Advanced search

Message boards : Number crunching : WTF

Author Message
Joe
Send message
Joined: 24 Nov 11
Posts: 3
Credit: 258,543,005
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwat
Message 42632 - Posted: 16 Jan 2016 | 9:06:53 UTC

I'm relatively new at this but have to say I'm very disappointed.

The server passes out work units that are never received by the host, thousands upon thousands of Linux work units fail for days on end and when there is finally a so called fix most of my machines lock up necessitating recovery and lost real world data.

I'm sorry this is not in any stretch of someones imagination "Real Science" its simply fools passing themselves of as "Scientists"

mikey
Send message
Joined: 2 Jan 09
Posts: 290
Credit: 2,035,466,115
RAC: 10,289,559
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 42633 - Posted: 16 Jan 2016 | 11:54:49 UTC - in response to Message 42632.

I'm relatively new at this but have to say I'm very disappointed.

The server passes out work units that are never received by the host, thousands upon thousands of Linux work units fail for days on end and when there is finally a so called fix most of my machines lock up necessitating recovery and lost real world data.

I'm sorry this is not in any stretch of someones imagination "Real Science" its simply fools passing themselves of as "Scientists"


I don't think your experience is normal, if you ask for some help, you would have to 'unhide' your computers first though, click on my name to see my pc's and see an example of what I could see if you did it, then perhaps some people here could help figure out what the problems is.

Dayle Diamond
Send message
Joined: 5 Dec 12
Posts: 84
Credit: 1,629,213,415
RAC: 672,941
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 42636 - Posted: 16 Jan 2016 | 17:22:50 UTC

YES, thousands and thousands of Linux work units were failing for days. That's shameful.

Profile caffeineyellow5
Avatar
Send message
Joined: 30 Jul 14
Posts: 225
Credit: 2,658,976,345
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwat
Message 42640 - Posted: 16 Jan 2016 | 20:49:24 UTC - in response to Message 42636.

No project has zero failed tasks, whether they report back to you that some failed or not. Furthermore, the more difficult the work, the more difficult it is to run. The more diverse the computer/GPU configuration a task can be run on means the more difficult it is to create a WU that will run to full time. Overclocking cars, even in cards called "super clocked" or otherwise that you bought overclocked can also be a factor on any GPU or CPU causing an error on a WU on any project. Even tasks that just hand out text files to analyze a simple mathematical equation can error out on a distributed project depending on any configuration or number of problems. Processor jitter (which happens more often than admitted), memory leakage from other software installed along side of the project, or internet packet loss can also contribute to anything. So when for a few days one project handing out WUs that mean so much and are this difficult to make (3-d simulated DNA involved interactions of chemicals and living cells)sends out a few WUs that are unstable or are not fully compatible with a few configurations of CPU/GPU, don't get bent out of shape. Even professionals in any field need to make small mistakes to learn how to eliminate all future ones. So just calm down, wait a few days, and try again, OR let the WUs fail and contribute more than you complain by letting the WUs teach the programmers how to better help the scientists. Your attitude is as if all life scientists have 20 years of experience with 3-d simulation of DNA structures and degrees in modern computing and everything related. Let them 'break a few eggs' so we can get breakfast, will you please?
____________
1 Corinthians 9:16 "For though I preach the gospel, I have nothing to glory of: for necessity is laid upon me; yea, woe is unto me, if I preach not the gospel!"
Ephesians 6:18-20, please ;-)
http://tbc-pa.org

Dayle Diamond
Send message
Joined: 5 Dec 12
Posts: 84
Credit: 1,629,213,415
RAC: 672,941
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 42642 - Posted: 16 Jan 2016 | 22:00:58 UTC - in response to Message 42640.

There's always someone on tech forums that takes things a little too personally. You don't get brownie points for defending the developers, hell or high water.

The thousands of work units weren't failing because of any of those random errors, anyway. And the project scientists had no comment for days.

I'm a donor. I bought graphics cards out of pocket to support this project, and if there's something I don't like, I'm going to say it.

No project has zero failed tasks, whether they report back to you that some failed or not. Furthermore, the more difficult the work, the more difficult it is to run. The more diverse the computer/GPU configuration a task can be run on means the more difficult it is to create a WU that will run to full time. Overclocking cars, even in cards called "super clocked" or otherwise that you bought overclocked can also be a factor on any GPU or CPU causing an error on a WU on any project. Even tasks that just hand out text files to analyze a simple mathematical equation can error out on a distributed project depending on any configuration or number of problems. Processor jitter (which happens more often than admitted), memory leakage from other software installed along side of the project, or internet packet loss can also contribute to anything. So when for a few days one project handing out WUs that mean so much and are this difficult to make (3-d simulated DNA involved interactions of chemicals and living cells)sends out a few WUs that are unstable or are not fully compatible with a few configurations of CPU/GPU, don't get bent out of shape. Even professionals in any field need to make small mistakes to learn how to eliminate all future ones. So just calm down, wait a few days, and try again, OR let the WUs fail and contribute more than you complain by letting the WUs teach the programmers how to better help the scientists. Your attitude is as if all life scientists have 20 years of experience with 3-d simulation of DNA structures and degrees in modern computing and everything related. Let them 'break a few eggs' so we can get breakfast, will you please?

fractal
Send message
Joined: 16 Aug 08
Posts: 87
Credit: 1,248,879,715
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 42645 - Posted: 16 Jan 2016 | 23:09:28 UTC

All projects have issues which get fixed. This project is no different.

This project is no different than any other University backed project in that things tend to be slow to fix during holidays.

The admin stated that the license for the code in the Linux application had expired resulting in 100% failures. This has nothing to do with overclocking. It is a configuration issue that is easy for someone with both the knowledge, access and the time to fix. The trick is finding the right "someone" when everyone is at home enjoying time with their families.

It took me a day or two to notice and move my machines to a different project. I am too stupid to figure out how to get backup projects to work reliably in Boinc for systems with multiple GPUs. Setting the priority to 0 doesn't do it. But, that's my problem and not theirs. I picked up 20 million cobbles over on another project before the license got updated. No biggie for me.

Post to thread

Message boards : Number crunching : WTF

//