Problems with MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu

Message boards : Number crunching : Problems with MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5126
Credit: 276,046,078
RAC: 462
Message 2028093 - Posted: 17 Jan 2020, 12:37:10 UTC

I just noticed these. They have a 900+ day crunching time estimate. Any ideas?

Tom


Application
Local: setiathome_v8 8.00 
Name
25mr13ab.8426.484291.8.35.198
State
Running
Received
Fri 17 Jan 2020 12:25:35 AM CST
Report deadline
Thu 06 Feb 2020 11:35:16 AM CST
Estimated computation size
5,875 GFLOPs
CPU time
00:27:34
CPU time since checkpoint
00:03:31
Elapsed time
00:27:42
Estimated time remaining
954d 11:09:47
Fraction done
0.002%
Virtual memory size
44.89 MB
Working set size
32.90 MB
Directory
slots/7
Process ID
17546
Executable
MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu




Application
Local: setiathome_v8 8.00 
Name
25mr13ab.8426.484291.8.35.208
State
Running
Received
Fri 17 Jan 2020 12:25:35 AM CST
Report deadline
Thu 06 Feb 2020 11:35:16 AM CST
Estimated computation size
5,875 GFLOPs
CPU time
00:27:30
CPU time since checkpoint
00:03:24
Elapsed time
00:27:36
Estimated time remaining
950d 13:46:57
Fraction done
0.002%
Virtual memory size
44.89 MB
Working set size
32.91 MB
Directory
slots/14
Process ID
17840
Executable
MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu

A proud member of the OFA (Old Farts Association).
ID: 2028093 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22526
Credit: 416,307,556
RAC: 380
United Kingdom
Message 2028095 - Posted: 17 Jan 2020, 12:45:00 UTC

Which of your computers?
Which processor on that computer?
Anything else running?
Without that information everyone is shooting in the dark
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2028095 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 2028098 - Posted: 17 Jan 2020, 12:55:41 UTC - in response to Message 2028093.  

I just noticed these. They have a 900+ day crunching time estimate. Any ideas?
Tom


. . I would suggest you suspend them both, let 2 more tasks begin and get well established then resume them. They should go waiting then restart on the next 2 CPU cores that become available. Monitor to see that they do not repeat the behaviour and they should run normally. If they do repeat the behaviour then 86 them :)

Stephen

. .

my 2c worth.
ID: 2028098 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 2028100 - Posted: 17 Jan 2020, 12:59:09 UTC - in response to Message 2028095.  

Looks like a computer problem, not a data problem. MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu looks like a homebrew / Anonymous Platform application, not one of Eric's. Can we move this query to another thread, please? I suggest he takes it up with the author of that app.
ID: 2028100 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5126
Credit: 276,046,078
RAC: 462
Message 2028106 - Posted: 17 Jan 2020, 13:37:39 UTC - in response to Message 2028095.  

Which of your computers?
Which processor on that computer?
Anything else running?
Without that information everyone is shooting in the dark


https://setiathome.berkeley.edu/show_host_detail.php?hostid=8853145

These are/were cpu tasks running on an Amd 3950x under Linux. The total threads I run are 28-29. Task manager claims ~90% cpu load. I am running 100% Seti@Home at the moment.

I had to reboot the computer because a lower cpu voltage experiment had made rendered it unresponsive after 15 odd minutes. Now, none of the tasks are showing extreme, long time estimates.

Have to run off to work in the midst of freezing rain. Will see if I can "find" these tasks after work.
Tom
A proud member of the OFA (Old Farts Association).
ID: 2028106 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 2028111 - Posted: 17 Jan 2020, 13:54:28 UTC

If a task has done 0.002% in 27 minutes 34 seconds, then linear extrapolation suggests that the whole task will take 83,100,000 seconds - or about 961.8 days. I reckon BOINC's estimate of 954 days 11 hours was pretty close, if a little low.

Alternatively: without volts, CPUs don't run very fast. Maybe your computer would have responded if you'd waited about 3 years?
ID: 2028111 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 2028144 - Posted: 17 Jan 2020, 18:09:19 UTC - in response to Message 2028106.  
Last modified: 17 Jan 2020, 18:16:56 UTC

Name
25mr13ab.8426.484291.8.35.198
Is this an Arecibo Shorty (VHAR)?
From the README_x41p_V0.98.txt included in the package;
8) The CPU Progress bar may stall with High AR Work Units, the task will finish normally, it's just the Progress Bar that's stalled.
If you aren't showing any Errors on those tasks I suggest you let them run and see how long it takes them to finish as per the ReadMe. It's Not just the Windows SoG App that doesn't like Arecibo (VHAR) Shorties, there are a number a Apps built from Raistmer's Code that don't respond well to Shorties.
ID: 2028144 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028148 - Posted: 17 Jan 2020, 19:12:53 UTC - in response to Message 2028144.  

Name
25mr13ab.8426.484291.8.35.198
Is this an Arecibo Shorty (VHAR)?
From the README_x41p_V0.98.txt included in the package;
8) The CPU Progress bar may stall with High AR Work Units, the task will finish normally, it's just the Progress Bar that's stalled.
If you aren't showing any Errors on those tasks I suggest you let them run and see how long it takes them to finish as per the ReadMe. It's Not just the Windows SoG App that doesn't like Arecibo (VHAR) Shorties, there are a number a Apps built from Raistmer's Code that don't respond well to Shorties.

Let's call them properly VHAR tasks with high angle range. The task in question has: WU true angle range is : 48.172349

An Arecibo "shorty" could just as well of been an early or late overflow with low crunch times.

