Posts by HAL9000

1) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1857204)
Posted 7 hours ago by Profile HAL9000
Post:
The project has NEVER promised to keep us fed with work, they have ALWAYS told us to have standby projects to fall back onto in the event of (extended) outages.

But it would be nice if it were possible to carry at least 24 hours of work, even 48. The option is there in the settings- it would be appreciated if it were possible.
Why not raise the server side limits, but restrict all users to a maximum of 2 days cache, leaving more work available to faster crunchers to see them through these outages, but still limiting the load on the database?

If there is no data available to crunch, then there's no work.
But If there is work available, but the project has issues, it would be nice to be able to continue processing it.

It has been said that there will (eventually) be way more work than the present user base can process in a reasonable time, and more crunchers are needed. But it doesn't matter how many crunchers there are, if they can't process the work. And even if the servers spend most of a week down, if in their uptime people are able to get enough work to tide them through to the next uptime, people will continue to crunch.

Being able to have a 24 cache of SETI@home work would be nice.
I don't think a project could set a day cache max. Given that is a BOINC wide setting instead of a project setting. I don't doubt they would implement such a setting, but with the lack of BOINC development I don't expect anything like that to happen soon.
Doubling the current 100 CPU / 100 GPU*N limit would probably not be a great idea. Given we have been in the 5-6 million range of in progress fairly recently. Previously the db server was falling over when we were hitting ~10-11 million. I believe it was stated at the time that the server wasn't hitting a hardware limitation, but more of a software or database limitation.
At one time it was mentioned they were considering switching from the current infomatrix software, but that was some time ago. With Nebula I'm not sure how priorities may be heading now.
2) Message boards : Number crunching : Win10 -> Radeon HD 5xxx and HD 6xxx CRIMSON Drivers don´t use OpenCl ? Well not anymore (Message 1856830)
Posted 3 days ago by Profile HAL9000
Post:
For sure, they might be work, but u have a lot of performance loss if u don't use the crimson drivers.

Try it under Win10.

I try not to use Win10 unless I am forced.
3) Message boards : Number crunching : Problem With New Machine on BoincStats 1.69 (Message 1856718)
Posted 3 days ago by Profile HAL9000
Post:
I found the problem - I had gotten the BOINC installer from elsewhere than SETI - though it was the correct version (7.6.33), it was 32bit, rather than 64bit, and was incompatible with BoincTasks 1.69.

When I installed it, BT found all the Tasks, etc. that it couldn't find before.

So all seems cool, for now!

Well that all makes more sense now. I was wondering what BOINCstats 1.69 was.
I still prefer BOINCview to watch over my systems. Plus it doesn't switch to the task view at random like BOINC tasks does.
4) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1856635)
Posted 4 days ago by Profile HAL9000
Post:
Checking out Tasks and Settings taking ages to respond again, and the Replica is falling behind as well.

Looks like it is going for orbit
5) Message boards : Number crunching : 25% badge (Message 1856343)
Posted 5 days ago by Profile HAL9000
Post:
. . When I select the project I just get stats for the project overall, nothing about my stats. ??

Stephen

<shrug>

I'm not sure why you would expect to see any information about your account when looking at the project stats.
The purpose of looking at the project stats was to get the current active number of participants. Which can be used to calculate where the borders for each of the RAC badges are at that point in time.

If you want to see your information on BOINCstats one of the many ways to get there is by using the links under Cross-project statistics from your account on each project.
6) Message boards : Number crunching : Asymetric "loads" on GTX 750Ti's (different brands) (Message 1856340)
Posted 5 days ago by Profile HAL9000
Post:
I have been seeing some GPU tasks on my R9 390X and 750ti with larger CPU times as well. Roughly about twice the normal CPU time. I figured it had to do with the data in the tasks since the task run time was about the same as normal.
7) Message boards : Number crunching : Reprocessing (Message 1855946)
Posted 6 days ago by Profile HAL9000
Post:
As the search did not generate satisfying results, I'm posting this question here:
At the moment, my machine is processing 2008 data and I've seen many times before that we're processing "old" data.
Question is: Is it really old data or are we reprocssing it. If so, for what specific reason (i.e. new algorithms etc.).

