I figured I would start a thread because even though it applies the discussion going on in the credit discussion it is also a different issue as well.

The longest workunits now are claiming I am seeing 70+ credits on C2D technology.Now there are concerns of not enough average credit due to these.

Edit-I guess MY question is how much longer can the work units get?

---

good question.

I must have one of these longer units running at the moment and I'm not sure what to do with it.

It's been running for nearly 4 hours but said it should complete in 3 hours and 9 minutes.

This on a Core 2 Quad.

Keep running ? Abort ?

---

Keep running of course:)! I have rarely seen a wu fail.Some are longer now and I am sure admin will address this soon. My longest are about 3 1/2 hours now on a C2D 2.66 xeon

My average cr/h has already dropped from 23/24 to 20, due to the newer longer WU's. I have had a number claiming more than we are being granted (longest so far 19,069.68 seconds (5 1/2 hours)- Claimed 67.02- Granted 50.00- Cr/h 9.44).

---

Stupid me...unit finished at 4:15.

Claimed 64.45 credits and granted 50.00

---

Hi,

I think you must adjust credits proportionaly with the lenght of WU.

Now more than 3 hours under Linux 64 for some WU with my 2.66 Ghz.
More than 90 claimed for only 50 Credits granted !!!

Please do something.

Thanks
Julien

---

Hell0
I think you must adjust credits proportionaly with the lenght of WU.
I have had claimed more than we are being granted. Look at the longest time of one WU!
CPU Time - Claimed - Granted
23,576.20 56.29 50.00
5,628.06 13.57 50.00

Please admin will you address this soon?

Regards,
Olivier.

LOL you guys crack me up.

Where was the clamoring for FEWER points on SHORT wu's? Now that some are (GASP!) longer all of a sudden there's 'concern' or a call for credit adjustment? Sorry, can't have it both ways.

Let's face it, this project has been more than generous over the past few months. Pretty stable too with few hiccups. It's high time Scott either keeps the units longer or drops the granted to a value that at least makes some attempt at meaningful comparison cross project, imperfect as it is.

The excuse of awarding disproportionately higher points to participants for a project in late alpha/early beta is over. It's chugging along just fine.

Hell0
I think you must adjust credits proportionaly with the lenght of WU.
I have had claimed more than we are being granted. Look at the longest time of one WU!
CPU Time - Claimed - Granted
23,576.20 56.29 50.00
5,628.06 13.57 50.00

Please admin will you address this soon?

----

Your 5,628.06 unit - how much LESS should the granted be?

There's no need to go around being reasonable about this.

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

I need a little more data to make the correct credit increase, so let's conduct a little poll: by what factor has the run-time increased for you?

EDIT: Sorry I didn't post sooner, but I've been *slammed* this week.

---

Scott Kruger
Project Administrator, Cosmology@Home

Runtime increase has been 30-50% for my farm.

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

I need a little more data to make the correct credit increase, so let's conduct a little poll: by what factor has the run-time increased for you?

EDIT: Sorry I didn't post sooner, but I've been *slammed* this week.

Hi,

I made averages on my C2Q. WUs are now 1.6 times longer than before.

---

Runtime increase has been 30-50& for my farm.

Seconded.

The slower the rig, the bigger the increase. Strange...

---

mic.

Hi,

I made averages on my C2Q. WUs are now 1.6 times longer than before.

+1

1.5 - 1.7 for me

75 - 85pts/wu will be fine

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

I need a little more data to make the correct credit increase, so let's conduct a little poll: by what factor has the run-time increased for you?

EDIT: Sorry I didn't post sooner, but I've been *slammed* this week.

Thanks for your reply Scott ;)

---

I need a little more data to make the correct credit increase, so let's conduct a little poll: by what factor has the run-time increased for you?

Since when are you giving this new longer ones out?
I still have not one that has claimed more than got granted.

---

Grüße vom Sänger

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

Interesting, if credits are to be increased does that mean the end, or reduction in, short work units? As it stands 50 granted on a wu taking 4600 seconds is quite generous.

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

Interesting, if credits are to be increased does that mean the end, or reduction in, short work units? As it stands 50 granted on a wu taking 4600 seconds is quite generous.

We're sort of at the mercy of one of Ben's grad students; we distribute whatever sort of parameters he needs for his research. Runtimes might go up or down, honestly.

It shouldn't happen too often, though.

---

Scott Kruger
Project Administrator, Cosmology@Home

Yes, of course I'll be increasing credits. We set up a new parameter file generator, but I was unsure as to how the WU run-time would change.

Interesting, if credits are to be increased does that mean the end, or reduction in, short work units? As it stands 50 granted on a wu taking 4600 seconds is quite generous.

We're sort of at the mercy of one of Ben's grad students; we distribute whatever sort of parameters he needs for his research. Runtimes might go up or down, honestly.

It shouldn't happen too often, though.

Scott is there any way to tell by the parameters loaded how long the results will take?

You could then adjust credit given based on a sub-set before or as the work goes out. I will ask the question again that I posted originally ...how long could the workunits take?

Will there be say 10 hour work down the road? (from the original longs of 2 hours now 3-1/2 as a reference.

Seems the grad student might have an idea if parameters would give short or long computing times.
How many more of the longer units are there to crunch?