Posts by HAL9000

21) Message boards : Number crunching : Is there a script to auto-abort a task if... (Message 1855089)
Posted 12 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.
22) Message boards : Number crunching : I have roughly 18 HP DL 360's from 2011-2013 (Message 1854904)
Posted 12 days ago by Profile HAL9000
Post:
On Monday day I can log into my PDU's and get a live reading. Im assuming roughly 4-500 for power

I had a Dell server with dual E5645's. Running 24 tasks at once the power draw was only ~300w.
Which is a bit more than my dual E5-2670 system running 32 tasks at once with a power draw of ~250w.

It is frustrating when perfectly good older hardware can't find a good use. I have a pretty nice dual X5470 system with 32GB & dual SAS controllers that I don't use because sitting at idle it draws ~200w.
23) Message boards : Number crunching : From FX to Ryzen (Message 1854893)
Posted 13 days ago by Profile HAL9000
Post:
I just posted a video on my experience with the R7 1700 so far. It includes benchmarks with the FX8370, i73930K, and i7-6950X. I have included the AVX MB app in my benchmarks. https://youtu.be/YpEOmxG3zV4

It looks like normal AR Arecibo tasks are running ~2hrs with the AVX app.
Is that at the 3.5GHz OC you mentioned?

I would have hoped for run times in the 1-1.5hr range at that speed. Given my dual E5-2670 system @ 3.0GHz takes ~2.5hrs when running 32 at a time.
24) Message boards : Number crunching : CPU leaderboards (Message 1854228)
Posted 14 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 believe the Mac BOINC client reports the CPU with the extra x86 information that the other clients do not.
25) Message boards : Number crunching : From FX to Ryzen (Message 1854224)
Posted 14 days ago by Profile HAL9000
Post:
I've been reading non-stop for the past couple of weeks about Ryzen forum builds and all the goings on. Everybody is concerned about the reported CPU temperatures that they are seeing. What I haven't seen is anyone pulling out a thermocouple or contactless IR temperature gun and shooting the chip socket or cooler to see what the real temps are. Rick, do you have a temp gun in your toolbox by chance and can get a temperature reading around the socket or back of the socket on the motherboard backside? I really think the readings that people are seeing in the BIOS or monitoring programs are flaky. A lot of people are commenting also that the temperature of the air coming through an air cooler or through a water radiator is only mildly warm. I can't see how a 65W or 95W chip can produce 90° C. temps. The thermal Tmax dissipation for Ryzen 7 is only 128W. It's not like these new chips are producing the kind of waste heat the 220W TDP FX-9570 did.

Got fed up with Amazon about no information on my motherboard order after 2 weeks and have tried to cancel the order. They can't figure out how because I ordered it in a bundle. Ordered the ASUS Prime X370 Pro from Newegg earlier as it was back in stock. Was $10 cheaper too. Hope I see it next week sometime so I can get started on my build. Don't now what I'm going to do for cooling yet since still no word on the AIO bracket. I might have to buy an air cooler to get the build up and running.

Toms Hardware has been using an IR camera in their reviews for a while.
http://www.tomshardware.com/reviews/amd-ryzen-7-1800x-cpu,4951-11.html
26) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1854193)
Posted 15 days ago by Profile HAL9000
Post:
The Wiki says it was first supported in 7.0.40+ versions. There is only 7.4.42 available on the BOINC website above that cutoff level. I was on the level and had the problem even before updating to the latest recommended 7.6.33 version.

Really you can download any version of BOINC you want to play around with on your hosts.

Some people have reported that a few of the older releases seem to be corrupt. YMMV
27) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1853976)
Posted 15 days ago by Profile HAL9000
Post:

Note: Deselecting both MB & AP will set the selection to (all apps). Selecting MB or AP, saving, and then selecting the 2nd app will not set the selection to (all apps). ¯\_(º¿º)_/¯

Test 1 & Test 4 resulted in no work being sent. As we have little to no AP work and the Other apps function is still broken.
I never got a "project has no tasks" message for those two tests. Just the "No tasks sent" "No tasks are available for AstroPulse v7" messages.



