Forgive me if this has been asked before..

I an currently crunching for 4 projects. I have the resource share set to the default 200 for each project.

How does this equate to each projects actual time crunching? Is it better to have a larger number or a lower number?

Thanks..

CW it's just a ratio. nothing more or less.


200
200
200
200

is the same as

20
20
20
20

or

2
2
2
2

or

100
100
100
100

by the way, the default is 100

does this help?

tony

[edit] each is 1/4 or 25% for each project

CW it's just a ratio. nothing more or less.


200
200
200
200

is the same as

20
20
20
20

or

2
2
2
2

or

100
100
100
100

by the way, the default is 100

does this help?

tony

[edit] each is 1/4 or 25% for each project

Don't worry about it you probably will connect to too many projects and scheduler will ignore it to finish some of the projects really early. It will cause massive debt accumalation so A project you might want to run won't run for a long time. You probably then run out of work for the other projects and can't get any new work because you accumulated too much debt somewhere.

---

by the way, the default is 100

does this help?

tony

[edit] each is 1/4 or 25% for each project

That is what I figured TOny, but I wanted to make sure. In a discussion I saw in Einstain I saw people using 2500, 200, 1000 and I wanted to make sure what I concluded was correct..

Thanks..

PS Yes it is 100.. It has not been a good week for me so far.. Having to be stuffed into a big glass tude and forced to breath compressed pure O2 for 4 hours is not fun and it appears I may have to do it again..

Don't worry about it you probably will connect to too many projects and scheduler will ignore it to finish some of the projects really early. It will cause massive debt accumalation so A project you might want to run won't run for a long time. You probably then run out of work for the other projects and can't get any new work because you accumulated too much debt somewhere.

Actually, John's scheduler does an excellent job (given sufficient time and patience) in preventing all the mayhem you are alluding to in your rather silly commentary. I've seen John try to explain it to you patiently on several occasions and you seem intent to repay that patience with vitriol and pure misinformation. You are quite entitled to your opinion but you should clearly indicate that it is just an opinion obviously tainted by your own lack of understanding of what is really happening.

Why are you buying into a thread that is really nothing to do with your personal gripes with the scheduler?
The simple answer to the original question was that each project will get 25% of the available time in the longer term with short term variations from that 25% in order to attempt to prevent any one project from having work expire due to deadline issues.

---

Cheers,
Gary.

Hi

Yep, just give it time.

greetz Mike

---

With each crime and every kindness we birth our future.

Just curious if anybody noticed that I ignored ksnash's response???

---

I'd rather speak my mind because it hurts too much to bite my tongue.

American Spirit BBQ Proudly Serving those that courageously defend freedom.

Just curious if anybody noticed that I ignored ksnash's response???

I certainly did notice, and I was ignoring it as well.

---

BOINC WIKI

Just curious if anybody noticed that I ignored ksnash's response???

I certainly did notice, and I was ignoring it as well.

Good I was just trying to stop the flame that seemed to be growing :)

---

I'd rather speak my mind because it hurts too much to bite my tongue.

American Spirit BBQ Proudly Serving those that courageously defend freedom.

Just curious if anybody noticed that I ignored ksnash's response???

Sadly, in the early days of the internet, no one defined NCTP.

---

Just curious if anybody noticed that I ignored ksnash's response???

Sadly, in the early days of the internet, no one defined NCTP.

Let me guess... "Noise Control Trimming Protocol"? :-)

I thought of less gracious one as well, but... ;-)

Alinator

Just curious if anybody noticed that I ignored ksnash's response???

Sadly, in the early days of the internet, no one defined NCTP.

Let me guess... "Noise Control Trimming Protocol"? :-)

I thought of less gracious one as well, but... ;-)

Alinator

Just curious if anybody noticed that I ignored ksnash's response???

Sadly, in the early days of the internet, no one defined NCTP.

