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

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 24
Posts: 24   Pages: 3   [ 1 2 3 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 3108 times and has 23 replies Next Thread
Taurus Oldbull
Advanced Cruncher
US
Joined: Nov 26, 2020
Post Count: 53
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

There seems to be OS-rev dependent computation results or OS-rev dependent validation procedures for SCC workunits run in MacOS (Darwin).

I have Darwin 17.7.0 - the highest rev my old hardware can handle.

After processing about 350 workunits of SCC, I find a pattern that seems Darwin OS-version dependent:

Workunits with Minimum Quorum of 1 validate successfully for my computer with OS 17.7.0.


Workunits with Minimum Quorum of 2 validate successfully for my computer with OS 17.7.0 if a second computer with an OS rev less than 22.4.0 also completes the workunit.

If a third computer with OS 22.4.0 or 22.5.0 also completes the workunit, it will fail as an Error or Invalid.
Examples:
SCC1_0004156_MyoD1-B_63433 22.4.0 Invalid
SCC1_0004112_MyoD1-A_16066 22.4.0 Error
SCC1_0004095_MyoD1-A_10891 22.4.0 Error
SCC1_0004128_MyoD1-B_17612 22.4.0 Error
SCC1_0004137_MyoD1-B_23322 22.4.0 Invalid
SCC1_0004096_MyoD1-A_45574 22.4.0 Invalid
SCC1_0004150_MyoD1-B_43473 22.4.0 Invalid
SCC1_0004079_MyoD1-A_63751 22.4.0 Invalid
SCC1_0004113_MyoD1-A_17661 22.4.0 Invalid
SCC1_0004147_MyoD1-B_7426 22.4.0 Error
SCC1_0004142_MyoD1-B_57422 22.4.0 Invalid
SCC1_0004150_MyoD1-B_55170 22.4.0 Invalid
SCC1_0004150_MyoD1-B_59187 22.4.0 Invalid
SCC1_0004079_MyoD1-A_56816 22.4.0 Invalid
SCC1_0004128_MyoD1-B_58612 22.4.0 Invalid
SCC1_0004150_MyoD1-B_50539 22.4.0 Invalid
SCC1_0004103_MyoD1-A_19163 22.4.0 Invalid
SCC1_0004147_MyoD1-B_15353 22.4.0 Invalid
SCC1_0004156_MyoD1-B_16237 22.4.0 Invalid
SCC1_0004096_MyoD1-A_27080 22.4.0 Invalid
SCC1_0004103_MyoD1-A_31181 22.4.0 Invalid 22.2.0 Error
SCC1_0004095_MyoD1-A_52040 22.4.0 Invalid
SCC1_0004147_MyoD1-B_39037 22.4.0 Invalid
SCC1_0004132_MyoD1-B_34711 22.4.0 Invalid 22.2.0 Error
SCC1_0004069_MyoD1-A_65705 22.4.0 Invalid
SCC1_0004100_MyoD1-A_27464 22.4.0 Invalid
SCC1_0004129_MyoD1-B_12630 22.4.0 Error
SCC1_0004150_MyoD1-B_11571 22.4.0 Error 22.1.0 Error
SCC1_0004156_MyoD1-B_10456 22.4.0 Invalid
SCC1_0004156_MyoD1-B_10061 22.4.0 Invalid 22.3.0 Error
SCC1_0004071_MyoD1-A_22351 22.4.0 Error
SCC1_0004114_MyoD1-A_44744 22.4.0 Invalid
SCC1_0004150_MyoD1-B_38193 22.4.0 Invalid
SCC1_0004074_MyoD1-A_33265 22.4.0 Invalid
SCC1_0004069_MyoD1-A_19465 22.4.0 Invalid
SCC1_0004113_MyoD1-A_26204 22.4.0 Invalid
SCC1_0004092_MyoD1-A_16806 22.4.0 Error
SCC1_0004092_MyoD1-A_24983 22.4.0 Invalid
SCC1_0004129_MyoD1-B_27385 22.4.0 Invalid
SCC1_0004142_MyoD1-B_32569 22.4.0 Invalid
SCC1_0004137_MyoD1-B_27848 22.4.0 Invalid
SCC1_0004079_MyoD1-A_14255 22.4.0 Invalid
SCC1_0004081_MyoD1-A_35844 22.4.0 Error
SCC1_0004095_MyoD1-A_17371 22.4.0 Invalid 22.2.0 Error
SCC1_0004129_MyoD1-B_21631 22.4.0 Invalid
SCC1_0004129_MyoD1-B_29244 22.4.0 Invalid
SCC1_0004113_MyoD1-A_4274 22.4.0 Invalid 22.3.0 Error
SCC1_0004143_MyoD1-B_6322 22.4.0 Invalid
SCC1_0004129_MyoD1-B_4986 22.4.0 Invalid
SCC1_0004156_MyoD1-B_20318 22.4.0 Error
SCC1_0004147_MyoD1-B_6120 22.4.0 Invalid
SCC1_0004150_MyoD1-B_1536 22.4.0 Invalid
SCC1_0004113_MyoD1-A_22687 22.4.0 Invalid
SCC1_0004112_MyoD1-A_17611 22.4.0 Invalid
SCC1_0004143_MyoD1-B_9786 22.5.0 Invalid (beta Darwin rev)
SCC1_0004156_MyoD1-B_8412 22.4.0 Error
SCC1_0004128_MyoD1-B_11549 22.4.0 Invalid 22.2.0 Error
SCC1_0004069_MyoD1-A_21080 22.4.0 Error
SCC1_0004069_MyoD1-A_18176 22.4.0 Invalid
SCC1_0004081_MyoD1-A_70060 22.4.0 Error
SCC1_0004114_MyoD1-A_1253 22.4.0 Invalid 22.1.0 Error


