Welcome to MilkyWay@home

GPU in non-continuous operation.

Questions and Answers : Windows : GPU in non-continuous operation.
Message board moderation

To post messages, you must log in.

Previous · 1 · 2

AuthorMessage
Profile Keith Myers
Avatar

Send message
Joined: 24 Jan 11
Posts: 696
Credit: 540,043,442
RAC: 86,688
Message 74826 - Posted: 17 Dec 2022, 8:47:46 UTC - in response to Message 74800.  
Last modified: 17 Dec 2022, 8:57:45 UTC


Another point of benefit is to consolidate their size... the tasks are too small. I think making a single task from 4 or 6 of them would make sense (while still being viable on older gpu's). Such a granularity not only affects performance of the client negatively but server/s are likely stressed more due to it too.

They already have implemented your suggestion for the Separation tasks. Your task is actually 5 single tasks bundled into one composite task. Look at the stderr.txt output of one of your tasks and you will see that there are 8 individual task parameters and 5 results in the output file.

You can see that is the case by looking at the task naming. All the Separation task have bundle in their name.
ID: 74826 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Keith Myers
Avatar

Send message
Joined: 24 Jan 11
Posts: 696
Credit: 540,043,442
RAC: 86,688
Message 74827 - Posted: 17 Dec 2022, 8:53:37 UTC

Really the easiest way to keep a fast gpu busy is the multiple clients approach in the provided link. Or this guide.

https://www.overclock.net/threads/guide-setting-up-multiple-boinc-instances.1628924/

There is a publicly available custom Windows Milkyway client by Joseph Stateson that gets around the 10 minute project backoff issue.

https://github.com/JStateson/MilkywayNewWork
ID: 74827 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
ace_quaker

Send message
Joined: 1 Feb 09
Posts: 4
Credit: 101,666,161
RAC: 0
Message 74855 - Posted: 26 Dec 2022, 7:41:56 UTC

Just picked up a used VII and had never ran into this issue until now. Pretty Lame. Don't want to bend over backwards fiddling with multiple instances/modded clients. Guess I'll run Einstein? Unless China takes Taiwan this problem is only ever going to get worse the faster cards get. Needs solved permanently IMO.
ID: 74855 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Joseph Stateson
Avatar

Send message
Joined: 18 Nov 08
Posts: 291
Credit: 2,461,693,501
RAC: 0
Message 74901 - Posted: 12 Jan 2023, 2:17:00 UTC
Last modified: 12 Jan 2023, 2:18:23 UTC

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.
ID: 74901 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,947,576
RAC: 22,161
Message 74903 - Posted: 12 Jan 2023, 11:02:36 UTC - in response to Message 74901.  

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?
ID: 74903 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Joseph Stateson
Avatar

Send message
Joined: 18 Nov 08
Posts: 291
Credit: 2,461,693,501
RAC: 0
Message 74904 - Posted: 12 Jan 2023, 13:07:07 UTC - in response to Message 74903.  
Last modified: 12 Jan 2023, 13:27:16 UTC

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?


That Einstein mod is in Linux and AFAICT makes an improvement in performance of Linux applications. There was also a mod to BOINC: 7.17 and 7.19

It was done by Petri and possibly others using Linux
This Einstein@home App (v1.0 by petri33) was built at: Apr 28 2022 18:47:15


All I have done is modify the scheduling algorithm in BOINC to bypass that 91 second delay that the Milkyway server wants.
ID: 74904 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,947,576
RAC: 22,161
Message 74906 - Posted: 13 Jan 2023, 11:32:48 UTC - in response to Message 74904.  

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?


That Einstein mod is in Linux and AFAICT makes an improvement in performance of Linux applications. There was also a mod to BOINC: 7.17 and 7.19

It was done by Petri and possibly others using Linux
This Einstein@home App (v1.0 by petri33) was built at: Apr 28 2022 18:47:15


All I have done is modify the scheduling algorithm in BOINC to bypass that 91 second delay that the Milkyway server wants.


And your mod works great!! Thank you!!
ID: 74906 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Keith Myers
Avatar

Send message
Joined: 24 Jan 11
Posts: 696
Credit: 540,043,442
RAC: 86,688
Message 74907 - Posted: 14 Jan 2023, 8:44:51 UTC - in response to Message 74903.  

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?

Not sure what is broken at Einstein that needs a "fix" Mikey.

Can you elucidate the problem?

I will try and compile Joseph's Linux BOINC client and drop it into the Lunatics AIO version of BOINC on a spare PC for beta testing his MW "fix"

Only problem is all I have are Nvidia cards and he mentions they aren't fast enough to trigger the problem at MW.

