SETI@home v8 beta to begin on Tuesday

Message boards : News : SETI@home v8 beta to begin on Tuesday
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 53 · 54 · 55 · 56 · 57 · 58 · 59 . . . 99 · Next

AuthorMessage
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58905 - Posted: 6 Jul 2016, 14:20:32 UTC
Last modified: 6 Jul 2016, 14:42:08 UTC

Don't think the new app is working any better than 8.14 for me.

I made sure it could only run guppi tasks. I wasn't able to watch it continuously but, the last time I checked, the memory usage was gradually climbing and had got to over 330Mb.

That task got to 35% complete and then stopped and another task is now running. I can only presume that's due to the same error as before:-

"ERROR: OpenCL kernel/call 'Enqueueing kernel:PC_find_pulse_kernel_cl' call failed (-6) in file ..\analyzePoT.cpp near line 3287"

EDIT: Monitored the 2nd task more often. The task reached 391Mb memory at 35% complete before erroring out and the first task re-starting.
ID: 58905 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 58906 - Posted: 6 Jul 2016, 14:59:51 UTC
Last modified: 6 Jul 2016, 15:50:22 UTC

Ah well, then I'm sure it's no idea that I try it on my iGPU HD4600 either. I'm too tired after this terrible day anyhow.

Thanks SusieQ for testing it, so I don't have to. Maybe I'll try it on Friday evening, or maybe there is another version to test then.

Edit: I'm trying it anyhow now on main. Let's see how it goes.

Edit 2: No, I agree with SusieQ. Same terrible memory counting up, until it fails (goes into a pause) with in this case:

ERROR: OpenCL kernel/call 'Enqueueing kernel:PC_find_pulse_kernel_cl' call failed (-6) in file ..\analyzePoT.cpp near line 3351.
Waiting 30 sec before restart...

For now, I give up, and let the iGPU work with r3430 on main.
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 58906 · Report as offensive
JLDun
Volunteer tester
Avatar

Send message
Joined: 23 May 07
Posts: 108
Credit: 127,898
RAC: 0
United States
Message 58907 - Posted: 6 Jul 2016, 15:13:27 UTC
Last modified: 6 Jul 2016, 15:16:54 UTC

Slight change of subject...
(I would post in the "New Android App" thread, but it's locked, and asking to unlock it MIGHT be too much trouble.)

Added Host 79482 this week. Lollipop 5.0, Intel Atom x86_64 quad core, BOINC 7.4.43 (having problems getting .51 and .53 to launch).

Task 24260798, (-9) overflow, has:
<core_client_version>7.4.43</core_client_version>
<![CDATA[
<stderr_txt>
stackdumps unavailable
setiathome_v8 8.00 Revision: 3320 i686-linux-android-g++ (GCC) 4.6 20120106 (prerelease)
libboinc: BOINC 7.7.0

Work Unit Info:
...............
WU true angle range is : 0.429508
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrumUnrolled 0.000847 0.00000
stackdumps unavailable
setiathome_v8 8.00 Revision: 3320 i686-linux-android-g++ (GCC) 4.6 20120106 (prerelease)
libboinc: BOINC 7.7.0

Work Unit Info:
...............
WU true angle range is : 0.429508
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrum 0.000634 0.00000
stackdumps unavailable
setiathome_v8 8.00 Revision: 3320 i686-linux-android-g++ (GCC) 4.6 20120106 (prerelease)
libboinc: BOINC 7.7.0

Work Unit Info:
...............
WU true angle range is : 0.429508
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrum 0.000616 0.00000
sse2_ChirpData_ak8 0.022742 0.00000
v_vTranspose4x16ntw 0.002833 0.00000
AK SSE folding 0.002327 0.00000
SETI@Home Informational message -9 result_overflow
NOTE: The number of results detected equals the storage space allocated.