Would be very interested to broaden my knowledge.

Greetings,
Michael

This post from Eric probably has much of the information you are looking for.
http://setiathome.berkeley.edu/forum_thread.php?id=78710#1752922
I'm sure if you want a more technical answer some of the developer guys can probably tell you more than you want to know about the code changes.
8) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855944)
Posted 6 days ago by Profile HAL9000
Post:
Otherwise, all the data-driven web pages would continue to run slow.

That's been the case for most of the last month or 2 anyway...

I have not seen the web pages running slow.
Are the people seeing the web pages being slow also the ones having issue keeping their cache filled?
9) Message boards : Number crunching : From FX to Ryzen (Message 1855854)
Posted 6 days ago by Profile HAL9000
Post:
Which one is the currently recommended optimized version for MB on Linux 64bit, http://lunatics.kwsn.info/index.php?action=downloads;cat=48 ?

Haven't done much SETI lately, but am right now trying to get an overview with which Intel CPUs my R7 @ 3.6GHz roughly compares with.

Thanks!

I'm not sure if anyone has done a comparison of the Linux apps yet.
I would probably start with the AVX app myself.
10) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855836)
Posted 6 days ago by Profile HAL9000
Post:
Seti is downloading a new executable ver. 8.05.
Does this mean we have to process 2008 *again*?
fk

I believe that is the new stock 64-bit windows app that has been under test at Beta.
11) Message boards : Number crunching : Win10 -> Radeon HD 5xxx and HD 6xxx CRIMSON Drivers don´t use OpenCl ? Well not anymore (Message 1855767)
Posted 7 days ago by Profile HAL9000
Post:
Since the release of Crimson drivers they have only officially supported GCN based GPUs.
The older Terascale based GPUs 15.12 is really the last driver to use. I've used the Windows 7 15.12 driver without problems in Windows 10 with my HD6870's.

I'm not sure if the Windows 10 version has OpenCL or not. It might be the same installer with a different name. I haven't checked.
http://support.amd.com/de-de/download/desktop/previous?os=Windows%2010%20-%2064
http://support.amd.com/de-de/download/desktop/previous?os=Windows%207%20-%2064
12) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855693)
Posted 7 days ago by Profile HAL9000
Post:
Some things not looking too good right now:
1) Replica DB offline,
2) AP work present but splitters not running,
3) Assimilators are down (due to 1 above, perhaps?)

Hopefully it's someone doing late work rather than a growing failure ...

Well the replica shows as disabled. Which is a better sign than not running.
There is a good chance that the loaded datasets have already been split for AP work. So i wouldn't get to excited about that.
13) Message boards : Number crunching : Are SoG harder to crunch for Nvidia ? (Message 1855455)
Posted 9 days ago by Profile HAL9000
Post:
There is no such thing as a CUDA or SoG workunit. What you are seeing is that application that was assigned to the task when the host requested work.

The SoG application is based on OpenCL which requires a bit more CPU support than do the CUDA applications.
If you have been using the app_config.xml to run 2 tasks at a time then you could start by setting the CPU usage for one CPU per task.

Additionally the SoG has tuning parameters that cam be added to the mb commandline txt to adjust how it runs.
You should have a "ReadMe_MultiBeam_OpenCL_NV_SoG.txt" in your \projects\setiathome.berkeley.edu\ folder that will describe the parameters.

For my GTX 750 ti I run 2 at once with the SoG app and the performance is pretty near the same as the CUDA5 app.
Except for GBT VLAR tasks. Which the SoG application handles much better.
14) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855412)
Posted 9 days ago by Profile HAL9000
Post:
Looks like some new datasets have made their way in for splitting.

Just in time too I only have 599/600 MB tasks!


