Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go ยป
No member browsing this thread
Thread Status: Active
Total posts in this thread: 55
Posts: 55   Pages: 6   [ Previous Page | 1 2 3 4 5 6 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 14958 times and has 54 replies Next Thread
rbotterb
Senior Cruncher
United States
Joined: Jul 21, 2005
Post Count: 401
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

Completed the last of my six WUs. All ran fine on my I7 Laptop with Win10. Started and stopped and restarted all of them several times with success. All six completed (five are now in Valid mode, one is still PV waiting on wingman).
[Jul 9, 2020 5:57:36 PM]   Link   Report threatening or abusive post: please login first  Go to top 
rbotterb
Senior Cruncher
United States
Joined: Jul 21, 2005
Post Count: 401
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

By the way, all six of my WUs ran in 2.5 to 2.7 Hrs on my laptop.
[Jul 9, 2020 5:58:04 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

The observations are curious. The first one I allowed to run alone with 7 others of assorted plumage

7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_6090_0 02:52:42 (02:38:05)

Then ran 2 concurrent with whatever in the corral, slightly faster

7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_5357_1 02:49:45 (02:43:55)
7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_6708_0 02:49:13 (02:42:50)

Finally 5 concurrent, and again slightly faster, very homogeneous 2:43 - 2:47

7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_7666_1 02:43:30 (02:41:25)
7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_8744_0 02:45:02 (02:43:15)
7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_3717_0 02:45:48 (02:43:55)
7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_2731_0 02:47:37 (02:45:07)
7.56 Beta - Mapping Cancer Markers BETA_MCM1_9890287_7191_0 02:46:58 (02:44:24)

Edit: Spellchecked.

Runtime, 99% efficient.
----------------------------------------
[Edit 1 times, last edit by Former Member at Jul 9, 2020 6:37:25 PM]
[Jul 9, 2020 6:10:27 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Sgt.Joe
Ace Cruncher
USA
Joined: Jul 4, 2006
Post Count: 7777
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

Then ran 2 concurrent with whatever in the coral, slightly faster

I am confused. What do little sea polyps have to do with the beta test ?
Cheers
----------------------------------------
Sgt. Joe
*Minnesota Crunchers*
[Jul 9, 2020 6:31:12 PM]   Link   Report threatening or abusive post: please login first  Go to top 
hchc
Veteran Cruncher
USA
Joined: Aug 15, 2006
Post Count: 837
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

The Linux on ARM binary is 32 bit. There were 140 results for Linux on ARM sent out.
Thanks,
Jonathan

I hope this isn't a big request, but it'd be really nice to see 64-bit binaries for MCM1 and OPN1 for Linux on ARM.

To quote DavidH's post from this thread:
Raspbian Buster is a 32 bit OS last I checked. If you have a RPi4 you can get up to a 50% performance boost switching to a 64 bit OS

https://www.cnx-software.com/2020/01/29/raspb...bit-and-64-bit-debian-os/

But if you use a 64 bit Linux OS with ARM (such as the raspberry pi 4), you will lose the ability to run OpenPandemics because they did not compile a 64 bit runtime

That link and some links within that show benchmarks that look promising.
1. 64-bit may be significantly faster than 32-bit in many crunching workloads. See this too.
2. 32-bit support on ARM is being phased out. The ARM Cortex A76 no longer supports it.

I'm looking to build out an "ARM Farm," but the biggest thing holding me back is the lack of 64-bit. Rosetta@home supports 64-bit, but not WCG at the moment.
----------------------------------------
  • i5-7500 (Kaby Lake, 4C/4T) @ 3.4 GHz
  • i5-4590 (Haswell, 4C/4T) @ 3.3 GHz
  • i5-3570 (Broadwell, 4C/4T) @ 3.4 GHz

----------------------------------------
[Edit 2 times, last edit by hchc at Jul 9, 2020 7:24:33 PM]
[Jul 9, 2020 7:02:51 PM]   Link   Report threatening or abusive post: please login first  Go to top 
DrMason
Senior Cruncher
Joined: Mar 16, 2007
Post Count: 153
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

Had 68 work units among all my machines. I can confirm that Windows runs about twice as efficient as Linux with this new binary. The work units seem to be well-behaved in general; stopping and restarting didn't lose a significant amount of progress or result in an error. There was only one non-nominal result, and there, the CPU time and elapsed time were very different. BETA_MCM1_9890287_4558_1 took 2.02 hrs cpu time and 4.11 hrs of elapsed time. (That was the sole result on the 3770k below, btw.)

Other than that, here are averages:
AMD Ryzen 5 3600 (linux) - 2.25 hrs
AMD Ryzen 5 3600 (windows) - 1.18 hrs
AMD Ryzen 7 1700 (linux) - 3.4 hrs
Epyc 7601 (linux) - 2.75 hrs
Epyc 7301 (linux) - 2.51 hrs
Intel 3770k (windows) - 2.02 hrs
Intel 4570 (windows) - 1.21 hrs
----------------------------------------

[Jul 9, 2020 7:56:10 PM]   Link   Report threatening or abusive post: please login first  Go to top 
armstrdj
Former World Community Grid Tech
Joined: Oct 21, 2004
Post Count: 695
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

I will take a look at the Linux performance and see if I can get a better performing build before I load the next batch.

Thanks,
Jonathan
[Jul 9, 2020 8:37:06 PM]   Link   Report threatening or abusive post: please login first  Go to top 
pvh513
Senior Cruncher
Joined: Feb 26, 2011
Post Count: 260
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

Got 44 of these WUs. Restarted most of them and they seem to be doing fine: 37 valid, 4 in PV, and 3 still ongoing. I am seeing very big differences in runtime on different systems though, far more than usual:

Epyc 7282: 2.7 hr
Epyc 7281: 4.3 hr
i7 6700K: 4.6 hr
Xeon Silver 4114: 6.3 hr
Xeon E5-2630 v3: 7.3 hr

All systems with hyperthreading/SMT enabled and fully loaded, and all running openSUSE Leap 15.1. It seems this binary prefers AMDs as they are significantly faster than the Intels....
[Jul 9, 2020 11:25:52 PM]   Link   Report threatening or abusive post: please login first  Go to top 
poppageek
Advanced Cruncher
Joined: Nov 16, 2004
Post Count: 99
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

2 of my RPi4 running Raspberry Pi OS have received a combined total of 4 MCM Beta work units. 2 are valid one is Pending and one active. I restarted Boinc on the one active and all seems OK. Work units running 11 to 12 hours and use about 159mb of RAM.

Cheers!
[Jul 9, 2020 11:56:41 PM]   Link   Report threatening or abusive post: please login first  Go to top 
TonyEllis
Senior Cruncher
Australia
Joined: Jul 9, 2008
Post Count: 265
Status: Recently Active
Project Badges:
Reply to this Post  Reply with Quote 
Re: Mapping Cancer Markers Beta Test July 8, 2020 [Issues thread]

Something different about running these Betas under Linux. Alerted by gkrellm showing large amounts of orange as well as green in the CPUs section began investigatiing. Normally WUs create virtually all green. On checking with 'top' the difference is very noticeable. Usually Wus show almost all of their CPU use under ni (nice), but these Beta WUs have a much lower ni component and show much more under sys (system). The production MCM WUs never like this, ran the same as all the others projects in this respect.
----------------------------------------
[Jul 10, 2020 3:13:11 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 55   Pages: 6   [ Previous Page | 1 2 3 4 5 6 | Next Page ]
[ Jump to Last Post ]
Post new Thread