Best GPU Projects?

Hey guys,

I wanted to start a list of the best GPU projects to work on while we wait for another WCG supported GPU project. Anyone have any suggestions to add? Or have a concise (< 3 sentence) summary describing what it is and how it helps advance science/humanity? (I'll update these as people contribute)

NVIDIA:

GPUGrid: biomedical research (full-atom molecular biology simulations)
*Notes: See "Everything You Need to Know About GPUGrid/BOINC" thread
*Requires 1 full CPU core on Kepler (only part of CPU core on other Nvidia cards)
*Give extra credit for <24h return times: set <report_results_immediately>1</report_results_immediately> in cc_config.xml

Einstein: Physics? (continuous gravitational-wave sources)
**searches radio telescope data (the data from SETI) for pulsars
*I'm planning on Reading Gandalf's link and expanding on this when I get a chance, but that takes more brain power than I currently have.
*Note: great for older generation cards
*Contributions: 24 Pulsars Found

ATI:

POEM: biomedical research (protein folding using Anfinsen's dogma)
*Note: GPU Client very new. Plenty of CPU tasks, but because of small project size, limited GPU tasks with no plans to expand.

Einstein: Physics? (continuous gravitational-wave sources)
*searches radio telescope data (the data from SETI) for pulsars
*I'm planning on Reading Gandalf's link and expanding on this when I get a chance, but that takes more brain power than I currently have.

MilkyWay: astrophysics (generate accurate three-dimensional dynamic models of stellar streams in the immediate vicinity of our Milky Way galaxy)
*Note: secondary objective is to develop and optimize algorithms for distributed computing
*Requires DP on GPU (58xx, 59xx, 69xx, 79xx)

SETI: Extra Terrestrial (analyze radio signals, searching for signs of extra terrestrial intelligence)
*IIRC, this had another purpose as well... anyone want to expand?

Thanks for the help guys

Quote:

---

Continuous gravitational-wave sources ... Someone want to explain this one?

---

http://www.astro.cardiff.ac.uk/resea...ge=gravsources

If you thought I'd put it into my own words, you've got another guess coming.

Quote Originally Posted by Otis11 View Post

Quote:

---

... Yeah, I know I gotta fix it... just don't remember the place and haven't cared enough to dig around for it. Asked SNURK, should be fixed shortly.

---

http://i1007.photobucket.com/albums/...sig/305484.gif ; Otis11

305484 is the number you need. It's magic! :up:

MilkyWay@Home requires double-precision, so it performs well if you have a higher-end GPU (like Radeon HD 79xx). On such GPUs you can complete a WU under 1 minute. Lower models are running DP at a considerably reduced clock, so the performance is relatively weak.

Einstein@Home uses by default the GPU up-to ~75% when running a single task, but you can easily configure multiple WUs/GPU in their control panel (no need to use app_config). BRP4 tasks support AMD via OpenCL, but nVidia tasks run currently using CUDA 3.2, which isn't optimal performance for later (Kepler) family. Apparently, CUDA 4.x/5.x has a bug which causes BRP4 tasks to fail, but nVidia didn't fix this yet.

GPUGrid tasks run GPU at high load, on Kepler GPUs they also require a full CPU core for feeding the GPU. There are long tasks, which can run several hours (more than 12h on slower GPUs) and short tasks (few hours). They have CUDA 3.x and 4.x tasks for older and newer families. They give credit bonus if you complete a task within 24h after download, so it's required to set a low WU cache and set <report_results_immediately>1</report_results_immediately> in cc_config.xml
It seems to me, that there are currently no tasks in queue...

Looks like for ATI and medical/science/humanity wise POEM seems top of my list right now.

Outer space stuff is great and all, but unless the aliens come here freely and willingly help cure diseases and solve humanitarian problems (and people don't kill them for trying), I'll stick to something that benefits the scientific/medical/humanitarian community. After all, I'm not that confident in our ability the get along peacefully with another species. Hell, people can't get along with each other half the time...stick fear and ignorance in the mix and you get wars.

