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: 169
Posts: 169   Pages: 17   [ Previous Page | 7 8 9 10 11 12 13 14 15 16 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 137155 times and has 168 replies Next Thread
Sightus@CAU
Cruncher
Joined: May 7, 2013
Post Count: 9
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

Damn...seems the 20yrs badge is beyond my reach.
----------------------------------------

[Jun 11, 2015 3:00:17 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: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

For those who do not like to be 'trollesque' pointed at as cheats [including those using free cloud servers ;o], revert to any client of the 6 series ** and you'll have unrestricted work buffering [well, unrestricted to the point where the client computes work wont finish in time], at least no +1000 restriction. If you're still using 5.10.45 on domain controllers, that's an even better choice to use on 'your' PC. (Went overboard on the Linux with the ncpus, and still only have 7 days worth, so might switch that box to Windows, to get the full 10. laughing

Bite me, I'm doing it, because I can tongue

Enjoy the day

P.S. Those running BOINC on Windows DCs, soon there's a 7 version that will install on these devices, the note implying, you could not be running the client as a service.

"Rom Walton [Mon, 11 May 2015 17:18:15 +0000]

WINBUILD: Remove the restriction that prevented BOINC from installing on domain controllers.

TODO: Block service install in the advanced settings dialog
"

well, u keep getting more WUs for yourselves...only for your personal gain - badge! & u r not sure if u'll get the science back in time - 'cause of the overload!
maybe that is NOT a cheat...but it's giving:
- WCG more job to do, 'cause science will be resend after ur not finishing in time!
- crunchers more jobs, to do after ur messing up with scripts...
- (most importantly) prolong the finish of the science part!
certainly DOESN'T look MORAL from this side!

& about CLOUD COMPUTING - it's a FREE trial...u can also use it for a month! why didn't u?

main difference between CLOUD COMPUTING & your hack with scripts - CLOUD COMPUTING works normaly & returns DATA on time...(exept when it's shut down on due date!)
cool

p.s. btw, u can run Linux with BOINC...so no problems with Win DC! wink

Not a single job has yet missed deadline, and they won't because v7 clients will auto-remove any single task that wont make it. And, all were single / zero redundant, because that's what the devices are rated for! As the climate research lady professor from Texas said, the boss gave us a brain, but not all are using with what they're equipped with ;>)
[Jun 11, 2015 4:23:51 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: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

" u r not sure if u'll get the science back in time - 'cause of the overload!" Not sure what facts that statement was based on but this is what I do know. ~85% of my hosts will be dry by 68 hours after that last WU is sent to me. The remaining will be dry by 102 hours. Well within the 10 day limit. Absolutely, positively, beyond a shadow of a doubt, ALL 13248 WUs will be returned before the 10 day limit. They only caveat to the above statement is if they increase the run times per unit but even at that, these hosts are managed daily.

As far as the cloud computing goes, I remember back last year when crunchers were using the Amazon images and then they were being pull out from under them while they were in use causing a disruption to the normal flow of WUs. There was a point I was computing as much as 35% to 50% resends.

"only for your personal gain - badge!" So what?, he asks. It gets the science done and isn't that why they were implemented? Motivation!
[Jun 11, 2015 5:10:03 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: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

The buffer limit is 10 days deep, and that is the deadline on all the tasks received for FAHV on my hosts, the repair tasks skipping forward in queue anyway, as the client is designed to do when the buffer is greater than 1.75 days [for FAHV]. All open and above board, not causing a single extra duplicate when not needed. As the techs indicated, the closure after last of the new has been submitted takes 2-3 weeks i.e. it's not an issue to them or anybody on the receiving side of of the data, the scientists for whom we're doing it.

As the saying goes, let's not waste further breath [on 'justifying'].
[Jun 11, 2015 6:16:03 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Eric_Kaiser
Veteran Cruncher
Germany (Hessen)
Joined: May 7, 2013
Post Count: 1047
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

Wasn't able to get more than ~1300 wu on my 12 core with ncpu=64. This equals a buffer for 3 days. That's even sufficiant for resends.
----------------------------------------

[Jun 11, 2015 7:17:10 PM]   Link   Report threatening or abusive post: please login first  Go to top 
NixChix
Veteran Cruncher
United States
Joined: Apr 29, 2007
Post Count: 1187
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

...revert to any client of the 6 series ** and you'll have unrestricted work buffering [well, unrestricted to the point where the client computes work wont finish in time], at least no +1000 restriction. If you're still using 5.10.45 on domain controllers, that's an even better choice to use on 'your' PC. (Went overboard on the Linux with the ncpus, and still only have 7 days worth, so might switch that box to Windows, to get the full 10. laughing

I still have 1 client running 6.10.58 and it too is limited to 35 WU per core. crying

Cheers coffee
----------------------------------------

[Jun 12, 2015 5:46:34 AM]   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: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

Nix, it's one of 2 limits, the server controlled 35 per core 'seen' by the server [overarching rule], or the +1000, which 6.10 does not have as client side restriction [with the right buttons pushed thru the cc_config.xml, if needed ;>]. Why Eric get's 1300 on one device is way beyond my comprehension. Wild guess, there's some wonky multiplicator in the work requested / needed, with GPGPU's somehow funking into the equation, maybe something in app_config????????????? (or he's using a pre-1000 restriction rule client ;?)
----------------------------------------
[Edit 1 times, last edit by Former Member at Jun 12, 2015 7:32:50 AM]
[Jun 12, 2015 7:25:33 AM]   Link   Report threatening or abusive post: please login first  Go to top 
KLiK
Master Cruncher
Croatia
Joined: Nov 13, 2006
Post Count: 3108
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

For those who do not like to be 'trollesque' pointed at as cheats [including those using free cloud servers ;o], revert to any client of the 6 series ** and you'll have unrestricted work buffering [well, unrestricted to the point where the client computes work wont finish in time], at least no +1000 restriction. If you're still using 5.10.45 on domain controllers, that's an even better choice to use on 'your' PC. (Went overboard on the Linux with the ncpus, and still only have 7 days worth, so might switch that box to Windows, to get the full 10. laughing

Bite me, I'm doing it, because I can tongue

Enjoy the day

P.S. Those running BOINC on Windows DCs, soon there's a 7 version that will install on these devices, the note implying, you could not be running the client as a service.

