with the new 32 bit Linux GPU applications...)

New features:
- Faster CPU calculations because of SSE2 intrinsics from Crunch3r. This should get a bit faster later when I get to some other stuff.

- The binaries now include the different SSE levels for the critical function (e.g. x87/SSE2/SSE3) all in one, and will use the appropriate one for the detected CPU's capabilities, so you don't need any special __sse2 version or anything. The problems on systems without SSE2 which happened sometimes for the GPU applications should also be fixed.

- Checkpointing for GPUs for both OpenCL/Nvidia and ATI/AMD for CAL. It will checkpoint no more frequently than after at least 10% progress, so it might not checkpoint as frequently as your settings specify if you have a particularly slow GPU. GPU checkpointing is a bit slow when it does happen, so if you don't want it, you can disable it with the flag --gpu-disable-checkpointing.

- The Nvidia OpenCL platform should now always be used avoiding issues if you had both AMD's and Nvidia's installed

- More reliable chunking for different GPUs. I didn't play with the time estimates so much, but I think the actual run times should now be closer to the estimates. There should also be fewer cases where a GPU will end up using the slowest possible work size option. I'll probably have to fiddle with this some more if people still aren't happy with the lag.

- Work around for a Catalyst driver problem where sometimes the GPU was reported as 0 Mhz, resulting in much more lag which I think was some peoples' problem.

- New flag that some people requested: --process-priority (-b). On Windows this is 0 (lowest) - 4 (highest) for overriding the process priority. On Linux this is the nice value.

- CAL specific: Removed --responsiveness-factor flag. Use --gpu-target-frequency or --non-responsive instead depending on what you want to do.

- Actually updated the 32-bit OS X application

- In the event of a crash on Windows, you should no longer be bothered with useless crash dialogs

Profile Simplex0
Avatar
Send message
Joined: 11 Nov 07
Posts: 232
Credit: 178,221,048
RAC: 0

Message 49208 - Posted: 12 Jun 2011, 21:20:37 UTC
Last modified: 12 Jun 2011, 21:31:04 UTC

No wu's for AMD GPU is downloading.
messages show...

2011-06-12 23:17:32 Milkyway@home Requesting new tasks for GPU
2011-06-12 23:17:34 Milkyway@home Scheduler request completed: got 0 new tasks
2011-06-12 23:17:34 Milkyway@home Message from server: No tasks sent
2011-06-12 23:17:34 Milkyway@home Message from server: Tasks for CPU are available, but your preferences are set to not accept them

I have also Resetting the project but I am not receiveing any files

TJ
Send message
Joined: 12 Aug 09
Posts: 262
Credit: 92,131,277
RAC: 8,547

Message 49209 - Posted: 12 Jun 2011, 21:28:57 UTC - in response to Message 49208.
Last modified: 12 Jun 2011, 21:41:59 UTC

I had the same but just when starting a post I saw a few WU's.
They seem to using less CPU time, so I guess less credits, but all are waiting ofr validation.

Now 21:28 UTC I do not get new WU's
____________
Greetings from,
TJ

Profile Werkstatt
Send message
Joined: 19 Feb 08
Posts: 350
Credit: 123,959,235
RAC: 4,538

Message 49211 - Posted: 12 Jun 2011, 21:37:48 UTC - in response to Message 49208.

No wu's for AMD GPU is downloading.
messages show...

2011-06-12 23:17:32 Milkyway@home Requesting new tasks for GPU
2011-06-12 23:17:34 Milkyway@home Scheduler request completed: got 0 new tasks
2011-06-12 23:17:34 Milkyway@home Message from server: No tasks sent
2011-06-12 23:17:34 Milkyway@home Message from server: Tasks for CPU are available, but your preferences are set to not accept them

I have also Resetting the project but I am not receiveing any files


Same here for me.

Profile wadnatze
Send message
Joined: 4 Jul 09
Posts: 2
Credit: 1,590,576
RAC: 0

Message 49213 - Posted: 12 Jun 2011, 21:39:25 UTC
Last modified: 12 Jun 2011, 21:40:26 UTC

I have the same Problem with Nvidia GTX 460.
5 WU's (from the Version befor 0.80) with the comment "Are not in Database".

cncguru
Avatar
Send message
Joined: 11 Jun 10
Posts: 329
Credit: 1,166,222,661
RAC: 0

Message 49214 - Posted: 12 Jun 2011, 22:10:14 UTC

OK yeah fine you updated all the seperation wu's.
Now could we please have some of them to crunch ???

Profile wadnatze
Send message
Joined: 4 Jul 09
Posts: 2
Credit: 1,590,576
RAC: 0

Message 49215 - Posted: 12 Jun 2011, 22:27:35 UTC

OK.

I got WU's now.

Profile Simplex0
Avatar
Send message
Joined: 11 Nov 07
Posts: 232
Credit: 178,221,048
RAC: 0

Message 49216 - Posted: 12 Jun 2011, 22:30:08 UTC

*LOL* This is just hilarious. Seti & Folding here I come.

swiftmallard
Avatar
Send message
Joined: 18 Jul 09
Posts: 289
Credit: 302,980,648
RAC: 0

Message 49217 - Posted: 12 Jun 2011, 23:16:55 UTC - in response to Message 49211.

No wu's for AMD GPU is downloading.
messages show...