The only thing that keeps me from loosing complete faith in humanity lately is belonging to a place like XS and seeing what the good folk in this small community achieves and accomplishes. What we do here together everyday amazes me!

I'm checking out F@H again.

I'm not convinced, that today's medicine is the only and right way to fight diseases like cancer. I know of alternative methods, which helped a lot of people, but it depends on type of cancer. Had some cases in the family and they tried alternative methods too, so it's hard to say if/which really helped. Chemotherapy is definitively a total overkill and trying to kill cells by shooting a radioactive beam straight into the body can result in even more problems. And once the body builds too many metastasis, it's too late. Also some diagnostic methods (like PET) are based on injecting radioactive material into the body - I measured over 600 uSv/h on my brother (external rate) few hours after the test. So basically radioactivity can cause cancer, though it's used in huge doses for the treatment.. Hm...
Unfortunately there are currently no other alternatives in 'modern' medicine, but I believe that in several years we will consider such treatment in a different light...

Quote:

---

Originally Posted by Mumak

MilkyWay@Home requires double-precision, so it performs well if you have a higher-end GPU (like Radeon HD 79xx). On such GPUs you can complete a WU under 1 minute. Lower models are running DP at a considerably reduced clock, so the performance is relatively weak.

---

Any ATI 58xx or 69xx card will work well on MW@H as they are DP capable.

i'll run SETI in case we find an alien race who can tell us how to conquer diseases(or our planet)

no but in all seriousness, maybe i'll run SETI, E@H and M@H until WCG comes out with something new :shrug:

Copycat. :up:

Quote:

---

Originally Posted by nanoprobe

Any ATI 58xx or 69xx card will work well on MW@H as they are DP capable.

---

I ran a 6950 on milkyway. Crunched pretty good.

Since no one seems to have answered your question; E@H searches radio telescope data (the data from SETI I think) for pulsars.
It does run on Nvidia cards. I have older ones that don't like GPUGrid that work great (or better) for E@H.
In terms of how "useful" it is to humanity, that's up to you. It has contributed quite a bit to science by finding a bunch of pulsars.
Space is big. That makes it worth checking out, IMO.

Sorry for the delay guys, I'll update the first thread as soon as I have time to read through all these responses in detail.

Also, don't know if it's been mentioned here yet, but I read in another thread (I think the HCC countdown) that POEM is having trouble getting enough work to support all the users - any more information on that?

Quote:

---

Originally Posted by Otis11

Sorry for the delay guys, I'll update the first thread as soon as I have time to read through all these responses in detail.

Also, don't know if it's been mentioned here yet, but I read in another thread (I think the HCC countdown) that POEM is having trouble getting enough work to support all the users - any more information on that?

---

Here is an excerpt from a POEM posting...

Quote:

---

Sort of...if you want cpu tasks you can get all you can handle, but the gpu tasks are being limited as they just don't have the manpower to review more then they are sending out right now. The Server CAN send out more units but that means they would just be put on the shelf for an indeterminate amount of time and that is not ideal. So they made the decision to only send out as many units as they can review and no more. With all the people wanting medical type projects as opposed to math type projects that means a shortage of gpu units.

---

Ok, OP updated as far as I can understand without more indepth research... please let me know if I misunderstood anything.

Also, keep the information flowing! Any more projects? Any more important notes on these projects? What are their contributions and goals?

Anyone familiar with:
Nvidia:
Collatz Conjecture
PrimeGrid
DistrRTgen
Ibercivis
SETI@home*

ATI:
Collatz Conjecture
PrimeGrid
DistrRTgen

Some may be listed above already or listed twice - I'm looking for projects specifics. Does it run better on one card over the other?

Thanks guys!

Quote:

---

Originally Posted by Evantaur

i'll run SETI in case we find an alien race who can tell us how to conquer diseases(or our planet)

no but in all seriousness, maybe i'll run SETI, E@H and M@H until WCG comes out with something new :shrug:

---

Exactly why I started this thread - I figure we have a few more days to figure out what's the best thing to run (each may have their own opinions, but better to have an informed opinion!). Then we can switch from HCC to one of these until WCG has more work for us!

