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: 30
Posts: 30   Pages: 3   [ Previous Page | 1 2 3 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 14758 times and has 29 replies Next Thread
Ian-n-Steve C.
Senior Cruncher
United States
Joined: May 15, 2020
Post Count: 180
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

IMO, true invalid results should not receive credit. you shouldn't be rewarded for providing incorrect results, regardless of time spent.

now if we're talking false invalids that are really still useful but only marked invalid for the validator being too strict, then I'm on the fence, it's up to the project to make the call. but a truely wrong result? no credit should be given. it promotes types of cheating where someone manipulates the application or calculation to make it run faster, but providing incorrect results. I agree with Richard, that tie-breakers should be employed here to see which is the true result. and they really need to allow cross validation between device types. having nvidia/nvidia, amd/amd, intel/intel always paired up is asking for trouble with application or device type that produces consistently invalid results, but they match so that validator would mark them both correct.
----------------------------------------

EPYC 7V12 / [5] RTX A4000
EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti
EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060
[2] EPYC 7642 / [2] RTX 2080Ti
[Apr 24, 2021 9:52:29 PM]   Link   Report threatening or abusive post: please login first  Go to top 
nanoprobe
Master Cruncher
Classified
Joined: Aug 29, 2008
Post Count: 2998
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

IMO, true invalid results should not receive credit. you shouldn't be rewarded for providing incorrect results, regardless of time spent.

now if we're talking false invalids that are really still useful but only marked invalid for the validator being too strict, then I'm on the fence, it's up to the project to make the call. but a truely wrong result? no credit should be given. it promotes types of cheating where someone manipulates the application or calculation to make it run faster, but providing incorrect results. I agree with Richard, that tie-breakers should be employed here to see which is the true result. and they really need to allow cross validation between device types. having nvidia/nvidia, amd/amd, intel/intel always paired up is asking for trouble with application or device type that produces consistently invalid results, but they match so that validator would mark them both correct.

We'll have to agree to disagree. If your resources are used to run the task until it's finished there should be some sort of thank you. Even an invalid task gives the researchers info. It's been done here in the past and I see no reason it can't be done again, especially with the amount of invalids for this project.
----------------------------------------
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.


----------------------------------------
[Edit 1 times, last edit by nanoprobe at Apr 24, 2021 9:59:40 PM]
[Apr 24, 2021 9:57:04 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Dayle Diamond
Senior Cruncher
Joined: Jan 31, 2013
Post Count: 450
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

Invalid results should not get credits.
I don't want to be thanked with credits, I want a fairly accurate accounting of the valid results I've done that can be compared across WCG projects.

Users who do not have an automated script running to request high-credit tasks every 2 minutes are barely impacted by running an invalid work unit for several minutes once every few days.
[Apr 24, 2021 10:16:45 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Ian-n-Steve C.
Senior Cruncher
United States
Joined: May 15, 2020
Post Count: 180
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

You don’t need a “thank you” for providing incorrect results. That’s an entitled attitude. No other project pays credit for invalids.
----------------------------------------

EPYC 7V12 / [5] RTX A4000
EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti
EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060
[2] EPYC 7642 / [2] RTX 2080Ti
[Apr 24, 2021 11:49:49 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Mike.Gibson
Ace Cruncher
England
Joined: Aug 23, 2007
Post Count: 12132
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

There is a big difference between an Invalid where you are the only one to return the unit as Invalid and the many case where all the copies are returned as Invalid. In the first scenario, you are deemed to be at fault and you should not get credits. In the second scenario, you should not be deemed to be at fault and should get credits.

Mike
[Apr 25, 2021 12:14:33 AM]   Link   Report threatening or abusive post: please login first  Go to top 
nanoprobe
Master Cruncher
Classified
Joined: Aug 29, 2008
Post Count: 2998
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

You don’t need a “thank you” for providing incorrect results. That’s an entitled attitude. No other project pays credit for invalids.

Entitled attitude? Really? Thank you Dr. Phil for giving me your attitude that I have an attitude. I never knew. And like I said, they have granted credits for invalid returns here before and considering that they are running about 10% ATM ............ The floor is all yours. I'm done.
----------------------------------------
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.


