474 new structures
log in

Advanced search

Message boards : News : 474 new structures

Author Message
Vlad
Project administrator
Project developer
Project tester
Project scientist
Help desk expert
Send message
Joined: 26 Oct 16
Posts: 303
Credit: 103,382
RAC: 0
Message 681 - Posted: 2 Feb 2019, 18:32:48 UTC
Last modified: 2 Feb 2019, 20:22:57 UTC

New COD entries are added.

Update. All WUs are sent to the hosts.

Profile [AF>Amis des Lapins] Phil1966
Send message
Joined: 12 Nov 16
Posts: 6
Credit: 33,583,439
RAC: 7,413
Message 682 - Posted: 2 Feb 2019, 18:41:56 UTC - in response to Message 681.

Very difficult to get GPU WU's.

Vlad
Project administrator
Project developer
Project tester
Project scientist
Help desk expert
Send message
Joined: 26 Oct 16
Posts: 303
Credit: 103,382
RAC: 0
Message 685 - Posted: 2 Feb 2019, 20:16:41 UTC - in response to Message 682.

Very difficult to get GPU WU's.

The fraction of GPU WUs among all WUs didn't changed since the last run.

bluestang
Send message
Joined: 22 May 18
Posts: 5
Credit: 3,132,419
RAC: 3,349
Message 686 - Posted: 3 Feb 2019, 1:03:20 UTC

What about the maximum amount of allowed GPU WUs a computer is allowed to have In Progress at once? Has that changed? It's like I can only have 25 WUs In Progress at a time an each PC I have. It's never been like that.

Vlad
Project administrator
Project developer
Project tester
Project scientist
Help desk expert
Send message
Joined: 26 Oct 16
Posts: 303
Credit: 103,382
RAC: 0
Message 688 - Posted: 3 Feb 2019, 10:55:24 UTC - in response to Message 686.

What about the maximum amount of allowed GPU WUs a computer is allowed to have In Progress at once? Has that changed? It's like I can only have 25 WUs In Progress at a time an each PC I have. It's never been like that.

Yes, it’s 25 per host since September. It was 25 per GPU before that but I had to switch to per host limit because some hosts reported that they had 5 or more GPUs installed to get more WUs. With the per host limit, the WUs are distributed more evenly among the hosts.

bluestang
Send message
Joined: 22 May 18
Posts: 5
Credit: 3,132,419
RAC: 3,349
Message 689 - Posted: 3 Feb 2019, 18:52:37 UTC - in response to Message 688.

What about the maximum amount of allowed GPU WUs a computer is allowed to have In Progress at once? Has that changed? It's like I can only have 25 WUs In Progress at a time an each PC I have. It's never been like that.

Yes, it’s 25 per host since September. It was 25 per GPU before that but I had to switch to per host limit because some hosts reported that they had 5 or more GPUs installed to get more WUs. With the per host limit, the WUs are distributed more evenly among the hosts.


Well, some of us do have 5 GPUs per host or more. Old mining rigs that I've acquired to crunch GPU projects instead of crappy mining are used. So 25 per GPU was nice.

Vlad
Project administrator
Project developer
Project tester
Project scientist
Help desk expert
Send message
Joined: 26 Oct 16
Posts: 303
Credit: 103,382
RAC: 0
Message 691 - Posted: 3 Feb 2019, 19:27:40 UTC - in response to Message 689.

Well, some of us do have 5 GPUs per host or more. Old mining rigs that I've acquired to crunch GPU projects instead of crappy mining are used. So 25 per GPU was nice.

Some hosts report 50 GPUs, http://xansons4cod.com/xansons4cod/show_host_detail.php?hostid=2356...
Of course, setting per GPU limit would not be a problem if the total number of WUs was not so limited.

bluestang
Send message
Joined: 22 May 18
Posts: 5
Credit: 3,132,419
RAC: 3,349
Message 692 - Posted: 3 Feb 2019, 21:13:50 UTC - in response to Message 691.

Yeah, those are evidently fake and gaming the system. I see your predicament with limited WUs. Thanks for the info.

Profile marsinph
Send message
Joined: 20 Apr 18
Posts: 23
Credit: 1,446,819
RAC: 6,015
Message 695 - Posted: 9 Feb 2019, 20:48:44 UTC - in response to Message 692.

Yeah, those are evidently fake and gaming the system. I see your predicament with limited WUs. Thanks for the info.