. . To me it seems to be related to the "Other Apps" function breakdown in that the symptoms emerged about the same time. But since I did not myself use that setting I may be wrong and that malfunction may have occurred earlier.

Stephen

¯\_(º¿º)_/¯ {sorry for pilfering but I liked it :) and cannot make it myself}

Wait... you stole my ASCII art! RAWR! (╯°□°)╯︵ ┻━┻

The issue does seem to have started for some users at the same time. With the supposed spaghetti code in the scheduler I wouldn't be surprised if it turned out to be something weird and seemingly unrelated effecting those users.
28) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1853929)
Posted 16 days ago by Profile HAL9000
Post:
It is a really odd issue. Since I wasn't having any issue I did some testing changing my settings all around and I was still not running into the issue.
I guess whatever the issue plaguing some users is there is another part related to it that I can not reproduce.

It may (or may not) be related to the BOINC manager version and/or whether or not you have an application for AP installed, or not.

The client version could be a variable. I was using 7.6.23 before switching to 7.6.33 on my Windows hosts.
I currently have 5 Windows hosts with 7.6.33 & 1 Linux AWS host with 7.4.42. They are all configured with an app_info.xml that include these apps:
8130144 MB: CPU, AP: none, Venue: default
5837483 MB: CPU & ATI, AP: ATI, Venue: home
7324426 MB: Nvidia, AP: IGPU, Venue: home
8146263 MB: iGPU & Nvidia, AP: none, Venue: school
7421701 MB: CPU, AP: CPU, Venue: default
8115499 MB: CPU, AP: none, Venue: default

My venues settings are:
Venue		default	home	school	work
Use CPU		yes	no	no	yes
Use ATI GPU	no	yes	yes	yes
Use NVIDIA GPU	no	yes	yes	yes
Use Intel GPU	no	yes	no	yes

Then I cycled through these options
Test		1	2	3	4	5	6
AstroPulse v7:	yes	no	yes	yes	no	yes
SETI@home v8:	no	yes	yes	no	yes	yes
Other apps:	no	no	no	yes	yes	yes
Note: Deselecting both MB & AP will set the selection to (all apps). Selecting MB or AP, saving, and then selecting the 2nd app will not set the selection to (all apps). ¯\_(º¿º)_/¯

For each test cycle I would change the setting on the site & update the client to ensure it had the latest config.
Then I would wait and see if the tasks on the host would go down.

Test 1 & Test 4 resulted in no work being sent. As we have little to no AP work and the Other apps function is still broken.
I never got a "project has no tasks" message for those two tests. Just the "No tasks sent" "No tasks are available for AstroPulse v7" messages.

Test 2, 3, 5, & 6 resulted in work being sent or a "This computer has reached a limit on tasks in progress" message.

As of this moment my venue settings are:
Venue		default	home	school	work
AstroPulse v7:	no	yes	no	yes
SETI@home v8:	yes	yes	yes	yes
Other apps:	no	no	no	no
29) Message boards : Number crunching : Panic Mode On (105) Server Problems? (Message 1853823)
Posted 16 days ago by Profile HAL9000
Post:
Hmm my Linux machine is once again struggling to get work, all other machines have full caches.
Down to 27 GPU WU now, I sure hope I get some work soon.....

Change the application settings, then hit Update in the BOINC manager to get the change to take effect. On the next Scheduler contact work should start to flow. Maybe.
There are a few Arecibo VLARs about, so that will complicate getting GPU work more than it already is.


You don't need to change application settings in order to claim new wu. Just click on 'Update' or run
./boinccmd --project http://setiathome.berkeley.edu/ update
Open your 'Event log' display to check new wu are coming....



. . Bonjour Laurent,

. . Merci for your attempt to help but this is not a lack of knowledge of the process. Of late there is a problem that does not, thankfully, affect everyone, but is affecting quite a few users. Agreed that sometimes kicking it in the pants by pressing update will score you a few tasks but often each attempt to obtain new work either by automatic or manual updates results in a "No new tasks available" message and no work. It has been plagueing (that looks wrong however I spell it ) some users for a couple of months now and there seems to be no fix in sight, and sadly no feedback. I hope you are never affected by it.

Stephen

.

It is a really odd issue. Since I wasn't having any issue I did some testing changing my settings all around and I was still not running into the issue.
I guess whatever the issue plaguing some users is there is another part related to it that I can not reproduce.
30) Message boards : Number crunching : Personal Goal (Message 1853506)
Posted 18 days ago by Profile HAL9000
Post:
Hello,
Just wanted to get it out there .
My personal goal is to reach 500,000 total credit by this year.
Should be easy. I'm currently at 4 Linux servers with plans for up to 16.

Eventually I want to reach 128 Linux servers (Virtual of course) running seti@home all day every day.

But yes - Let us continue our search

Thank You

With the 4 you have running average tasks in ~5 hours. They would produce around 1,500,000 credit after a year if running 24/7.
31) Message boards : Number crunching : PCIe x1 GPU cards? (Message 1853490)
Posted 18 days ago by Profile HAL9000
Post:
My J1900-ITX seems to feed a GTX 750 ti without any issues using a PCIe x1 to PCIe x16 adapter.

Since the J1900-ITX only has PCIe 2.0 a very fast GPU would probably seem some slowdown if used in a PCIe x1 slot.
If anyone is using a x60 or x70 GPU in a x1 slot the amount of slowdown could be determined.
32) Message boards : Number crunching : Grabbing the Total Credit and Recent average credit (Message 1853489)
Posted 18 days ago by Profile HAL9000
Post:
Hello,
Not sure where to post this so I will leave it here.

I'm writing some software based around seti@home as a fun side project.
How can I grab from my account the Total Credit and Recent average credit ?

Let us say I write a webpage and I want it to grab this information to display on the webpage
Does seti@home provide this type of information grabbing ?

Thank You

Stats are generated, almost, daily by the server and can be downloaded from here: http://setiathome.berkeley.edu/stats/

Alternatively you can use the local stats file in your BOINC folder for each project you want to track.
33) Message boards : Number crunching : 2 GTX 750ti- 24/7 why is credit falling (Message 1853253)
Posted 19 days ago by Profile HAL9000
Post:
Perhaps the effect is greater on much faster GPUs?

I'd say that's the case.
The faster the GPU crunches the work, the more support it needs from the CPU to keep it fed, and the greater the impact of any delays in the GPU sending & getting data.

That makes sense. The CPU seems to have very little effect on the final run time of tasks for my 750 ti and 950m. On the notebook I had to drop below 1.8GHz before any real changes occurred in run time vs its normal 3.2GHz.
34) Message boards : Number crunching : From FX to Ryzen (Message 1853250)
Posted 19 days ago by Profile HAL9000
Post:
[url]I have a suspicion that the sudden dearth of any motherboard for immediate purchase might be because the manufacturers have pulled all stock back from the distributors.[/url]

That is a pretty accurate interpretation of what I was told by my vendor yesterday (a major UK component distributor). The majority of the motherboard manufacturers had pulled stock, while for the others they were being faced with log-jammed phones to resolve BIOS issues.

Yes electrical instability is being touted as the main reason behind mobo shortages (several small independent review sites have now already flagged this issue).

Cheers.

