OK so where do we start.

Right let us start with the server

1. Is the server a dedicated sever or is it used for something else as well?

2. Is the server at the moment getting hammered by?

A. Someone like that has 2 computers with 2 duo cores who is running an updater script say every 20 mins.
B. Someone that has a farm of computer?
C. Someone that has a farm of computers with ATI cards that complete work units in about 1.30 min?
( this was one persons cobstones for today 354,135.96 on milkyway)

3. If the server is not dedicated them i would recomend this thread


Now lets move onto the work units themselves

1. How long after a work unit has been validated dose the science of that work unit become invalid?

2. Is there a way to increase the amount of work units on the server as it took about 5 hours before one of my computer was able to get work even though im running this update script?

3. Can the cpu limit be increased again because with these ATI and optimised app the work is getting done quicker?

---

Smoke me a kipper ill be back for breakfast

3. Can the cpu limit be increased again because with these ATI and optimised app the work is getting done quicker?

Like I've requested on a previous thread, if the server sends CPU tasks that can be done in an hour and GPU tasks that can also be done in an hour, we wont have GPUs sucking WUs dry.

Case in point: seti@home (CPU & Nvidia) & GPU GRID (ps3 and nvidias)

The server is dedicated to specifically the milkyway project. It can handle quite a bit but C is probably stressing it the most currently because the workunits complete so quickly.

---

Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute

It can handle quite a bit but C is probably stressing it the most currently because the workunits complete so quickly.

When you increased the length of the work units there was remarkable improvement in the response.
It would appear that the number of shorter work units (S7x and S8x) generated lately increased.
This put the stress back on the server as well the traffic on the network.
Wouldn’t another increase in length of the work units help?

---

A BLAST FROM YOUR PAST

If the WUs were say 4x would that make a marked difference? You could still keep the deadline and everyone should be able to meet it with the small cache settings?

Could the ATI and CPU apps process different work units? Then make the GPU ones like 10x longer and everyone is happy...well mostly anyway ;)

---

I am not sure that the actual times are for completing a MW task on the GPU, but, compared to my CPUs which take from 15-40 minutes to do the work with optimized versions the increase of task length on GPUs by a factor of ten means only that it would take about 120 seconds to compete a task rather than 12 seconds. The multiplier would need to be at least 100 to achieve the effect desired ... that would push the task time to 20 minutes ...

An alternative would be a task "bundle" where the GPU would get a package of tasks instead of single tasks. BUt, as I understand the feedback and task creation method none of these may be practical methods.

I'll add:

1. Can we get an update on how everything is Dave?

2.(edit) Dave is Travis AWOL?

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

2. Is Travis AWOL?

All those who aren't here would you please raise your hand?

2. Is Travis AWOL?

All those who aren't here would you please raise your hand?

I ment that for Dave, as Travis doesn't seem to be around.

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

2. Is Travis AWOL?

All those who aren't here would you please raise your hand?

I ment that for Dave, as Travis doesn't seem to be around.

Do I have to request leave when I'm not around?

Two weeks starting Monday please, no mobile phone calls including Retired BT. Oh go on then, phone me in the Caribbean if you must.

---

Ice in the Caribbean?!

XD

2. Is Travis AWOL?

All those who aren't here would you please raise your hand?

I ment that for Dave, as Travis doesn't seem to be around.

I know...but it just seemed funny to me...

This is mid-term / spring break time for colleges and universities, plus he had mentioned he had been sick...

haha ya travis has not been feeling well. also our spring break was last week. hopefully he'll be back in business this week.

---

Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute

haha ya travis has not been feeling well. also our spring break was last week. hopefully he'll be back in business this week.

Well that's No excuse!! He should be spending all his time here! ;p

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

Ice in the Caribbean?!

XD

He's captain of the Jamaican bobsled team.

---

me@rescam.org

Ice in the Caribbean?!

XD

He's captain of the Jamaican bobsled team.

I'm very good at bobbing and sledding, but this year I'll be doing the Barbados Ice Chill! :)

---

I'll add:

1. Can we get an update on how everything is Dave?

2. Dave is Travis AWOL?

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

Ice in the Caribbean?!

XD

He's captain of the Jamaican bobsled team.

I'm very good at bobbing

That is so wrong.

---

me@rescam.org

Ice in the Caribbean?!

XD

He's captain of the Jamaican bobsled team.

I'm very good at bobbing

That is so wrong.

OK, weaving then.

---

I'm not sure where Travis is. I haven't talked to him in days but have seen emails pass by that he wrote to various people that I was copied on. However, everything seems to be running fairly smoothy except for the workunit distribution issue. But I'm almost certain that the server can't keep up with all the requests from the users with graphics cards.

---

Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute