News:

Members can see all forum boards and posts. Non members can only see a few boards.  If you have forgotten your password use this link to change it.
https://forum.boinc-australia.net/index.php?action=reminder

Main Menu

Project Updates

Started by Daniel, July 17, 2015, 01:05:59 AM

Previous topic - Next topic

chooka03

LOL...morning mate  :wave:
I just saw that about the credits. You were lucky to be an early adopter  :thumbsup:
I'm not worried about being number one. If I was, I would have been putting more effort into Universe & NFS @ Home but I have other things I want to do :) It's all good.

I think I was more worried that I spent a bit of time to try get the team into the top 30, only to watch us slip out again and not really being able to help with non NVIDIA GPU's.I doubt my CPU's would have been enough to save us with the high credits being awarded for GPU work.

Guess we'll wait and see but they are certainly flipping and flopping all over the place in regards to credits. Sounds like Eric didn't put a whole lot of forethought into it before implementation.

:rocks

tazzduke

Morning Chooka

Well things are looking better, its a pity I cannot edit my last post, but Eric is still fine tuning the credit system, so now its moved from Credit New, to a more fair system.

Yeah just having a bit of a joke, currently our world position is 27 for Numberfields.

Cheers and have a beaut day :-)



 AA 24 - 53 participant

Dingo

#17
Quote from: tazzduke on March 27, 2019, 08:00:54 AM
Morning Chooka

Well things are looking better, its a pity I cannot edit my last post, but Eric is still fine tuning the credit system, so now its moved from Credit New, to a more fair system.

Yeah just having a bit of a joke, currently our world position is 27 for Numberfields.

Cheers and have a beaut day :-)

You should be able to modify your own post. I had a look at the permissions and you have the same as everyone else. There should be a "modify" tab on the post on the top right?

EDIT:  I just checked my normal user login and the "Modify" tad was there on a post I created as that user.


Radioactive@home graph
Have a look at the BOINC@AUSTRALIA Facebook Page and join and also the Twitter Page.

Proud Founder and member of BOINC@AUSTRALIA

My Luck Prime 1,056,356 digits.
Have a look at my  Web Cam of Parliament House Ottawa, CANADA

tazzduke

HI Dingo

Thankyou for the info.



 AA 24 - 53 participant

tazzduke

Greetings All

A further update is that the CPU credits have been adjusted for the time being, so at the moment this is how it is.

a CPU workunit - 8000 credits

a GPU workunit - 4000 credits

This may change again, once Eric has released an AMD (OpenCL) app.

Regards



 AA 24 - 53 participant

Dataman

Quote from: tazzduke on March 28, 2019, 07:44:56 AM
Greetings All

A further update is that the CPU credits have been adjusted for the time being, so at the moment this is how it is.

a CPU workunit - 8000 credits

a GPU workunit - 4000 credits

This may change again, once Eric has released an AMD (OpenCL) app.

Regards

Thanks!


chooka03

Thanks for the update Tazzduke!

chooka03

Ohhhh.... I only just noticed that the GPU tasks only work with Linux systems.  :thumbdown:
I hope when they roll out work for ATI cards that it's Windows supported.

It looks like Eric keeps changing the credits too. You may end up with a MASSIVE total Tazzduke for being an early adopter  :wink

I have too many other GPU challenges ahead of me anyway  :crazy

tazzduke

Greetings All

I have moved on to Primegrid now, as I got to my goal of 10 Million @ Numberfields.

They have adjusted the credits again at Numberfields, its now 800 per GPU workunit.

Not to sure on what CPUs are getting though.

Cheers



 AA 24 - 53 participant

Dataman

Adios GPU
From the project:

So there have been some new developments over the last week. It's both good and bad.

First of all, some history. The reason I waited so long to develop a GPU app is because the calculation was heavily dependent on multi-precision libraries (gmp) and number theoretic libraries (pari/gp). Both of these use dynamically allocated memory which is a big no-no in GPUs. I found a multi-precision library online that I could use by hard coding the precision to the maximum required (about 750 bits), thereby removing the dependence on memory allocations. The next piece of the puzzle was to code up a polynomial discriminant function. After doing this, I could finally compile a kernel for the GPU. That is the history for the current GPU app. It is about 20 to 30 times faster than the current cpu version (depends on WU and cpu/gpu speeds).

But then I got thinking... my GPU polynomial discriminant algorithm is different from the one in the PARI library (theirs works for any degree and mine is specialized to degree 10). So to do a true apples-to-apples comparison, I replaced the PARI algorithm with mine in the cpu version of the code. I was shocked by what I found... the cpu version was now about 10x faster than it used to be. I never thought I was capable of writing an algorithm that would be 10x faster than a well established library function. WTF? Now I'm kicking myself in the butt for not having done this sooner!

This brings mixed emotions. On one side, it is great that I now have a cpu version that is 10x faster. But it also means that my GPU code is total crap. With all the horsepower in a present day GPU I would expect it to be at least 10x faster than the equivalent cpu version. Compared with the new cpu version, the gpu is only 2 to 3 times faster. That is unacceptable.

So the new plan is as follows:
1. Deploy new cpu executables. Since it's 10x faster, I will need to drop the credit by a factor of 10. (Credits/hour will remain the same for the cpu but will obviously drop for the GPU)
2. Develop new and improved GPU kernels.

I don't blame the GPU users for jumping ship at this point. Frankly, the inefficiency of the current GPU app just makes it not worth it (for them or the project).

For what it's worth, I did have openCL versions built. Nvidia version works perfectly. The AMD version is buggy for some reason, as is the windows version. Since I will be changing the kernels anyways, there is no point in debugging them yet.


:wavey2:


chooka03

Jesus. This has all been pretty haphazard.
Meh... I have plenty of other projects to crunch so no biggy.


Daniel

#26
NumberFields has added some more beta GPU applications.
Earlier today I got some workunits for my Nvidia GPUs on Windows.

It looks like the new application is running ok.
Eric Driver (the project admin) posted ...

QuoteThe Nvidia windows version seems to be doing pretty well. Many successful results from multiple users. Only 2 compute errors.

Good news.  :thumbsup:

JugNut

#27
Yep they do work OK but..

A WU took 25 mins and received 35 cr on my gtx 970's.   The 970's may not be the most powerful of cards around these days but still that's pretty bad.  So at the moment there probably not worth it.

PS: If you'd like you can try this an app_config to utilize your GPU's better. The below app_config is for running 2 at once.
Run through a few before setting up the app_config as there are some big difference between how each WU behaves.  Some WU's hardly use the GPU at all but others use it heaps.

<app_config>

<app>
<name>GetDecics</name>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>

</app_config>



EDIT: I've packed it in as they don't seem worth it for now. 

How did it go for other team members?


 - Participated in AA's 27 - 55 & Team Challenge # 1.
My team (Boinc@Australia) stat's
My personal stat's


     Crunching today for a better tomorrow...

chooka03

Well I've only got Radeon cards AND I'm using Windows not Linux so no GPU work for me to try.

JugNut

Yea the version I tried was a Nvidia app for windows but so far there's no AMD version.  But they are working on it so it's possible the AMD cards will perform much better? But time will tell.


 - Participated in AA's 27 - 55 & Team Challenge # 1.
My team (Boinc@Australia) stat's
My personal stat's


     Crunching today for a better tomorrow...