Workunits with Minimum Quorum of 2 fail as Invalid for my computer with Darwin 17.7.0 if 2 other computers with Darwin 22.4.0 or 22.5.0 also complete the work unit. The 2 other computers will validate successfully.
Examples:
SCC1_0004079_MyoD1-A_41440 22.4.0 Valid 22.5.0 Valid 17.7.0 Invalid
SCC1_0004071_MyoD1-A_51071 22.4.0 Valid 22.5.0 Valid 17.7.0 Invalid
SCC1_0004069_MyoD1-A_12833 22.4.0 Valid 22.5.0 Valid 17.7.0 Invalid
SCC1_0004095_MyoD1-A_15922 22.4.0 (2x) Valid 17.7.0 Invalid
SCC1_0004113_MyoD1-A_13354 22.4.0 (2x) Valid 17.7.0 Invalid

This majority-OS-version-prevails pattern seems to indicate that there is a either a difference in the computation results or the validation process between a workunit run with MacOS (Darwin) earlier than 22.4.0 and the same workunit run with OS 22.4.0 or 22.5.0.

I have not seen this pattern with MCM or OPN work units.
[May 5, 2023 5:33:40 AM]   Link   Report threatening or abusive post: please login first  Go to top 
adriverhoef
Master Cruncher
The Netherlands
Joined: Apr 3, 2009
Post Count: 2089
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

First a focus on the Error. It is not normal that a third computer kicks in, so there must me some kind of error either on your side or on the side of your first wingman. Maybe the task didn't start before its deadline (causing the error). I can't tell, so it's just a guess. Can you rule this out, confirming that your machine completes all tasks in a timely fashion? Can you also rule out that your wingmen with the Error status complete their SCC1-tasks before their deadlines?

Adri
[May 5, 2023 8:37:36 AM]   Link   Report threatening or abusive post: please login first  Go to top 
TPCBF
Master Cruncher
USA
Joined: Jan 2, 2011
Post Count: 1932
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

Rather than being an OS revision issue, I suspect that this is much rather a problem of dissimilar CPU types.
On the 17.x system, you are most likely running straight x86 code on an x86 CPU.
On the 22.x system(s), there is most likely x86 code running on Apple's Rosetta x86 CPU emulator, with the physical CPU being an ARM based CPU (Apple's M1 or M2). And in that case, the problem is likely a peculiarity/bug in that Rosetta emulation that is causing this.