[Apr 25, 2021 12:43:58 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Ian-n-Steve C.
Senior Cruncher
United States
Joined: May 15, 2020
Post Count: 180
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

Well like I said. No other project pays for invalids. WCG shouldn’t either. If they decide they want to be nice and give some credits, more power to them. But they shouldn’t in any way be pressured or guilted into it by the users who think their time is more important than correct results.
----------------------------------------

EPYC 7V12 / [5] RTX A4000
EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti
EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060
[2] EPYC 7642 / [2] RTX 2080Ti
[Apr 25, 2021 1:06:55 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Dayle Diamond
Senior Cruncher
Joined: Jan 31, 2013
Post Count: 450
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

Nano, your signature line is substantial.
I'm sure you wouldn't like if it people thought it reflected the amount of errors you submitted.

If you want a 'thank you' for submitting invalid tasks, here it is.
Thank you.
[Apr 25, 2021 2:58:09 AM]   Link   Report threatening or abusive post: please login first  Go to top 
adriverhoef
Master Cruncher
The Netherlands
Joined: Apr 3, 2009
Post Count: 2085
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

Invalid tasks have been here in the past, too. It's nothing new.
See this post from 2013 by knreed, e.g., to quote:
2) The second change is that if your result is invalid, then the max credit you get is 50% of the computed credit or your claimed credit, whichever is less.


Even when OPNG tasks aren't included, there are cases nowadays:
Pages: 2
workunit 627670627
Project Name: OpenPandemics - COVID 19
Created: 04/19/2021 20:24:21
Name: OPN1_0042653_01142
Minimum Quorum: 2
Replication: 3
OPN1_0042653_01142_3-- Linux 717 Server Aborted 4/25/21 07:00:08 4/25/21 14:22:50 0.00 0.0 / 0.0
OPN1_0042653_01142_2-- Linux Ubuntu 717 Valid 4/25/21 07:00:06 4/25/21 11:29:33 1.28 68.2 / 72.7
OPN1_0042653_01142_1-- Linux openSU 717 Valid 4/21/21 20:08:03 4/25/21 06:59:30 2.11 77.2 / 72.7
OPN1_0042653_01142_0-- Red Hat Ente 717 Invalid 4/21/21 20:07:14 4/24/21 16:59:06 3.63 67.0 / 67.0

workunit 627670620
Project Name: OpenPandemics - COVID 19
Created: 04/19/2021 20:24:21
Name: OPN1_0042653_01043
Minimum Quorum: 2
Replication: 3
OPN1_0042653_01043_3-- Linux openSU 717 Server Aborted 4/25/21 07:00:08 4/25/21 11:16:26 0.00 0.0 / 0.0
OPN1_0042653_01043_2-- Linux Ubuntu 717 Valid 4/25/21 07:00:06 4/25/21 09:47:24 1.39 73.5 / 76.1
OPN1_0042653_01043_1-- Linux openSU 717 Valid 4/21/21 20:08:03 4/25/21 06:59:30 2.19 78.7 / 76.1
OPN1_0042653_01043_0-- Red Hat Ente 717 Invalid 4/21/21 20:07:14 4/24/21 16:59:06 3.79 69.8 / 69.8

workunit 627672005
Project Name: OpenPandemics - COVID 19
Created: 04/19/2021 20:24:26
Name: OPN1_0042716_00458
Minimum Quorum: 2
Replication: 3
OPN1_0042716_00458_3-- Linux Ubuntu 717 Valid 4/25/21 07:00:06 4/25/21 09:47:24 1.40 73.9 / 76.0
OPN1_0042716_00458_2-- Linux Fedora 717 Server Aborted 4/25/21 07:00:05 4/25/21 09:52:14 0.00 0.0 / 0.0
OPN1_0042716_00458_1-- Linux openSU 717 Valid 4/21/21 20:08:03 4/25/21 06:59:30 2.20 78.1 / 76.0
OPN1_0042716_00458_0-- Red Hat Ente 717 Invalid 4/21/21 20:07:14 4/24/21 16:59:06 3.86 71.1 / 71.1

----------------------------------------
[Edit 4 times, last edit by adriverhoef at Apr 25, 2021 4:47:15 PM]
[Apr 25, 2021 11:03:04 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Richard Haselgrove
Senior Cruncher
United Kingdom
Joined: Feb 19, 2021
Post Count: 360
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Advanced test batches

I think this is the longest-running task I've seen yet: 16:30 on half a GTX 1660 Ti under Linux - OPNG_0013362_00034
[Apr 25, 2021 7:27:20 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 30   Pages: 3   [ Previous Page | 1 2 3 | Next Page ]
[ Jump to Last Post ]
Post new Thread