Moo: New Version

Good news for Trigggl?

Quote:

---

I just released v1.01 of our wrapper application. In addition to previous 32-bit Windows versions there's now versions for 64-bit Linux. Both Stream and CUDA are available for both.

Some of the major changes in this release:


* Overactive hang detection code fixed and it now takes into account relative speed of the used card. This should fix many failures and even slowness seen by users especially on CUDA platform.
* Limited retries (no checkpoint or premature exit) to ten (10) so that there's no chance for an endless loop.
* Added support for setting core used by the client. I'll add a preference for this soon. Allows advanced users to set best performing core based on their knowledge or client benchmarks.
* There's now a default Distributed.net ID used if one is not specified in preferences. For now it's simply "u_<userid>@moowrap.net" where <userid> is your User ID of your account on this project. Yes, you can still set a custom one in your preferences.
* And some other minor improvements and fixes around the code.

---

Official request to Admin crew: Can we turn on the RSS feed for this project?

Quote:

---

Originally Posted by Maxwell

Official request to Admin crew: Can we turn on the RSS feed for this project?

---

After having asked DrPop we decided to turn on this RSS feed.

I just added it. (WOOT- first task as admin)

If it does not work, or look strange please let us know.

Looks good. Much appreciated, Admin Beer!

Admin. Beer?? That mean Free Beer? :p

One question about this project. I just put in the second 5870 and now have run times of 600+ gpu, 600+ cpu. Credits are 6k sometimes and 4k other times. My box with one 5970 is doing 900+ gpu, cpu 9.5 for 9k. Credits may be similar but large cpu usage. Worth it? I'm ony using 3 of my 4 cores now. I'll try it running only 2 cores and leaving 2 cores for the gpu, see what happens.

I checked and I have the new ap now, but I get a message saying that this machine has reached it's daily quota of 2 tasks and I'm not getting new work.

Quote:

---

Originally Posted by Crazybob

I checked and I have the new ap now, but I get a message saying that this machine has reached it's daily quota of 2 tasks and I'm not getting new work.

---

I had that message too for awhile a few days ago. It's from too many workunits errorring out or being aborted. It resets after 24 hours and starts getting new work again. It'll get one WU, when that one completes successfully, it'll ask for two, etc. and keeping doubling up to the max number of workunits allowed by the project, with each successful return. Everytime a workunit errors out or gets aborted, it cuts that by 1 again. At least in theory that's how it works...

Yep, that seems to be the case. 1 machine must have got bad WU's and I had to put it back on MW. Sucks. Just put a call into my son so he could check his. Looks like he hasn't connected since Friday night.

Yep, my computer had rebooted and stuck on the logon screen (sorry didn't get to my computer all weekend I know I'm horrible). Logged on, all my Moo WU's errored out, reset the project all wu's still computation error. Back to Collatz for now.

- Jason

I posted over on their boards, we'll see if anyone has a fix.

http://moowrap.net/forum_thread.php?id=30#185

14.2 mil credits in 7 days, even with one day lost to errors. I'm loving it. :cool:

Finally got my HD 5870 back online in the middle of the night. Woke up this morning to way more credits than i expected from Moo. I didn't even overclock it. Wow...

Somebody needs to send me a PM or hit me upside the head or something when this kind of news comes out. |-) :-o

My 5850 still keeps hitting computation errors. Tried resetting the project and everything. Not sure what the deal is.... Haven't tried detaching and reattaching yet but going to try that this afternoon.

I remember reading that you need ATI driver version 10.4+ for it to work as properly as it can, considering. So if you've got that, only thing left to do is wait for them to fix their code, which they have admitted has problems.

Quote:

---

Originally Posted by John P. Myers

I remember reading that you need ATI driver version 10.4+ for it to work as properly as it can, considering. So if you've got that, only thing left to do is wait for them to fix their code, which they have admitted has problems.