"Rom Walton [Mon, 11 May 2015 17:18:15 +0000]

WINBUILD: Remove the restriction that prevented BOINC from installing on domain controllers.

TODO: Block service install in the advanced settings dialog
"

well, u keep getting more WUs for yourselves...only for your personal gain - badge! & u r not sure if u'll get the science back in time - 'cause of the overload!
maybe that is NOT a cheat...but it's giving:
- WCG more job to do, 'cause science will be resend after ur not finishing in time!
- crunchers more jobs, to do after ur messing up with scripts...
- (most importantly) prolong the finish of the science part!
certainly DOESN'T look MORAL from this side!

& about CLOUD COMPUTING - it's a FREE trial...u can also use it for a month! why didn't u?

main difference between CLOUD COMPUTING & your hack with scripts - CLOUD COMPUTING works normaly & returns DATA on time...(exept when it's shut down on due date!)
cool

p.s. btw, u can run Linux with BOINC...so no problems with Win DC! wink

Not a single job has yet missed deadline, and they won't because v7 clients will auto-remove any single task that wont make it. And, all were single / zero redundant, because that's what the devices are rated for! As the climate research lady professor from Texas said, the boss gave us a brain, but not all are using with what they're equipped with ;>)


The main thing is that you can't see how IMMORAL & UNETHICS this kind of your behavior is! & what it will bring with others who read your posts...

So yes, it's a true shame when someone is not using it's brain in a positive way! Only selfishness resides in you...& that's so sad! Ur blinded by it!
cool
----------------------------------------
oldies:UDgrid.org & PS3 Life@home


non-profit org. Play4Life in Zagreb, Croatia
[Jun 12, 2015 8:32:59 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Eric_Kaiser
Veteran Cruncher
Germany (Hessen)
Joined: May 7, 2013
Post Count: 1047
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

No, all Linux and Windows computers are running 7.2.42 client.
The 4 servers and my Windows PC (all 12 cores, but ncpu set to 64) received up to ~1300 wu each.
The two amd 4 core were levelled out to a 8 day buffer (~800 wu).
In total I had ~9000 wu in buffer (5x1300=7500, 2x800=1600, rest Android).
ncpu was not applied to androids.
app_config was used to restrict to the physical available cores.
----------------------------------------

[Jun 12, 2015 8:35:01 AM]   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: FA@H Progress Tracking: 99%, 29 APR 2015; finish estimate: mid JUN 2015

For those who do not like to be 'trollesque' pointed at as cheats [including those using free cloud servers ;o], revert to any client of the 6 series ** and you'll have unrestricted work buffering [well, unrestricted to the point where the client computes work wont finish in time], at least no +1000 restriction. If you're still using 5.10.45 on domain controllers, that's an even better choice to use on 'your' PC. (Went overboard on the Linux with the ncpus, and still only have 7 days worth, so might switch that box to Windows, to get the full 10. laughing

Bite me, I'm doing it, because I can tongue

Enjoy the day

P.S. Those running BOINC on Windows DCs, soon there's a 7 version that will install on these devices, the note implying, you could not be running the client as a service.

"Rom Walton [Mon, 11 May 2015 17:18:15 +0000]

WINBUILD: Remove the restriction that prevented BOINC from installing on domain controllers.

TODO: Block service install in the advanced settings dialog
"

well, u keep getting more WUs for yourselves...only for your personal gain - badge! & u r not sure if u'll get the science back in time - 'cause of the overload!
maybe that is NOT a cheat...but it's giving:
- WCG more job to do, 'cause science will be resend after ur not finishing in time!
- crunchers more jobs, to do after ur messing up with scripts...
- (most importantly) prolong the finish of the science part!
certainly DOESN'T look MORAL from this side!

& about CLOUD COMPUTING - it's a FREE trial...u can also use it for a month! why didn't u?

main difference between CLOUD COMPUTING & your hack with scripts - CLOUD COMPUTING works normaly & returns DATA on time...(exept when it's shut down on due date!)
cool

p.s. btw, u can run Linux with BOINC...so no problems with Win DC! wink

Not a single job has yet missed deadline, and they won't because v7 clients will auto-remove any single task that wont make it. And, all were single / zero redundant, because that's what the devices are rated for! As the climate research lady professor from Texas said, the boss gave us a brain, but not all are using with what they're equipped with ;>)


The main thing is that you can't see how IMMORAL & UNETHICS this kind of your behavior is! & what it will bring with others who read your posts...

So yes, it's a true shame when someone is not using it's brain in a positive way! Only selfishness resides in you...& that's so sad! Ur blinded by it!
cool

laughing
[Jun 12, 2015 8:48:28 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 169   Pages: 17   [ Previous Page | 7 8 9 10 11 12 13 14 15 16 | Next Page ]
[ Jump to Last Post ]
Post new Thread