2011-06-12 23:17:32 Milkyway@home Requesting new tasks for GPU
2011-06-12 23:17:34 Milkyway@home Scheduler request completed: got 0 new tasks
2011-06-12 23:17:34 Milkyway@home Message from server: No tasks sent
2011-06-12 23:17:34 Milkyway@home Message from server: Tasks for CPU are available, but your preferences are set to not accept them

I have also Resetting the project but I am not receiveing any files


Same here for me.

That's what happening for me also.

TJ
Send message
Joined: 12 Aug 09
Posts: 262
Credit: 92,131,277
RAC: 8,547

Message 49218 - Posted: 12 Jun 2011, 23:28:46 UTC

Hi Matt what have you done...

CPU work is comming, but GPU taskes no longer soon after the update to 0.80
The few I have got are all pending.
____________
Greetings from,
TJ

Profile arkayn
Avatar
Send message
Joined: 14 Feb 09
Posts: 999
Credit: 74,932,619
RAC: 0

Message 49219 - Posted: 12 Jun 2011, 23:34:59 UTC

Question on the Open CL apps, do they require a specific version of CUDA installed (ie. cuda23, cuda31, etc) or will they work as just plain cuda in an app_info?
____________

Profile arkayn
Avatar
Send message
Joined: 14 Feb 09
Posts: 999
Credit: 74,932,619
RAC: 0

Message 49220 - Posted: 13 Jun 2011, 0:52:39 UTC

I created versions of the new apps with app_info's.

I hope they are correct as I guessed on the CUDA OpenCL versions.
http://www.arkayn.us/forum/index.php?action=downloads;cat=11
____________

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 49221 - Posted: 13 Jun 2011, 1:00:47 UTC - in response to Message 49206.

The problem is the separation assimilator decided to break when I removed the old versions

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 49222 - Posted: 13 Jun 2011, 1:08:57 UTC - in response to Message 49221.

Separation work should be back

swiftmallard
Avatar
Send message
Joined: 18 Jul 09
Posts: 289
Credit: 302,980,648
RAC: 0

Message 49223 - Posted: 13 Jun 2011, 1:23:47 UTC
Last modified: 13 Jun 2011, 1:24:51 UTC

Got one WU, crunched it, now nothing happening.

edit - there we go

cncguru
Avatar
Send message
Joined: 11 Jun 10
Posts: 329
Credit: 1,166,222,661
RAC: 0

Message 49224 - Posted: 13 Jun 2011, 1:32:13 UTC

caches full all around, thank you Matt!

Len LE/GE
Send message
Joined: 8 Feb 08
Posts: 261
Credit: 104,050,322
RAC: 0

Message 49225 - Posted: 13 Jun 2011, 1:43:51 UTC
Last modified: 13 Jun 2011, 2:01:52 UTC

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 SSE2 path


The cpu is SSE3 able (Phenom II X4)

Estimated iteration time 152.839606 ms
Target frequency 30.000000 Hz, polling mode 1
Dividing into 4 chunks, initially sleeping for 0 ms
Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Using 4 chunk(s) with sizes: 400 400 400 400
Integration time = 102.426011 s, average per iteration = 160.040642 ms
Integral 0 time = 103.622172 s
Likelihood time = 2.450853 s


It was running without command line params.
gpu load 99%
cpu 25% (equal to 1 core) in busy wait.
system response sluggish up to boinc manager not responding.

Maybe you forgot to set the initial sleep time?


Estimated iteration time 152.839606 ms
Target frequency 30.000000 Hz, polling mode 1
Dividing into 4 chunks, initially sleeping for 0 ms
Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Using 4 chunk(s) with sizes: 400 400 400 400
Integration time = 102.783538 s, average per iteration = 160.599279 ms
[quote]


Tried to set --gpu-target-frequency 40, all WUs gave computation errors
Number of parameters doesn't make sense

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 49226 - Posted: 13 Jun 2011, 1:59:23 UTC - in response to Message 49225.

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
Using SSE2 path
The cpu is SSE3 able (Phenom II X4)
Not sure about this. I don't have any AMD systems so it's possible that might not be working correctly there.


It was running without command line params.
gpu load 99%
cpu 25% (equal to 1 core) in busy wait.
system response sluggish up to boinc manager not responding.

Maybe you forgot to set the initial sleep time?
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.

Estimated iteration time 152.839606 ms
Target frequency 30.000000 Hz, polling mode 1
Dividing into 4 chunks, initially sleeping for 0 ms
Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Using 4 chunk(s) with sizes: 400 400 400 400
Integration time = 102.783538 s, average per iteration = 160.599279 ms
It looks like the time estimates still isn't working out so great for you which explains the lag.

Profile cedricdd
Send message
Joined: 30 Jan 09
Posts: 13
Credit: 140,613,318
RAC: 0

Message 49227 - Posted: 13 Jun 2011, 2:04:32 UTC

When I put something in the <cmdline></cmdline> every WUs directly ends up with an error ::

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

paris
Avatar
Send message
Joined: 26 Apr 08
Posts: 75
Credit: 23,854,011
RAC: 14,231

Message 49228 - Posted: 13 Jun 2011, 2:13:49 UTC

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:

<core_client_version>6.10.56</core_client_version>
<![CDATA[
<message>
process got signal 5
</message>
<stderr_txt>
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.
____________

Plus SETI Classic = 21,082 WUs

1 · 2 · 3 · Next
Post to thread

Message boards : News : Separation updated to 0.80


Main page · Your account · Message boards


Copyright © 2018 AstroInformatics Group