Braggart! LOL

Stephen

:-P

It looks like my statement was actually not true. I was only looking at my In progress MB tasks. As I wasn't expecting to have any AP tasks in progress, but I seem to have snagged one. I had nearly forgotten what they look like.
15) Message boards : Number crunching : CPU leaderboards (Message 1855403)
Posted 9 days ago by Profile HAL9000
Post:
@ Gabriel Siegel:

Not only what you have noticed, but there are many instances of the same CPU being listed in two places. I counted 126 of these. For example,

line 6 (Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz [x86 Family 6 Model 60 Stepping 3] ) and
line 21 (Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz [Family 6 Model 60 Stepping 3] )

are the same CPU, but listed with slightly different specs (notice the "x86" in front of "Family").

I haven't found who/how to address such things since there is no contact info on the webpage.


. . I think that is the difference between users running a 32 bit OS versus a 64 bit OS. They identify differently, the lines with x86 in them are running 32 bit. Yet strangely they seem to outperform the same machines running on 64bit. All so counter-intuitive.

. . Also I think the strange values for number of cores is because some users run with hyperthreading off. So when you have 2,400 i7-6700K machines listed and 242 of them (not actual figures, just using them for example) are not running with HT and therefore reporting only 4 cores not 8, then the average will be an odd value like 7.52666. Sorry Hal but I don't think it is to do with the manager configuration to control the number of cores crunching. The benchmarks are run independently of that.

Stephen

:)

You are correct. In that changing "Use at most n% of the CPUs" will not change the number of reported processors by BOINC. However if you read my post I indicated using the <ncpus> value in the cc_config.xml. Which when modified will cause BOINC to report the number of processors you entered. So if you set a value such as <ncpus>100</ncpus> on your i5-6400 system then it would appear on the website and in the stats as:
GenuineIntel
Intel(R) Core(TM) i5-6400 CPU @ 2.70GHz [Family 6 Model 94 Stepping 3]
(100 processors)


I feel this image may explain what I was referring to regarding the "[x86 Family" processor information.
Additionally "Darwin" is how BOINC reports OS X.
16) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855395)
Posted 9 days ago by Profile HAL9000
Post:
Looks like some new datasets have made their way in for splitting.

Just in time too I only have 599/600 MB tasks!
17) Message boards : Number crunching : 25% badge (Message 1855386)
Posted 9 days ago by Profile HAL9000
Post:
Well you do have to select the project. Which requires you scroll down to where it says SETI@home and click on it.




Also the default BOINC RAC badges are per project. So I'm not sure why you would think it has anything to do with your BOINC wide stats.

Also "Accumulated more credit than %" would be irrelevant. Since that is total credit and the badges, once again, are based on RAC. If you pop over to your SETI@home Beta account. You will notice you don't have a badge there. Since your RAC is ~0.
18) Message boards : Number crunching : 25% badge (Message 1855330)
Posted 10 days ago by Profile HAL9000
Post:
I have earned a 25% badge but I am not sure what it means. I am a member of a team. Does this mean I am among the top 25% in daily credit for the team or for the entire SETI project?


. . Go to your account page, click on "BOINCstats" or "BOINC Combined Statistics" and you will see lots of info about your "standing" amongst BOINC and SETI participants. But the short answer is option number 2, you are among the top 25% of all BOINC contributors. Don't puff your chest out too far though, most of the BOINC clients out there are either inactive or have tried it, didn't like it and then left. So top 25% sadly just translates to something like "you are one of the truly active BOINC members", but welcome to the club. :) About 70% of all listed hosts have a RAC below 10, so it's that top 30% or thereabouts that really matters :)

. . Have fun and keep crunching

Stephen

:)