I've been seeing lots of these "stalled" tasks lately from all the old Arecibo data we have been reprocessing. Other than the task getting shifted into high priority mode, they finish fine in the expected length for easy tasks. Just ignore the progress bar and the estimation.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028148 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028161 - Posted: 17 Jan 2020, 20:07:55 UTC - in response to Message 2028100.  

Looks like a computer problem, not a data problem. MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu looks like a homebrew / Anonymous Platform application, not one of Eric's. Can we move this query to another thread, please? I suggest he takes it up with the author of that app.

I'm not sure what other thread is more appropriate. The thread is specifically about the latest alternative optimized cpu app that is provided in TBar's latest AIO. Just another in the line of AVX2 enabled cpu applications he has provided over the years.

Anybody run the new app through the benchmarks yet with a comparison run against the default r3711 SSE41 cpu app? Tom did you?
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028161 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22526
Credit: 416,307,556
RAC: 380
United Kingdom
Message 2028162 - Posted: 17 Jan 2020, 20:12:15 UTC - in response to Message 2028161.  

The post from Richard was the one that triggered the question's move from its initial location to a thread all of its own.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2028162 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028164 - Posted: 17 Jan 2020, 20:19:12 UTC - in response to Message 2028162.  

The post from Richard was the one that triggered the question's move from its initial location to a thread all of its own.

Oh. Wasn't aware the thread started elsewhere. Nevermind.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028164 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 2028168 - Posted: 17 Jan 2020, 20:33:33 UTC - in response to Message 2028161.  

Anybody run the new app through the benchmarks yet with a comparison run against the default r3711 SSE41 cpu app?
After spending DAYS trying to get the AKv8 OpenCL App to compile in the newer Linux systems it has come to my attention a large number of Optimizations were left out of the CPU Apps. It May be possible to build a New CPU App with better speed using those said Optimizations. Right now I'm trying to find the System calls named by Eric & RueiKe that should be left out of the OpenCL App. Not having much luck. Anyone know just WHERE in the code "funsafe-math-optimizations" & "unsafe-fp-opt" are lurking? The New OpenCL App is a little faster than r3584, I suspect a Newer CPU App can be faster than r3711 as well.
ID: 2028168 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22526
Credit: 416,307,556
RAC: 380
United Kingdom
Message 2028174 - Posted: 17 Jan 2020, 20:43:52 UTC

Can we stick to the problem about the AVX2 app, which is a CPU app, not a GPU app.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2028174 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028177 - Posted: 17 Jan 2020, 20:50:25 UTC - in response to Message 2028168.  

It should be in the configure file for the app. Either on the CPPFLAGS= line or CFLAGS= line or on the configure line itself. At least that is where I found and modified the math optimizations for my aarch_64 zi3v app.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028177 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 2028178 - Posted: 17 Jan 2020, 20:51:40 UTC - in response to Message 2028174.  

Not and have any input from Me.
ID: 2028178 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028179 - Posted: 17 Jan 2020, 20:51:40 UTC - in response to Message 2028174.  

Can we stick to the problem about the AVX2 app, which is a CPU app, not a GPU app.

Don't think there is any problem with the AVX2 app. Simply Tom not understanding that like all the cpu apps, it does not move the progress bar normally with VHAR tasks.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028179 · Report as offensive
Ian&Steve C.
Avatar

Send message
Joined: 28 Sep 99
Posts: 4267
Credit: 1,282,604,591
RAC: 6,640
United States
Message 2028181 - Posted: 17 Jan 2020, 20:53:30 UTC - in response to Message 2028168.  

those appear to be compiler options. not in the code itself.
Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

ID: 2028181 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22526
Credit: 416,307,556
RAC: 380
United Kingdom
Message 2028183 - Posted: 17 Jan 2020, 20:56:58 UTC - in response to Message 2028179.  

Then why has it taken so many posts for you to explain that to him (in public).
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2028183 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 2028200 - Posted: 17 Jan 2020, 21:46:47 UTC - in response to Message 2028181.  
Last modified: 17 Jan 2020, 21:56:26 UTC

those appear to be compiler options. not in the code itself.
I've searched a couple versions of the sah_v7_opt folder and can't find those lines anywhere in the files. I even moved the current r4101 sah_v7_opt to my Mac, which has a Much better search routine, and still can't find unsafe-fp-opt or funsafe-math-optimizations anywhere in the folder. On a Mac, if those lines were in the sah_v7_opt folder, it would find them.

For instance, if I search for CalcChirpData_kernel_df64_cl in the sah_v7_opt folder , the two files analyzeFuncs.cpp & MultiBeam_Kernels.cl instantly pop up, meaning those two files contain the line CalcChirpData_kernel_df64_cl in them. Searching for unsafe-fp-opt & funsafe-math-optimizations finds nothing...
ID: 2028200 · Report as offensive
Ian&Steve C.
Avatar

Send message
Joined: 28 Sep 99
Posts: 4267
Credit: 1,282,604,591
RAC: 6,640
United States
Message 2028209 - Posted: 17 Jan 2020, 22:27:40 UTC - in response to Message 2028200.  

Perhaps these were compiler options that Eric was adding himself, but not in the default configuration. So there may not be anything for you to remove.

“unsafe-fp-opt” doesn’t appear to be a valid compiler option. Google returns 0 results when trying for an exact match. Whatever option you’re looking for, it’s likely called something else.
Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

ID: 2028209 · Report as offensive
1 · 2 · 3 · Next

Message boards : Number crunching : Problems with MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu


 
©2024 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.