Hello Bluestang
I can understand your frustration.
But I also fully defend Vlad.
Not forget Xanson have very few WU to crunch.
It is known from the start of the project.
Very clear.
The limitations give everyone a possibility to crunch.
Very few, I agree. But it is the same for everyone around the world.
After very fast calculation, with my weak host, I am able to crunch all WU
in only two days.
Yes, midrange host, can crunch all WU in only two days.

Before you accuse Vlad to fake and gaming, I suggest you read again the rules of this project !

Imagine instead of my weak hosts, I come with tenth of 64cores, all with 4x108Ti or better, All wu will be returned in a few minuts. And nothing for other. OK so far. For the project it is better.
Scientist are happy to receive results. It is the main goal.
The research increase. Also OK

But it is not the phylosophy of this project.

Vlad, try to give everyone a chance to crunch. Always very few by user/host.
I fully understand, agree and support his way of thinking. Everyone the same.

I personnally very like this project. I am also like you frustrated to not receive more than 25 WU.
But it are the rules. And Project admin have very clear explain why.

Then before complaints, you processed 2077 COD for a total credit of 2,781,024 CR (Situation on 09feb20:39CET

Personnaly I crunched 3797 COD for only 1,042,915CR (situation time the same) !

So please do not do not accuse admin (vlad) about "fake" or "gaming".

Personnaly, (but it is only me), I think he try to do his best for everyone.
Scientist are happy. The main goal.

Ranking .... is not important. Only for some who give more interest to ranking than to research.

So Bluestang, please reconsider your opinion (vision). For sure againt the admin
Best regards
____________

Profile marsinph
Send message
Joined: 20 Apr 18
Posts: 23
Credit: 1,446,819
RAC: 6,015
Message 696 - Posted: 16 Feb 2019, 17:26:44 UTC - in response to Message 691.

Well, some of us do have 5 GPUs per host or more. Old mining rigs that I've acquired to crunch GPU projects instead of crappy mining are used. So 25 per GPU was nice.

Some hosts report 50 GPUs, http://xansons4cod.com/xansons4cod/show_host_detail.php?hostid=2356...
Of course, setting per GPU limit would not be a problem if the total number of WUs was not so limited.



Well well well.
50 GPU on the same host !
And tyhis host is managed by a i7-2600K !?!? The same as 3 from my hosts.
And this host have onlmy 8Gb RAM (also like me)The only possible is to run on several virtual machine.
But with only 8Gb RAM I am not sure to be able to run more than 4 VM together.
Then the "little" 2600K will not be able to follow and manage it

I want so mainboard able to manage 50 GPU !!!
And eventually the speed of the FSB (not "russian secret service" his new name of course) Smile.


I really would known the manufacturer and model of this mainboard able to manage 50 GPU. As far as I know, the maximum is four GPU SLI or Crossfire.
If it is a server, OK, but then 2600K can NOT manage and Win10 also not.
Where to buy ???

xii5ku
Send message
Joined: 10 Mar 18
Posts: 6
Credit: 23,692,317
RAC: 3,186
Message 700 - Posted: 16 Feb 2019, 19:32:21 UTC
Last modified: 16 Feb 2019, 19:39:07 UTC

VLad wrote:
Some hosts report 50 GPUs, http://xansons4cod.com/xansons4cod/show_host_detail.php?hostid=2356...

bluestang wrote:
Yeah, those are evidently fake and gaming the system.

A system with 50 logical AMD RX series GPUs is of course nonsense.

But if you spot a host with many big Nvidia GPUs, then a likely explanation for this is that its owner is also SETI@home participant and configured this pseudo number of GPUs in order to bridge SETI's weekly maintenance window. The people who do this don't want to "game the system", they merely want highest possible uptime of their compute clients. This directly benefits the SETI@home project, and I applaud these users for their dedication and ability.

Anyway; in case of XANSONS for COD with its low number of tasks per month, I can understand Vlad's choice for a simple per-host limit, instead of anything more complicated.

--------
Edit:
marsinph wrote:
So Bluestang, please reconsider your opinion

Bluestang disapproved of the user with the abnormal host config, not of Vlad's countermeasure against hosts like this.

Profile marsinph
Send message
Joined: 20 Apr 18
Posts: 23
Credit: 1,446,819
RAC: 6,015
Message 701 - Posted: 16 Feb 2019, 19:54:36 UTC - in response to Message 700.

VLad wrote:
Some hosts report 50 GPUs, http://xansons4cod.com/xansons4cod/show_host_detail.php?hostid=2356...

bluestang wrote:
Yeah, those are evidently fake and gaming the system.