---

IIRC MOO, like PrimeGrid, needs OpenCL to crunch. If you are using older ATI drivers you can install the ATI/AMD OpenCL SDK or just update the drivers. Just remember that 4xxx series and below do not fully implement OpenCL so they are out of luck.

Quote:

---

Originally Posted by John P. Myers

I remember reading that you need ATI driver version 10.4+ for it to work as properly as it can, considering. So if you've got that, only thing left to do is wait for them to fix their code, which they have admitted has problems.

---

Are you referring to ATI Driver 10.4 or Catalyst version?

I have catalyst version 11.3 and Driver Version 6.14.10.7171

This ran just fine before this past weekend. Whatever they did on their end with the application jacked everything up. Have no clue what is going on since it seems as though everyone else is running just as they were before.

Quote:

---

Originally Posted by Crazybob's Son

Are you referring to ATI Driver 10.4 or Catalyst version?

I have catalyst version 11.3 and Driver Version 6.14.10.7171

This ran just fine before this past weekend. Whatever they did on their end with the application jacked everything up. Have no clue what is going on since it seems as though everyone else is running just as they were before.

---

One more thing you could try. Completely shut down BOINC. Find the MOO project folder, its named moowrap.net, and delete it. Restart BOINC. If you still have problems after that then its something in the project app.

Quote:

---

Originally Posted by Fire$torm

One more thing you could try. Completely shut down BOINC. Find the MOO project folder, its named moowrap.net, and delete it. Restart BOINC. If you still have problems after that then its something in the project app.

---

Nope, that didn't work either. Suspending and running Collatz until they say it's fixed.

Returned my first CUDA task. Waiting to see what it gives me.

EDIT: 9k, nice!

The first one seems to take a long time on CUDA. Thinking maybe I should stick to PG on CUDA while the getting is good.

Quote:

---

Originally Posted by denim

The first one seems to take a long time on CUDA. Thinking maybe I should stick to PG on CUDA while the getting is good.

---

My first one took 3 hours. Up til now, the credit given to me has been proportional to the length of the task. Of course, I run Linux. OS could be a factor.

Finnaly figured out my problems with Moo Wrapper and dual 5870s in CF. I ran the following.

"C:\Users\All Users\BOINC\projects\moowrap.net\dnetc518-win32-x86-stream.exe" -bench
Actually the above command is dependent where your Bonic project files are.

This came back and said core 3 was the best option for Moo Wrapper. I was running with the default which was core 0. I modified this using dnetXX --config. Didn't work. I then changed my Moo Wrapper preferences to use core 3. Now I've gone from 3600 to complete a work unit to 900 seconds, netting 12K.

Dan

Will that result hold for single 5870s?

Quote:

---

Originally Posted by Maxwell

Will that result hold for single 5870s?

---

It might. Seems that the DNet app is s'pose to bench mark and pick the best core when running. Core 0 is the worst core. You need to run the benchmark to find out which core is best suited for running DNet with your card.

Dan

Quote:

---

Originally Posted by dan

Core 0 is the worst core.

---

Not necessarily. On my 5770, core 1 was the worst, core 2 was the best and core 0 (which it uses by default) fell in between.

Quote:

---

Originally Posted by Mumps

Not necessarily. On my 5770, core 1 was the worst, core 2 was the best and core 0 (which it uses by default) fell in between.

---

Sorry. On my machine core 0 was the worst.

Dan

Quote:

---

Originally Posted by dan

Sorry. On my machine core 0 was the worst.

Dan

---

Perhaps you should swap your cores in order of usefulness.

EDIT: On the motherboard to be clear

I ran it on 4 machines:

2x5870: 3 was the best by far
1x5870: 3 was just barely better than 0
1x4870: 0 was best
1x4870: 0 was best

-1 is supposed to auto-select the best. I don't trust it.

So I set up a home and a work preference, 3 and 0 respectively. Put the 5870s on home, and the 4870s on work.