Flopcounter: 22121050747.570072

Spike count: 30
Autocorr count: 0
Pulse count: 0
Triplet count: 0
Gaussian count: 0
01:40:08 (5653): called boinc_finish(0)

</stderr_txt>
]]>


Task 24245694 , normal task with stops, restarts, an me using the tablet at the same time, includes:
<core_client_version>7.4.43</core_client_version>
<![CDATA[
<stderr_txt>
stackdumps unavailable
stackdumps unavailable
setiathome_v8 8.00 Revision: 3320 i686-linux-android-g++ (GCC) 4.6 20120106 (prerelease)
libboinc: BOINC 7.7.0

Work Unit Info:
...............
WU true angle range is : 0.429559
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrumUnrolled2 0.000559 0.00000
sse2_ChirpData_ak8 0.017137 0.00000
v_vTranspose4x16ntw 0.005607 0.00000
BH SSE folding 0.002456 0.00000
stackdumps unavailable
Restarted at 1.79 percent.
stackdumps unavailable
Restarted at 1.79 percent.
stackdumps unavailable
stackdumps unavailable
stackdumps unavailable
stackdumps unavailable
stackdumps unavailable
stackdumps unavailable
Restarted at 3.88 percent.
stackdumps unavailable
stackdumps unavailable
Restarted at 3.88 percent.
stackdumps unavailable
stackdumps unavailable
stackdumps unavailable
Restarted at 14.84 percent.
stackdumps unavailable
Restarted at 21.78 percent.
stackdumps unavailable
Restarted at 24.07 percent.
stackdumps unavailable
(and so on...)

ID: 58907 · Report as offensive
Zalster
Volunteer tester

Send message
Joined: 30 Dec 13
Posts: 258
Credit: 12,340,341
RAC: 0
United States
Message 58911 - Posted: 6 Jul 2016, 17:53:27 UTC - in response to Message 58906.  

@Susie & Tut

What version of SoG are you both running?

I know Raistmer just released the new version on his mail.ru site but I don't think it's made it's way to Beta yet (Someone please correct me if I"m wrong)

His new version is r3482

I'll post a link

https://cloud.mail.ru/public/J5f8/HGuG3Vp4R

I'll be back here tonight to run some 4 bit on his new version and see what they do with it
ID: 58911 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58914 - Posted: 6 Jul 2016, 20:18:00 UTC - in response to Message 58911.  

I thought SoG was just the NVidia version?

Tut and I were both running the 3482 Intel GPU client. Raistmer had created a new version he thought solved the memory leak and he wanted some pre-Beta release testing if possible.
ID: 58914 · Report as offensive
Zalster
Volunteer tester

Send message
Joined: 30 Dec 13
Posts: 258
Credit: 12,340,341
RAC: 0
United States
Message 58915 - Posted: 6 Jul 2016, 20:25:29 UTC - in response to Message 58914.  
Last modified: 6 Jul 2016, 21:11:59 UTC

@Susie you may be right.

I tend to just cluster most of his recent apps as SoG and they very well maybe others, sorry about that.

Ok, running r3482 now here on Beta

First BLC 8bit result
WU true angle range is : 0.008604
https://setiweb.ssl.berkeley.edu/beta//result.php?resultid=24266470

WU true angle range is : 0.429590
https://setiweb.ssl.berkeley.edu/beta//result.php?resultid=24266809

SETI@Home Informational message -9 result_overflow
WU true angle range is : 0.429511
https://setiweb.ssl.berkeley.edu/beta//result.php?resultid=24265402

sorry I forgot the -v6 and -use_sleep, I'll put those in now

Edit...

WU true angle range is : 0.429512
https://setiweb.ssl.berkeley.edu/beta//result.php?resultid=24267214

WU true angle range is : 0.008604
https://setiweb.ssl.berkeley.edu/beta//result.php?resultid=24266865
ID: 58915 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 58916 - Posted: 7 Jul 2016, 8:46:41 UTC - in response to Message 58906.  

Ah well, then I'm sure it's no idea that I try it on my iGPU HD4600 either. I'm too tired after this terrible day anyhow.

Thanks SusieQ for testing it, so I don't have to. Maybe I'll try it on Friday evening, or maybe there is another version to test then.

Edit: I'm trying it anyhow now on main. Let's see how it goes.

Edit 2: No, I agree with SusieQ. Same terrible memory counting up, until it fails (goes into a pause) with in this case:

ERROR: OpenCL kernel/call 'Enqueueing kernel:PC_find_pulse_kernel_cl' call failed (-6) in file ..\analyzePoT.cpp near line 3351.
Waiting 30 sec before restart...

For now, I give up, and let the iGPU work with r3430 on main.


What Intel APUs you checked and what OpenCL driver versions were?
Such different behavior for iGPU app on different hosts...
My host uses:

 Core(TM) i5-3470 CPU @ 3.20GHz [Family 6 Model 58 Stepping 9]

INTEL Intel(R) HD Graphics 2500 (1195MB) OpenCL: 1.2

Driver version:				 10.18.10.3621

News about SETI opt app releases: https://twitter.com/Raistmer
ID: 58916 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 58919 - Posted: 7 Jul 2016, 8:58:47 UTC

@Tutankhamon and SusieQ

Could you try to run v8.12 (r3430) of iGPU MB with -use_sleep option enabled for some time, please.

Will it show similar memory leak with -use_sleep enabled?
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 58919 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58920 - Posted: 7 Jul 2016, 9:01:42 UTC

My host is:

Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz [Family 6 Model 94 Stepping 3]

Intel(R) HD Graphics 530 (4859MB) OpenCL: 2.0

Driver version: 20.19.15.4326
ID: 58920 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 58921 - Posted: 7 Jul 2016, 9:06:40 UTC - in response to Message 58920.  

My host is:

Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz [Family 6 Model 94 Stepping 3]

Intel(R) HD Graphics 530 (4859MB) OpenCL: 2.0

Driver version: 20.19.15.4326

And 10.18.10.3621 will just not support your type of device?
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 58921 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58922 - Posted: 7 Jul 2016, 9:07:20 UTC - in response to Message 58919.  

@Tutankhamon and SusieQ

Could you try to run v8.12 (r3430) of iGPU MB with -use_sleep option enabled for some time, please.

Will it show similar memory leak with -use_sleep enabled?

Will happily run this if you can tell me how - I usually only run non-customised stock apps.

My first foray into using customised files didn't quite work out as planned. When I restarted BOINC manager I'd lost all the tasks that had been waiting to run and it only downloaded GPU tasks so my CPU sat doing nothing :-)
ID: 58922 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58923 - Posted: 7 Jul 2016, 9:09:14 UTC - in response to Message 58921.  

My host is:

Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz [Family 6 Model 94 Stepping 3]

Intel(R) HD Graphics 530 (4859MB) OpenCL: 2.0

Driver version: 20.19.15.4326

And 10.18.10.3621 will just not support your type of device?

Will see if I can download this from the Intel site and give it a go.

I had previously tested 10.18.15.4279 when trying to sort out the invalid/errors, and found that it was giving me more problems.
ID: 58923 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 58925 - Posted: 7 Jul 2016, 9:21:26 UTC - in response to Message 58922.  

@Tutankhamon and SusieQ

Could you try to run v8.12 (r3430) of iGPU MB with -use_sleep option enabled for some time, please.

Will it show similar memory leak with -use_sleep enabled?

Will happily run this if you can tell me how - I usually only run non-customised stock apps.

My first foray into using customised files didn't quite work out as planned. When I restarted BOINC manager I'd lost all the tasks that had been waiting to run and it only downloaded GPU tasks so my CPU sat doing nothing :-)


For the sake of this particular testing you did OK actually (w/o CPU apps iGPU will completetasks faster ;) ).
But if you want to do different types of work you need to add other blocks into app_info.xml too. Each aistub contains pre-prepared app_info.xml file for only binary in that pack. Combining few packs would require copy/past corresponding blocks w/o outermost tags for each pack into app_info.xml.
More info about app_info.xml format: http://boinc.berkeley.edu/wiki/Anonymous_platform

For requested test you could just download this pack: https://cloud.mail.ru/public/HYxd/yLNx5jzoj
do exactly same as with r3482
and then add -use_sleep between <cmdline></cmdline>
tags. And restart BOINC.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 58925 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58927 - Posted: 7 Jul 2016, 9:30:02 UTC

I've got it running as you described so hopefully it should start producing results soon.