But I used to run into the MW Scheduler problem all the time back when I was running MW at 100% resource share. Why I asked our dev to come up with the report_delay parameter for our Pandora client which he kindly put into the cs_scheduler.cpp for me even though I was the only team member running MW at 100% at that time.
ID: 74907 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,947,576
RAC: 22,161
Message 74908 - Posted: 14 Jan 2023, 12:09:27 UTC - in response to Message 74907.  

For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?

Not sure what is broken at Einstein that needs a "fix" Mikey.

Can you elucidate the problem?


Poor choice of words, the speed-up would have been a better choice I think

I will try and compile Joseph's Linux BOINC client and drop it into the Lunatics AIO version of BOINC on a spare PC for beta testing his MW "fix"[/quotd]

WOO HOO!!

[quote]Only problem is all I have are Nvidia cards and he mentions they aren't fast enough to trigger the problem at MW.

But I used to run into the MW Scheduler problem all the time back when I was running MW at 100% resource share. Why I asked our dev to come up with the report_delay parameter for our Pandora client which he kindly put into the cs_scheduler.cpp for me even though I was the only team member running MW at 100% at that time.


Hmmmm that does present a problem
ID: 74908 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Joseph Stateson
Avatar

Send message
Joined: 18 Nov 08
Posts: 291
Credit: 2,461,693,501
RAC: 0
Message 74909 - Posted: 14 Jan 2023, 14:43:30 UTC - in response to Message 74908.  
Last modified: 14 Jan 2023, 14:45:04 UTC

[quote]For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?

Not sure what is broken at Einstein that needs a "fix" Mikey.

Can you elucidate the problem?


Poor choice of words, the speed-up would have been a better choice I think

I will try and compile Joseph's Linux BOINC client and drop it into the Lunatics AIO version of BOINC on a spare PC for beta testing his MW "fix"[/quotd]

WOO HOO!!

Only problem is all I have are Nvidia cards and he mentions they aren't fast enough to trigger the problem at MW.

But I used to run into the MW Scheduler problem all the time back when I was running MW at 100% resource share. Why I asked our dev to come up with the report_delay parameter for our Pandora client which he kindly put into the cs_scheduler.cpp for me even though I was the only team member running MW at 100% at that time.


Hmmmm that does present a problem



I had resources set to 0 for the Milkyway project as I run Einstein at %100 on my single Linux system. As Keith mentioned, the share needs to be set to %100 to trigger the problem and be able to see if the mod works properly. AFAICT the Linux version works. The executable can be downloaded or can be built. It does not do anything special over what the 7.15 mod does, but it does work with the 7.20 Berkeley manager AFAICT.
ID: 74909 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,947,576
RAC: 22,161
Message 75029 - Posted: 7 Feb 2023, 10:48:15 UTC - in response to Message 74909.  

[quote]For what it is worth, I updated the Milkyway fix "mod" using BOINC version 7.20.2
https://github.com/JStateson/Milkyway-7-21
If you are running 7.15.0 there is no Milkyway advantage other than having 7.20.2 functionality instead of 7.14
Any problems post as an issue over at github.


Per chance does your 'fix' include the Lunatics 'fix' for Einstein as well, or is yours solely focused on Milkyway?

Not sure what is broken at Einstein that needs a "fix" Mikey.

Can you elucidate the problem?


Poor choice of words, the speed-up would have been a better choice I think

I will try and compile Joseph's Linux BOINC client and drop it into the Lunatics AIO version of BOINC on a spare PC for beta testing his MW "fix"[/quotd]

WOO HOO!!

Only problem is all I have are Nvidia cards and he mentions they aren't fast enough to trigger the problem at MW.

But I used to run into the MW Scheduler problem all the time back when I was running MW at 100% resource share. Why I asked our dev to come up with the report_delay parameter for our Pandora client which he kindly put into the cs_scheduler.cpp for me even though I was the only team member running MW at 100% at that time.


Hmmmm that does present a problem



I had resources set to 0 for the Milkyway project as I run Einstein at %100 on my single Linux system. As Keith mentioned, the share needs to be set to %100 to trigger the problem and be able to see if the mod works properly. AFAICT the Linux version works. The executable can be downloaded or can be built. It does not do anything special over what the 7.15 mod does, but it does work with the 7.20 Berkeley manager AFAICT.


Thank you very much!!
ID: 75029 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 · 2

Questions and Answers : Windows : GPU in non-continuous operation.

©2024 Astroinformatics Group