A bit of a rocky launch. Maybe after the BIOS & socket issues are resolved there will be updated reviews.
So far I've been most disappointed in the power consumption and temps under full load of the 1800X.
35) Message boards : Number crunching : From FX to Ryzen (Message 1853118)
Posted 19 days ago by Profile HAL9000
Post:
Apparently there is an issue where AMD forgot to supply all the details about the changes to the AM4 socket for Ryzen to all of the MB & cooling hardware manufactures.
I imagine that may be the reason some MBs are not shipping.
36) Message boards : Number crunching : 2 GTX 750ti- 24/7 why is credit falling (Message 1853117)
Posted 19 days ago by Profile HAL9000
Post:
Hi all,

I installed SoG version 3584 and have it running but the CPU utilisation seems really high with these tasks.
I'm currently running two at the same time on a GTX 770 and often they both chew up a CPU core each:
[img ]https://i.imgsafe.org/b68880cb88.png[/img]

Is this normal?

You can use the -use_sleep option to reduce the CPU usage.
The readme does say that sleep can increase the run time for tasks, but I have not seen a great deal of change on my systems when using it.
Perhaps the effect is greater on much faster GPUs?
37) Message boards : Number crunching : CPU leaderboards (Message 1852991)
Posted 20 days ago by Profile HAL9000
Post:
I'm not too sure just how they come up with some of the numbers (eg 3.93 average cores per computer?).
.

Probably something as simple as:
Total number of reported CPUs/Total number of computers
There are a number of users that restrict the number of cores they use for BOINC via the cc_config.xml option <ncpus>.
So you will see their CPU information reported something like
GenuineIntel
Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz
(3 processors)
38) Message boards : Number crunching : How much RAC from a RX480? (Message 1852158)
Posted 23 days ago by Profile HAL9000
Post:
I'm looking for a pure GPU number, surely 20-25k includes the CPU, right? If not that would be an amazing RAC :)

Thanks!

My R9 390X seems to be good for ~27-29K running 1 task at a time. So I would estimate a RX 480 at ~22-23K.

Some have said their 1050 ti's are good for ~15k. That feels a bit high to me. Since a 750 ti is good for ~10k. I feel ~12-13k may be more accurate.
Wiggo has reported their GTX 1060's run ~18-19k.
39) Message boards : Number crunching : 2 GTX 750ti- 24/7 why is credit falling (Message 1851973)
Posted 24 days ago by Profile HAL9000
Post:
Thanks, I have installed the SoG app but still see Cuda 32 as the application running in Boinc manager. Is that normal?
Also stopped Collaz.

[img ]https://i.imgsafe.org/4c8cdbf1f0.jpg[/img]

It seems really odd to me to run the CUDA 3.2 app on a Maxwell GPU, but that irrelevant to the SoG app.

If you added the SoG app to your app_info.xml and left the CUDA app in. Then the tasks are likely still using the CUDA application. You can look in task manager and see which exe is actually running.
Probably the easiest way to switch apps is to run the GPU queue empty and then remove the CUDA app from your app_info.
If you are more adept at modifying xml files you can remove the CUDA app from your app_info and update your client_state.xml. However messing up in the client_state.xml can lead to the loss of all BOINC tasks/projects on the host.
40) Message boards : Number crunching : 2 GTX 750ti- 24/7 why is credit falling (Message 1851969)
Posted 24 days ago by Profile HAL9000
Post:
Guess I have been here too long. I have the latest app from Lunatics. I will have to try to find Mike's World. Any leads on where that is?
As of Collatz it's set to run 10% of the time. So it really should not bee that big a drain. I will turn it off for a bit and see if that helps. If a 750TI should do 7-10K then I should be doing a lot better with 2 of them.

BOINC does its best to try to run as requested. Short term it doesn't appears to work as expected, but if you look at it on the timescale of a year it does a pretty good job... mostly.


If you want some single GPUs for comparison.
I have a host with a 750 ti. Which runs 2 MB tasks at once with the SoG app 0 CPU tasks.
It is an EVGA FTW model running at 1345MHz.

I also have a notebook running the same configuration.
It has a GTX 950m, but that is really a rebadged and downclocked 750 ti running at 927MHz.


Previous 20 · 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.