A system with 50 logical AMD RX series GPUs is of course nonsense.

But if you spot a host with many big Nvidia GPUs, then a likely explanation for this is that its owner is also SETI@home participant and configured this pseudo number of GPUs in order to bridge SETI's weekly maintenance window. The people who do this don't want to "game the system", they merely want highest possible uptime of their compute clients. This directly benefits the SETI@home project, and I applaud these users for their dedication and ability.

Anyway; in case of XANSONS for COD with its low number of tasks per month, I can understand Vlad's choice for a simple per-host limit, instead of anything more complicated.



Hello xii5ku,
I see (read) I am not the only who agree with Vlad his way of doing.
Then you comparaison with SETI.
I am from past century on SETI. Everyone know the weekly outage.
Sometimes very long.
If I good remember in the latest 20 years on SETI, there were two main outage period. (I not count just after Arecibo "crash".

You applaude for such invest. OK if it is honnest.

So why to artificial mount 50 GPU. Like you say : nonsense.
It is abusing the system.

Look on SETI : host 8480062 !!!
Great : on one host 48 (fourty-eight) GTX1080Ti !!!
And running on a i7-6850K !!!
OK, Linux.
It only show virtualization.

I not write. It is not the place.

I only fully agree with your opinion about Xanson and the decision of Vlad

Best regards






But, evryone is free to download WU up to 10 days (on SETI)

Profile marsinph
Send message
Joined: 20 Apr 18
Posts: 23
Credit: 1,446,819
RAC: 6,015
Message 702 - Posted: 16 Feb 2019, 19:55:07 UTC - in response to Message 700.

VLad wrote:
Some hosts report 50 GPUs, http://xansons4cod.com/xansons4cod/show_host_detail.php?hostid=2356...

bluestang wrote:
Yeah, those are evidently fake and gaming the system.

A system with 50 logical AMD RX series GPUs is of course nonsense.

But if you spot a host with many big Nvidia GPUs, then a likely explanation for this is that its owner is also SETI@home participant and configured this pseudo number of GPUs in order to bridge SETI's weekly maintenance window. The people who do this don't want to "game the system", they merely want highest possible uptime of their compute clients. This directly benefits the SETI@home project, and I applaud these users for their dedication and ability.

Anyway; in case of XANSONS for COD with its low number of tasks per month, I can understand Vlad's choice for a simple per-host limit, instead of anything more complicated.



Hello xii5ku,
I see (read) I am not the only who agree with Vlad his way of doing.
Then you comparaison with SETI.
I am from past century on SETI. Everyone know the weekly outage.
Sometimes very long.
If I good remember in the latest 20 years on SETI, there were two main outage period. (I not count just after Arecibo "crash".

You applaude for such invest. OK if it is honnest.

So why to artificial mount 50 GPU. Like you say : nonsense.
It is abusing the system.

Look on SETI : host 8480062 !!!
Great : on one host 48 (fourty-eight) GTX1080Ti !!!
And running on a i7-6850K !!!
OK, Linux.
It only show virtualization.

I not write. It is not the place.

I only fully agree with your opinion about Xanson and the decision of Vlad

Best regards






But, evryone is free to download WU up to 10 days (on SETI)

xii5ku
Send message
Joined: 10 Mar 18
Posts: 6
Credit: 23,692,317
RAC: 3,186
Message 706 - Posted: 16 Feb 2019, 20:45:38 UTC
Last modified: 16 Feb 2019, 20:47:38 UTC

Sorry for straying off-topic, but: SETI@home has a quota of 100 tasks in progress per GPU. This is far less than what a good mainstream GPU can complete during a typical S@h maintenance downtime.

motqalden
Send message
Joined: 27 Jan 19
Posts: 1
Credit: 1,968,709
RAC: 12,668
Message 707 - Posted: 17 Feb 2019, 0:59:24 UTC

This "fake" user who is "abusing the system" has completed almost 10 million points. Bravo to this guy for such dedication. If the program allow you to do it its not cheating / gaming or anything like that. If this shouldn't be allowed then the project software itself shouldn't allow it. But really this / these users are getting tons of work done so good on them as far as i'm concerned.

Message boards : News : 474 new structures


Main page · Your account · Message boards


© 2019 Vladislav Neverov (NRC 'Kurchatov institute'), Nikolay Khrapov (Institute for Information Transmission Problems of RAS)