I think you meant SETI@home contributors rather than BOINC contributors.
Since the badge is for RAC it only includes users "with recent credit". I don't recall what the server considers "with recent credit" off the top of my head, but it is documented somewhere.
Also for reference BOINCstats reports SETI@home Total users: 1,674,642, Active users: 141,863 (8.47%)
For 141863 active users the badges levels are something like:
1% badge: Top 1419 users
5% badge: Top 7093 users
25% badge: Top 35466 users
19) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1855325)
Posted 10 days ago by Profile HAL9000
Post:
When you have been in correspondence with Matt/Jeff or Eric have you discussed the lack of Astropulse work? If so is there been any answer as to when/if there will be a steady flow of work? I was thinking about emailing one of them and asking however I did not want to do this if you have already raised the question

My understanding on this is that AP splits much faster, provides fewer WUs than MB, and therefore.. gets processed and returned faster.... and we've basically run out of AP almost entirely, until some brand new tapes arrive. We keep redoing older tapes that never got done with MB's newest autocorr check, so that's why those are all being redone, but AP hasn't really changed since it began. So finding tapes we haven't done with AP is few and far between.

The secondary problem is that starting at some point in 2015 and basically all of 2016, something with the tapes appear to be broken, either at the recording end or the splitting end, because 99% of them are 100% blanked when it comes to AP.

There was mention last year that GBT would begin being processed with AP at some point in the future, but it seems like they need to sort some other gremlins out with the supply of tapes first.

Also I believe Arecibo was offline much of 2015 or 2016.
20) Message boards : Number crunching : Is there a script to auto-abort a task if... (Message 1855089)
Posted 11 days ago by Profile HAL9000
Post:
I tossed this together really quickly and there are probably better ways to do it.
This will get the job done if you want to abort tasks that run over a specific limit.

@ECHO OFF
pushd %~dp0
cls
set outfile1=_task-name.txt
set outfile2=_task-project.txt
set outfile3=_task-checkpoint.txt
set skipcount=0
set bnccmd=boinccmd --get_tasks
set killproject=http://setiathome.berkeley.edu/
set runlimit=18000

@ECHO SOF SOF SOF SOF > %outfile1%
@ECHO SOF SOF SOF SOF > %outfile2%
@ECHO SOF SOF SOF SOF > %outfile3%
%bnccmd% | find "name:" | find /V "WU name:">> %outfile1%
%bnccmd% | find "URL:" >> %outfile2%
%bnccmd% | find "time:" | find /V "checkpoint CPU time:" | find /V "final CPU time:" >> %outfile3%
@ECHO EOF EOF EOF EOF >> %outfile1%
@ECHO EOF EOF EOF EOF >> %outfile2%
@ECHO EOF EOF EOF EOF >> %outfile3%

@ECHO Starting task run time check %date% %time%
:miniloop
set /a skipcount=%skipcount%+1
FOR /F "skip=%skipcount% tokens=4* delims= " %%a in (%outfile3%) do set runtime=%%a & goto checktime
:checktime
if %runtime% equ EOF goto end
if %runtime% equ 0.000000 goto miniloop
set /a runtime/=1
if %runtime% LSS %runlimit% goto miniloop
FOR /F "skip=%skipcount% tokens=3* delims= " %%a in (%outfile2%) do set projectname=%%a & goto checkproject
:checkproject
if %projectname% NEQ %killproject% goto miniloop
FOR /F "skip=%skipcount% tokens=2* delims= " %%a in (%outfile1%) do set taskname=%%a & goto killtask
:killtask
@ECHO Run time limit reached %runtime%/%runlimit% seconds. Killing task %taskname%
@ECHO %date% %time%, %runtime%/%runlimit%, %taskname% >> _kill_log.txt
boinccmd --task %projectname% %taskfile% abort
goto miniloop
:end
@ECHO Stopping task run time check %date% %time%
del %outfile1% & del %outfile2% & del %outfile3%


bnccmd - The boinccmd command line. Can be modified with --host --passwd values to run on remote hosts. See the BOINC wiki for full details on using boinccmd and configuring remote hosts.
killproject - The project you want to check
runlimit - The number of seconds that will trigger an abort.
_kill_log.txt - The log of tasks that have been aborted.


Next 20


 
©2017 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.