Posts by Richard Haselgrove

1) Message boards : Technical News : Fermi LAT Gamma-ray pulsar search #4 "FGRP4" (Message 140511)
Posted 4 days ago by Richard Haselgrove
Last night (co-incidence?) my photon mobo or cpu expired, and it was a few days away from retirement (something new in the pipeline). Apologies to my patient wing-men, but i could not get a peep off the bios, just a fan spinning silently and a sad red power light. If a moderator wants to abort any pending tasks for it, feel free, it has crunched its last cobblestone, it has gone to meets its maker....

If the hard disk is still OK, and the replacement arrives in time, you can copy the BOINC data folder to the new host and crunch the remaining tasks there.
2) Message boards : Technical News : Binary Radio Pulsar Search (Parkes PMPS XT) "BRP6" (Message 140423)
Posted 6 days ago by Richard Haselgrove
Sorry, copy/paste error. Should work now.

BM

It worked for a couple of days, but I came back from a trip away to find my Haswell hadn't received new work since early Friday morning. Similar messages about Beta applications in the server log, got a new task immediately when I re-enabled Beta apps in preferences.

All tasks for computer 8864187

Edit - same for my Ivy Bridge: computer 5744895
3) Message boards : Problems and Bug Reports : einsteinbinary_BRP4G_1.39_i686-apple-darwin__BRP4G-opencl-ati-lion (Message 140422)
Posted 6 days ago by Richard Haselgrove
Ok, we'll look into this after the weekend...

v7.4.42 displays an estimate of progress immediately, based on the initial estimate of total runtime, and increases that estimate asymptotically towards 100%. I wrote this up as "pseudo-progress" for FGRP4.

Pseudo-progress is replaced by real reported progress in the display at the first checkpoint. So Moldr's observations are consistent with the application never starting any real computation at all. It may be helpful to inspect the contents of the slot directory, starting with stderr.txt
4) Message boards : Cruncher's Corner : Parkes PMPS XT on GPUs - a general problem? (Message 140362)
Posted 8 days ago by Richard Haselgrove
The Intel issue apparently is a known problem (although I could not extract a solution from the posting below). I use a Dell system where I cannot install any other driver than those provided by Dell because I use other custom graphics hardware whose proper function might be affected by manually changing the drivers. Only a week ago the latest Dell system driver update allowed me to at least receive WUs for the Intel HD 4000 IGP from the Einstein@home project, because before this IGP was not recognized as an OpenCL capable device.

I run the iGPU on a Dell platform (Optiplex 9020 with Haswell i5), and I can - eventually - use the generic Intel driver installers to switch between drivers at will. To get started, you may need to force one to install manually: Intel's 'Installation ReadMe' file is actually pretty good, if you follow the instructions exactly. I found the 'have disk' route the best way to get started.

I've got to rush out now, but if you search the boards for "10.18.10.3621" - the most reliable driver found so far - your should be able to find a download link. Or some kind soul might post it here.
5) Message boards : Technical News : Binary Radio Pulsar Search (Parkes PMPS XT) "BRP6" (Message 140271)
Posted 11 days ago by Richard Haselgrove
Yup, I get that too.

2015-04-14 18:34:25.9262 [PID=27077] [mixed] sending non-locality work second
2015-04-14 18:34:25.9565 [PID=27077] [version] Checking plan class 'BRP6-opencl-ati'
2015-04-14 18:34:25.9565 [PID=27077] [version] beta test app versions not allowed in project prefs.
2015-04-14 18:34:25.9565 [PID=27077] [version] Checking plan class 'BRP6-opencl-intel_gpu'
2015-04-14 18:34:25.9566 [PID=27077] [version] beta test app versions not allowed in project prefs.
2015-04-14 18:34:25.9566 [PID=27077] [version] Checking plan class 'BRP6-cuda32'
2015-04-14 18:34:25.9566 [PID=27077] [version] beta test app versions not allowed in project prefs.
2015-04-14 18:34:25.9566 [PID=27077] [version] Checking plan class 'BRP6-cuda32-nv301'
2015-04-14 18:34:25.9567 [PID=27077] [version] beta test app versions not allowed in project prefs.
2015-04-14 18:34:25.9567 [PID=27077] [version] Checking plan class 'BRP6-opencl-ati'
2015-04-14 18:34:25.9567 [PID=27077] [version] beta test app versions not allowed in project prefs.
2015-04-14 18:34:25.9567 [PID=27077] [version] no app version available: APP#29 (einsteinbinary_BRP6) PLATFORM#9 (windows_x86_64) min_version 0
2015-04-14 18:34:25.9567 [PID=27077] [version] no app version available: APP#29 (einsteinbinary_BRP6) PLATFORM#2 (windows_intelx86) min_version 0
2015-04-14 18:34:25.9691 [PID=27077] [debug] [HOST#8864187] MSG(high) No work sent
2015-04-14 18:34:25.9691 [PID=27077] [debug] [HOST#8864187] MSG(high) see scheduler log messages on http://einstein5.aei.uni-hannover.de/EinsteinAtHome/host_sched_logs/8864/8864187
2015-04-14 18:34:25.9692 [PID=27077] Sending reply to [HOST#8864187]: 0 results, delay req 60.00
2015-04-14 18:34:25.9702 [PID=27077] Scheduler ran 0.403 seconds

Putting the Beta flag back on gets tasks with no problem.

2015-04-14 18:46:31.0403 [PID=31881] [mixed] sending non-locality work second
2015-04-14 18:46:31.0604 [PID=31881] [send] [HOST#8864187] will accept beta work. Scanning for beta work.
2015-04-14 18:46:31.0810 [PID=31881] [version] Checking plan class 'BRP6-opencl-ati'
2015-04-14 18:46:31.0811 [PID=31881] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2015-04-14 18:46:31.0811 [PID=31881] [version] No ATI devices found
2015-04-14 18:46:31.0811 [PID=31881] [version] Checking plan class 'BRP6-opencl-intel_gpu'
2015-04-14 18:46:31.0811 [PID=31881] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2015-04-14 18:46:31.0812 [PID=31881] [version] [HOST#8864187] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 10.18.10.3621; platform version: OpenCL 1.2; device version: OpenCL 1.2
2015-04-14 18:46:31.0812 [PID=31881] [version] driver version 1018103621, min: 0, max: 1018103906
2015-04-14 18:46:31.0813 [PID=31881] [version] Peak flops supplied: 1.84e+11
2015-04-14 18:46:31.0813 [PID=31881] [version] plan class ok
2015-04-14 18:46:31.0813 [PID=31881] [version] Checking plan class 'BRP6-cuda32'
2015-04-14 18:46:31.0813 [PID=31881] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2015-04-14 18:46:31.0813 [PID=31881] [version] No CUDA devices found
2015-04-14 18:46:31.0813 [PID=31881] [version] Checking plan class 'BRP6-cuda32-nv301'
2015-04-14 18:46:31.0814 [PID=31881] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2015-04-14 18:46:31.0814 [PID=31881] [version] No CUDA devices found
2015-04-14 18:46:31.0814 [PID=31881] [version] Checking plan class 'BRP6-opencl-ati'
2015-04-14 18:46:31.0814 [PID=31881] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2015-04-14 18:46:31.0814 [PID=31881] [version] No ATI devices found
2015-04-14 18:46:31.0814 [PID=31881] [version] Best version of app einsteinbinary_BRP6 is 1.52 ID 724 BRP6-opencl-intel_gpu (15.34 GFLOPS)
2015-04-14 18:46:31.0815 [PID=31881] [send] [HOST#8864187] [WU#216193597 PM0016_02971_76] using delay bound 1209600 (opt: 1209600 pess: 1209600)
6) Message boards : Technical News : Binary Radio Pulsar Search (Parkes PMPS XT) "BRP6" (Message 140262)
Posted 11 days ago by Richard Haselgrove
So, when I shut off Beta I get no units.

I haven't "shut off" Beta, but I'm certainly getting the new v1.52 app allocated as production tasks for Intel GPU (and had to revise both an app_config.xml and Process Lasso settings to keep up).

Edit - and for NVidia GPU. (checked another machine)
7) Message boards : Cruncher's Corner : BRP4G WU (Message 140179)
Posted 13 days ago by Richard Haselgrove
Oh! E@H is now involved in race for discoveries, isn't it?
There was no hurry for last 10 years seen by volunteers, especially between PALFA and E@H. Instead almost all the time we heard about scrupulous analysis being held over existing data to find something new in what is already superficially analyzed. Can you explain what for this hurry is when you have to move aside the main project purpose while concentrating on rather permanent and endless observation data crunching?

I don't think that's a fair objection. The primary purpose of this project is to study (potential) gravity wave detections by - yes - searching through the permanent and potentially endless observational output of the LIGO detectors. That goes on in the background, using the S6Bucket application(s), which for the time being only run on CPUs. We just don't talk about it much.

Alongside that, we can use different compute resources - GPUs - to search through different types of observation. It doesn't seem unreasonable to occasionally participate in the initial scan of brand-new observations from Arecibo, alongside the long-term re-analysis of the Parkes data - especially since neither of them is in competition with the resources used for the primary (LIGO) search.
8) Message boards : Technical News : Binary Radio Pulsar Search (Parkes PMPS XT) "BRP6" (Message 140178)
Posted 13 days ago by Richard Haselgrove
New driver 10.18.10.4176 from Intel. Did a bunch of BRP4's which seemed to be successful, now trying 1.52 beta app. First task is here

According the Mark's other thread, this one seems to be working for 3rd. Generation processors (Intel's preferred search term for Ivy Bridge). But don't use the matching 4th. Generation (Haswell) v10.18.14.4156 driver for Einstein, unless you find a way of elimination the 100% validation failure I saw when I tested it.
9) Message boards : Cruncher's Corner : Intel 10.18.10.4176 driver for HD 4000 graphics? (Message 140154)
Posted 14 days ago by Richard Haselgrove
I tried the previous v10.18.10.4061 driver on an HD 4000, and it was worse than the recommended (by us - for Einstein) v10.18.10.3621: slower,and more invalid results.

Perhaps you could test this one, and report the results in the Binary Radio Pulsar Search (Parkes PMPS XT) "BRP6" thread, since that's the primary focus for getting these apps/drivers/hardware incompatibilities sorted out.
10) Message boards : Problems and Bug Reports : Uninstall before upgrade (Message 140121)
Posted 15 days ago by Richard Haselgrove
A new version of what? If you mean BOINC, it's never been necessary to uninstall first - the installer does the uninstall operation first, if needed.

Personally, I shut down BOINC and exit the Manager myself before installing an upgrade, but even that's not strictly necessary.


Next 10

Home · Your account · Message boards

This material is based upon work supported by the National Science Foundation (NSF) under Grants PHY-1104902, PHY-1104617 and PHY-1105572 and by the Max Planck Gesellschaft (MPG). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the investigators and do not necessarily reflect the views of the NSF or the MPG.

Copyright © 2015 Bruce Allen