Highly unlikely that this is anything that can be fixed on the WCG/BOINC side, but much rather a problem that Apple would have admit to and fix on their end. Which might be equally unlikely...


Ralf
----------------------------------------

[May 5, 2023 3:18:56 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Unixchick
Veteran Cruncher
Joined: Apr 16, 2020
Post Count: 859
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

I got a 1% error rate on SCC WUs. since my queue is very short, I know I completed the WUs. I think you are right it is a majority win situation which could be a wrong valid. It is worth mentioning to the scientists who run SCC, so they can look at it.

https://www.worldcommunitygrid.org/contribution/workunit/288250899

I predict when this one is finished, it will be invalid for me.
----------------------------------------
[Edit 1 times, last edit by Unixchick at May 6, 2023 1:26:23 AM]
[May 6, 2023 1:20:18 AM]   Link   Report threatening or abusive post: please login first  Go to top 
TPCBF
Master Cruncher
USA
Joined: Jan 2, 2011
Post Count: 1932
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

I got a 1% error rate on SCC WUs. since my queue is very short, I know I completed the WUs. I think you are right it is a majority win situation which could be a wrong valid. It is worth mentioning to the scientists who run SCC, so they can look at it.

https://www.worldcommunitygrid.org/contribution/workunit/288250899

I predict when this one is finished, it will be invalid for me.
Well, out of curiosity, what is the CPU in your machine?

Ralf
----------------------------------------

[May 6, 2023 3:11:00 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Unixchick
Veteran Cruncher
Joined: Apr 16, 2020
Post Count: 859
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

Here is another one that went invalid - I think falsely.
https://www.worldcommunitygrid.org/contribution/workunit/295486916

I'm on a mac mini .
Dual -Core Intel Core i5 (2.6 GHz).

I'm back to running ARP, so I'm not running SCC at the moment.
[May 6, 2023 2:07:33 PM]   Link   Report threatening or abusive post: please login first  Go to top 
TPCBF
Master Cruncher
USA
Joined: Jan 2, 2011
Post Count: 1932
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

Here is another one that went invalid - I think falsely.
https://www.worldcommunitygrid.org/contribution/workunit/295486916

I'm on a mac mini .
Dual -Core Intel Core i5 (2.6 GHz
OK, was just a guess due to the OS version is a high chance that you could have been running on an ARM based chip and that the Rosetta x86 emulator has a problem with a certain series of (FPU) instructions.I am aware that there were issue with those kind of things that took Apple ages to acknowledge and fix...

Ralf
----------------------------------------

[May 7, 2023 12:27:29 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Ian C
Cruncher
Joined: Jul 28, 2022
Post Count: 27
Status: Offline
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

What MacOS version is your old kit running? Does it use Apple's new security model (from Catalina on, iirc)? I (2018 In tell Mac mini) had a fair few errors until I made sure that Boinc had full disk access enabled in Security & Privacy in Settings (or Preferences as it was before Ventura).

Still get the odd one though.
[May 9, 2023 12:31:51 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Taurus Oldbull
Advanced Cruncher
US
Joined: Nov 26, 2020
Post Count: 53
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

I'm running masOS High Sierra version 10.13.6 on a
Mid-2011 27" iMac + 2x external 27" monitors
3.4 GHz Intel Core i7-2600 w/ 4 cores + Hyperthreading
32 GB RAM, 4 TB SSD, AMD Radeon HD 6970M 2GB
(I really like my maxed-out old setup! Still a pleasure to use.)
[May 9, 2023 1:48:04 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Ian C
Cruncher
Joined: Jul 28, 2022
Post Count: 27
Status: Offline
Reply to this Post  Reply with Quote 
Re: SCC computation results (or validation process) for MacOS (Darwin) seems OS-rev dependent

Not the new security model then.

Yeah, High Sierra was really solid.

I presume you've done all the usual Mac troubleshooting (Disk Utility/ reset SMCs etc) just to rule out issues at your end.
[May 9, 2023 10:11:58 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 24   Pages: 3   [ 1 2 3 | Next Page ]
[ Jump to Last Post ]
Post new Thread