BillR
Born Again Cynic
- Joined
- Feb 17, 2002
- Messages
- 18,535
In the past few weeks it would seem that we have all seen some interesting changes in the folding at home project. Changes are not uncommon with a project of this scope; in fact they are to be expected. As for myself I dont mind change, its part of day to day life and has a tendency to keep things interesting.
Even though we should expect change no logic in the world says you have to like it. Some changes are for the good and some not so good. Virtually the only time I have issues with change are when the changes I face are born of utter stupidity and are driven by someones hidden agenda or change brought about by total and complete incompetence with no thought what so ever given to the people involved.
Im not sure what prompted these new, rushed and untested code changes and no explanation was ever given why today or tomorrow suddenly became a hard deadline and cut off point. Once again a deadline given regardless of the consequences to the users or if you like members of all the teams.
In spite of all the above I have managed to change over both my SMP clients and GPU clients and even better I have been able to make them all stable and they are producing work.
So, your next question has to be, Bill, why point out the obvious?
Glad you asked actually. We were promised new higher output work and if not promised certainly this concept was pushed as a hint.
An hour ago one of my GPU clients, the new one and one that had already turned in two work units since install caught my eye. It seemed to be frozen in time. Stopped, dead, deceased etc. My first thought was heat and a computer lock up however the SMP window was open and happily chewing on a 2665. I stopped the GPU client and restarted it and as it should it came back to its checkpoint and I ignored it for a while. I looked at it again, about a half hour later and I was on frame 3. I checked my other clients and they were rolling along at a nice clip but not this one.
I like to keep things simple so rather then get upset with the machine I decided to check here:
http://fah-web.stanford.edu/psummary.html
This for those who dont know is Stanfords project summery page where we are able to see the specs of the given projects.
I call your attention to project p4109 and you will see it is valued at 213 points with a preferred deadline of only 39 days. The fact is the 4109 folds more slowly then an SMP 2665 and we are all aware of how productive those are.
A quick glance around the GPU stats will show you this is not the only one like this, there are several. Clearly these units have suffered the same fate as the new code etc and nobody could have ever checked this on a benchmark machine.
While I am a strong proponent of the points dont matter its the science idea it is obvious there is a serious problem here and once again, where is Stanfords accountability? Who is really in charge if anyone and to whom do they answer?
This falls squarely into the abused folder category and I will under no circumstances police my boxen every few hours to delete these ridiculous work units.
As of this moment its unclear to me what decisions I have to make but I know that we as a team and me as an individual care more about this project then the folks running the project.
Your opinions and ideas and thoughts, please.
Even though we should expect change no logic in the world says you have to like it. Some changes are for the good and some not so good. Virtually the only time I have issues with change are when the changes I face are born of utter stupidity and are driven by someones hidden agenda or change brought about by total and complete incompetence with no thought what so ever given to the people involved.
Im not sure what prompted these new, rushed and untested code changes and no explanation was ever given why today or tomorrow suddenly became a hard deadline and cut off point. Once again a deadline given regardless of the consequences to the users or if you like members of all the teams.
In spite of all the above I have managed to change over both my SMP clients and GPU clients and even better I have been able to make them all stable and they are producing work.
So, your next question has to be, Bill, why point out the obvious?
Glad you asked actually. We were promised new higher output work and if not promised certainly this concept was pushed as a hint.
An hour ago one of my GPU clients, the new one and one that had already turned in two work units since install caught my eye. It seemed to be frozen in time. Stopped, dead, deceased etc. My first thought was heat and a computer lock up however the SMP window was open and happily chewing on a 2665. I stopped the GPU client and restarted it and as it should it came back to its checkpoint and I ignored it for a while. I looked at it again, about a half hour later and I was on frame 3. I checked my other clients and they were rolling along at a nice clip but not this one.
I like to keep things simple so rather then get upset with the machine I decided to check here:
http://fah-web.stanford.edu/psummary.html
This for those who dont know is Stanfords project summery page where we are able to see the specs of the given projects.
I call your attention to project p4109 and you will see it is valued at 213 points with a preferred deadline of only 39 days. The fact is the 4109 folds more slowly then an SMP 2665 and we are all aware of how productive those are.
A quick glance around the GPU stats will show you this is not the only one like this, there are several. Clearly these units have suffered the same fate as the new code etc and nobody could have ever checked this on a benchmark machine.
While I am a strong proponent of the points dont matter its the science idea it is obvious there is a serious problem here and once again, where is Stanfords accountability? Who is really in charge if anyone and to whom do they answer?
This falls squarely into the abused folder category and I will under no circumstances police my boxen every few hours to delete these ridiculous work units.
As of this moment its unclear to me what decisions I have to make but I know that we as a team and me as an individual care more about this project then the folks running the project.
Your opinions and ideas and thoughts, please.