My present WU has been running 14 hours and is only at 21%. This has not happened before. What's going on?

Occasionally an nbody wu is extra long. I've had up to 40 hours before, my current one is going to be ~25 hours.

---

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

My present WU has been running 14 hours and is only at 21%. This has not happened before. What's going on?

Are you crunching on cpu or gpu ? If cpu then it is ok. For gpu it depends on your card but usually under 30 minutes max

My present WU has been running 14 hours and is only at 21%. This has not happened before. What's going on?

Are you crunching on cpu or gpu ? If cpu then it is ok. For gpu it depends on your card but usually under 30 minutes max

If you look her only computer doesn't have a gpu.

---

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

I've seen it now.

My advice is to add 6950 card if she can offred it as that will improve the outcome a lot but in MW only.

HassanShebli wrote

My advice is to add 6950 card if she can offred it as that will improve the outcome a lot but in MW only.

It's on my list, along with a quad core and some other stuff. :) Just wanted to know if this was a different type of work unit, since my previous ones weren't so slow.

Yeah, I noticed the same thing. Also, MW@Home began to send me programs that require 4.00 CPUs. When those begin to run everything slows way down. They are labor intensive so other programs cannot run when the computer is involved with them; I cannot see them finishing before the due dates.

Anybody else seeing this?

^..^~~

I will finish this one by the due date 11/15, but it is at the expense of other projects, so I won't be asking for more tasks after this.

I will finish this one by the due date 11/15, but it is at the expense of other projects, so I won't be asking for more tasks after this.

Is that any way for the User of the day to be talking?

---

Aw, I'm sorry.

Actually I had hoped to know more about the nature of this long task. Can anyone explain that?

Sorry, since I don't run MW CPU WUs. But when I see long GPU WUs, I know that they are stuck and I have to stop and restart them.

---

HassanShebli wrote

My advice is to add 6950 card if she can offred it as that will improve the outcome a lot but in MW only.

It's on my list, along with a quad core and some other stuff. :) Just wanted to know if this was a different type of work unit, since my previous ones weren't so slow.

I would not advice an ATI card as they can only be used for MilkyWay (or Conjuncture, but that is not a University, but private project).
With nVidia cards, also Einstein@home and GPUGRID (medicine) can be done as some others.
I have two ATI's in a workstation but when MW is dowm I can power my system down as well.

---

Greetings from,
TJ

I agree with you, AMD supported projects are very limited… actually it is only one (Milkyway) others are useless .. . to me at least.

I am so frustrated that other scientific projects didn't provide AMD support for their projects yet, and I think it is not that hard. They should do it as the second most important project on the net is MW as all AMD scientific – oriented owners are participating in.

So dear Kathryn :
If you're interested in MW only so you better buy AMD card with double precision support, and the best bargain is to get 2 6950 which will exceed 6990 but cheaper. This one is my favorite as it has dual fan so it is so quiet :
http://www.newegg.com/Product/Product.aspx?Item=N82E16814127588

But if you want other project buy Nvidia, my advice to get 2 570 such as this one:
http://www.newegg.com/Product/Product.aspx?Item=N82E16814127582&Tpk=570

finally:
make sure that your board supports the card and have enough power to run your system.

Aw, I'm sorry.

Actually I had hoped to know more about the nature of this long task. Can anyone explain that?

The n-body are random lengths, only way to know is to let it run a bit and calculate it out to see the length. The ps tasks generally run at the pre-estimated time.

---

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

There is a good chance that SETI will have a ATI app when they release the V7 app, it will be OpenCL but it will work.

---

WCG has a project using OpenCL, hence support ATI/AMD. Beta starts in Jan.

---