Posts by JohnMD
|
1)
Message boards :
Number crunching :
Updated GPU Requirements (Currently not supporting GPU tasks)
(Message 75372)
Posted 4 May 2023 by  JohnMD
Post: The first thing to learn about "bang for my buck" is "bang for my electricity bill". How many kwh/$ will you sacrifice each day ? Some projects require double-precision. In general AMD er better here. Some projects require minimum video memory. There's no simple answer - you need to be more precise about your interests and goals. |
|
2)
Message boards :
News :
News General
(Message 75370)
Posted 4 May 2023 by JohnMD
Post: I have been monitoring run times for the gpu app, and they have been slowly increasing over the last couple of weeks. My guess is that WU's are getting more demanding - but only around 5% to date. An odd thing about them is that my bigger card (1650 max-q 4GB 1024 shaders) uses about double so much cpu as my smaller (mx350 2GB 640 shaders). We all know that gpu WU's get "normal" cpu priority contra cpu WU's "low" priority. I don't have other normal jobs running, and the 2 machines (intel 10 & 11) are running at similar speeds. Is it because the bigger card requires more exchange of data ? |
|
3)
Message boards :
Number crunching :
Large surge of Invalid results and Validate errors on ALL machines
(Message 68778)
Posted 22 May 2019 by JohnMD
Post: I have only CPU tasks - about 15% failure. One common factor is that ALL my affected WUs give 244 credits. In addition, many of the WUs have other error tasks - mostly other GPUs but also CPUs. This leads me to the conclusion that the application is not consistent for the larger WUs in different environments, where differing precision in intermediate work fields could be the culprit. |
|
4)
Message boards :
Number crunching :
error posting
(Message 68777)
Posted 22 May 2019 by JohnMD
Post: error |
|
5)
Message boards :
News :
New runs of MilkyWay Nbody out
(Message 68706)
Posted 8 May 2019 by JohnMD
Post: Thanks for GPUGrid preferences, Mr. Myers. It really helped. I believe I've found the main problem with why the preferences were lost. In the project_specific_prefs.inc file used in BOINC, there is a line of code which is necessary to show the preferences: Hello Eric. 40 years ago I was a system programmer charged with a local upgrade of IBM VM. We had a range of changes and additions to the previous version of the operating system - particularly resource accounting. One did not change the supplied application/operating system; one created source update files(add/del/repl) All these update files were checked and - if necessary - adjusted to the new upgrade. 40 years ago there was no recognised IT education - have we learnt anything since ? |
|
6)
Questions and Answers :
Web site :
How do I chose Modfit workunits
(Message 68694)
Posted 4 May 2019 by JohnMD
Post: I have been cancelling Nbody WUs to concentrate on Modfit. Now the project has stopped sending me WUs. Is this because of all these so-called failures ? |
|
7)
Questions and Answers :
Getting started :
Project download error
(Message 68335)
Posted 22 Mar 2019 by JohnMD
Post: Several days BOINC Manager can't download milkyway_1.46_windows_x86_64__opencl_ati_101.exe. Speed = 0.0KBps always. Status - active. Reload or refreash project not helped. Other projects has no such problems.Firewall - off (exeption had same result) You have hidden your machine(s) so don't expect hardware help. |
|
8)
Message boards :
News :
New Server Update
(Message 68325)
Posted 21 Mar 2019 by JohnMD
Post: I think I found the issue with the credit assignment. It seems that the newest BOINC server version no longer supports how we would set the credit per workunit. Should be a pretty easy fix, I am working on an implementation now. I got "old" credit for one of my WUs - https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=174248695 But all the others are still getting "new" credit |
|
9)
Message boards :
News :
Changing Workunit Priority to 1 from 0
(Message 67201)
Posted 4 Mar 2018 by JohnMD
Post: Hey EG, The phrase "Use Reliable Host" might explain it all. New hosts and reawakened old hosts have zero history and therefore no reliability. The object of the exercise is "Don't Use Unreliable Host" - can you find a parameter for that instead ? I too would like to rejoin... |
|
10)
Message boards :
News :
Running Modfit on MilkyWay@home
(Message 64759)
Posted 26 Jun 2016 by JohnMD
Post: Hi Jake, Have you any idea how many 32-bit Intel Core 2 machines there are out here ? Many upgraded to Windows 10. I would hardly call them extremely old systems If it's XP you can't support, so filter on NT 6 or higher instead of 64-bit. Do-it-yourself solutions are NOT the aim of BOINC. |
|
11)
Message boards :
Number crunching :
SSE4.1 slower than SSE3 ?
(Message 64360)
Posted 5 Mar 2016 by JohnMD
Post: Is it possible to force SSE3 on cpu with SSE4.x capability ? |
|
12)
Message boards :
Number crunching :
app 'milkyway_nbody', plan class 'mt' doesn't match any app
(Message 64352)
Posted 28 Feb 2016 by JohnMD
Post: This "app_version" used to work. Can someone tell me how to find the new version ? Thanks in advance |
|
13)
Message boards :
Number crunching :
Guidence from Project Team Requested
(Message 4229)
Posted 17 Jul 2008 by JohnMD
Post: After seeing some of the times, I'm thinking that the "medium" length ones will probably work out the best for everyone, so figure on timings roughly around what you're getting with the 373... series. It sounds like you know pretty well how many cycles are required for each unit. Whether they last 12 minutes or 12 hours on my P4 doesn't bother me as long as these cycles get translated into realistic (say, factor 2 ?) runtime estimates. One thing that does bother me is that the project can apparently choose unit sizes quite arbitrarily - are they getting 60 times the science from 12-hour units compared with 12-minute ones ? |
©2024 Astroinformatics Group