Let me guess... "Noise Control Trimming Protocol"? :-)

Network Clue Transport Protocol

---

Well let's see they did plan for Boinc based on established working models and they leave out a lot of useful utility. Scheduler uses bad data to go into overdrive too much and you have to a kludge against a kludge. You can barely get it to work right and you go around patting yourself on the back saying a correct form of scheduler can't be done. You apply a sledge haame when you need a hammer. a lot of people are saying the same thing and all you say is something that doesn't work works great. The emperor is not wearing any cloths.

---

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

---

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

Once again. It does EXACTLY what you list in the four points. Your definition of deadline imminent leaves out modem users. Once the CPU scheduler has determined that there is a problem, it is late enough that it has to dedicate the entire CPU the process (it is being optimistic that the process will work itself out until it is extremely obvious that it isn't). It does return to0 normal mode within one timeslice of when it can.

---

BOINC WIKI

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

Once again. It does EXACTLY what you list in the four points. Your definition of deadline imminent leaves out modem users. Once the CPU scheduler has determined that there is a problem, it is late enough that it has to dedicate the entire CPU the process (it is being optimistic that the process will work itself out until it is extremely obvious that it isn't). It does return to0 normal mode within one timeslice of when it can.

Want option to not be based on connection rate. connection rate is not the deadline.

---

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

I think he's rejecting your "facts" -- especially 3 and 4.

Your major objection is that if you have six work units, call them A, B, C, D, E and F, then they must be done in ABCDEF order.

The scheduler may do FCDBAE to make sure work is returned on time. Why does the order matter?

---

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

Once again. It does EXACTLY what you list in the four points. Your definition of deadline imminent leaves out modem users. Once the CPU scheduler has determined that there is a problem, it is late enough that it has to dedicate the entire CPU the process (it is being optimistic that the process will work itself out until it is extremely obvious that it isn't). It does return to0 normal mode within one timeslice of when it can.

Want option to not be based on connection rate. connection rate is not the deadline.

The more I read these threads, the more I think "panic mode" is a poor choice of terms.

We aren't talking about "panic" like you'd see in the typical Japanese monster movie when everyone in the city is running while Godzilla is kicking it to pieces.

What we're talking about is looking at the queue, and if there is some doubt about making schedules, doing the work in order by deadline so that it's ready for the next connection.

That is followed by a mechanism that works long-term to make sure that BOINC honors resource share -- that a user with 60% SETI and 40% E@H does 60 hours of SETI for every 40 hours of Einstein.

At the end of the 100 hours, does it really matter which ones went to each project?

---

People want the option to not use your scheduler as is. So you are saying it is out of your ability to make a change so it:
1: goes into panic mode only if deadline is imminent.
2: only readjusts timeshare to deadline project, not dedicate everything to it.
3: Not go overboard when the wu is not going past deadline.
4: Return to normal immediately when no deadline has problem.

Once again. It does EXACTLY what you list in the four points. Your definition of deadline imminent leaves out modem users. Once the CPU scheduler has determined that there is a problem, it is late enough that it has to dedicate the entire CPU the process (it is being optimistic that the process will work itself out until it is extremely obvious that it isn't). It does return to0 normal mode within one timeslice of when it can.

Want option to not be based on connection rate. connection rate is not the deadline.

The more I read these threads, the more I think "panic mode" is a poor choice of terms.

We aren't talking about "panic" like you'd see in the typical Japanese monster movie when everyone in the city is running while Godzilla is kicking it to pieces.

What we're talking about is looking at the queue, and if there is some doubt about making schedules, doing the work in order by deadline so that it's ready for the next connection.

That is followed by a mechanism that works long-term to make sure that BOINC honors resource share -- that a user with 60% SETI and 40% E@H does 60 hours of SETI for every 40 hours of Einstein.

At the end of the 100 hours, does it really matter which ones went to each project?

but when at the end of 100 hours you have done 95 hours of E@H (@ 40% timeshare) and 5 hours of S@H that's not good and this is what is happening. Projects are accumulating weeks of debt.

---

<a href="http://tinyurl.com/9hemz"><img src="http://www.boincsynergy.com/images/stats/comb-1441.jpg"></img></a>