I got mostly non-guppi tasks but I have suspended those for the moment. Will keep an eye on memory usage and report back.
ID: 58927 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58928 - Posted: 7 Jul 2016, 10:33:40 UTC

So far, so good. The running guppi task has gone beyond 35% and hasn't stopped yet. Even better the memory hasn't climbed, in fact it's actually dropped about 20Mb from when the task started.

This is not exactly a like-for-like test though as, when I was running the 3482 app yesterday, I didn't have any command line switches.

PS Not able to find any older drivers to test, although I have tracked down a couple of newer ones. Will leave it running on 20.19.15.4326 for now though as the memory leak error occurred with this driver when running 8.14 and 3482.
ID: 58928 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 58929 - Posted: 7 Jul 2016, 10:55:02 UTC - in response to Message 58916.  

Ah well, then I'm sure it's no idea that I try it on my iGPU HD4600 either. I'm too tired after this terrible day anyhow.

Thanks SusieQ for testing it, so I don't have to. Maybe I'll try it on Friday evening, or maybe there is another version to test then.

Edit: I'm trying it anyhow now on main. Let's see how it goes.

Edit 2: No, I agree with SusieQ. Same terrible memory counting up, until it fails (goes into a pause) with in this case:

ERROR: OpenCL kernel/call 'Enqueueing kernel:PC_find_pulse_kernel_cl' call failed (-6) in file ..\analyzePoT.cpp near line 3351.
Waiting 30 sec before restart...

For now, I give up, and let the iGPU work with r3430 on main.


What Intel APUs you checked and what OpenCL driver versions were?
Such different behavior for iGPU app on different hosts...
My host uses:

 Core(TM) i5-3470 CPU @ 3.20GHz [Family 6 Model 58 Stepping 9]

INTEL Intel(R) HD Graphics 2500 (1195MB) OpenCL: 1.2

Driver version:				 10.18.10.3621

I run :

Name: Intel(R) HD Graphics 4600
Vendor: Intel(R) Corporation
Driver version: 10.18.10.3907
Version: OpenCL 1.2
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 58929 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 58930 - Posted: 7 Jul 2016, 10:59:09 UTC - in response to Message 58919.  
Last modified: 7 Jul 2016, 11:01:09 UTC

@Tutankhamon and SusieQ

Could you try to run v8.12 (r3430) of iGPU MB with -use_sleep option enabled for some time, please.

Will it show similar memory leak with -use_sleep enabled?

Easy, since I already run r3430 on the iGPU on main. When I get home, I will remove all commandline options I already have for the iGPU, and only run it with -use_sleep. -use_sleep isn't an option I have in the command file now.

Edit: Did you really mean r3430, as you wrote above, or did you mean the latest version??
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 58930 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 58931 - Posted: 7 Jul 2016, 11:28:39 UTC - in response to Message 58930.  


Edit: Did you really mean r3430, as you wrote above, or did you mean the latest version??

3430 - the one that works OK, w/o memory leak.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 58931 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 58932 - Posted: 7 Jul 2016, 11:34:11 UTC - in response to Message 58931.  


Edit: Did you really mean r3430, as you wrote above, or did you mean the latest version??

3430 - the one that works OK, w/o memory leak.


OK, will test that one with sleep when I get home.
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 58932 · Report as offensive
SusieQ
Volunteer tester

Send message
Joined: 12 Nov 10
Posts: 1148
Credit: 32,460,657
RAC: 1
United Kingdom
Message 58933 - Posted: 7 Jul 2016, 11:51:45 UTC

The first task has completed - took about 50% longer than previous guppi tasks.

Unfortunately ended as Inconclusive but at least it finished without restarting and without a memory leak.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=24264444
ID: 58933 · Report as offensive
Previous · 1 . . . 53 · 54 · 55 · 56 · 57 · 58 · 59 . . . 99 · Next

Message boards : News : SETI@home v8 beta to begin on Tuesday


 
©2021 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.