So where did everyone end up going?

9860 SETI@home 5.5.2013 18:50:39 Sending scheduler request: Requested by user.
9861 SETI@home 5.5.2013 18:50:39 Requesting new tasks for ATI
9862 SETI@home 5.5.2013 18:50:42 Scheduler request completed: got 0 new tasks
9863 Einstein@Home 5.5.2013 18:50:47 Sending scheduler request: Requested by user.
9864 Einstein@Home 5.5.2013 18:50:47 Requesting new tasks for ATI
9865 Einstein@Home 5.5.2013 18:50:52 Scheduler request completed: got 0 new tasks
9866 Einstein@Home 5.5.2013 18:50:52 No work sent
9867 Einstein@Home 5.5.2013 18:50:52 see scheduler log messages on http://einstein.phys.uwm.edu//host_s...s/7191/7191633
9868 Einstein@Home 5.5.2013 18:50:52 Gamma-ray pulsar search #2 is not available for your type of computer.

not looking good, SETI is dry and Einstein doesn't have linux apps :(

I'm running Milkyway and POEM - only selected ATI at POEM cuz WCG gets my CPUs :up:

I'm am also a bit of a stats whore ;)(not just points but my overall stats)

- I'm pushing 1 new project to 25K (Docking@home) and a couple of others will go higher ... looking to optimize my position in as many *Mega Milestone* categories for team XS as I can
http://stats.free-dc.org/stats.php?p...4492e2cd4a2518
:sonic:

Just dusted off my Nvidia cards & moved back to GPUgrid. Couple of my 580's don't want to wake from their hibernation :shakes:

My 460 is back on GPUGrid as well... that's a shame some of them 580s don't want to run!

For ATI I guess Einstein is the way to go if you crunch Windows, otherwise it's MilkyWay?

Snow - How's POEM working? any apparent shortages? Is Docking@Home GPU based?

Thanks guys!

My 7950's are now doing MW@H. Others are doing Einstein, maybe will continue with GPUGrid too.
I'm also working on some benchmarks with different GPUs, can post them here later...

I'm doing both Einstein@H on one computer and MW@H on another. Of course, WCG has my CPUs. I am learning which @Home projects can't seem to support WU downloads. It's kind of hard to crunch a project, when they don't or can't send you WUs. So, I guess it is Einstein and Milkyway for my GPUs.

Edit: Otis - Milkyway does ATI. :up:

managed to get Einstein@home to run on linux :up:

but now it's running cpu tasks even when i told it to not use my CPU :shrug:
and the one gpu task it runs barely uses any gpu

AMD Radeon HD 7800 Series (:0.0)
engine clock 975MHz, memory clock 1500MHz, core voltage 1.256VDC, performance level 3, utilization 30%

In your Einstein Account's Preferences for this project option, does Use CPU say No?
If it does, then you may just need to clear your cache. It worked for me.

So some results

Milkyway@Home:
----------------
HD7950: ~37 / ~47 secs per WU; ~80 / ~107 pts per WU

Einstein@Home BRP4:
---------------------
650 Ti: ~3000 secs for 2 WUs
660 Ti: ~1200 secs for 1 WU, ~2200 secs for 2 WUs
HD7770: ~2300 secs for 1 WU, ~4000 secs for 2 WUs
HD7950: ~1050 secs for 1 WU
Each GPU WU is worth 500 pts

GPUs are clocked:
HD7950 - 1120 MHz
HD7770 - 1120 MHz
660 Ti - default
650 Ti - 1020 MHz

Mumak - do you run these WUs virgin or do they go through the app_config file?
If app_config, what settings are you using?

Quote:

---

Originally Posted by Gandalf

Mumak - do you run these WUs virgin or do they go through the app_config file?
If app_config, what settings are you using?

---

For MW@H I run a single WU per GPU, so no app_config.

For E@H there's no need to use app_config for running multiple WUs/GPU, since you can specify this in the E@H Preferences page - "GPU utilization factor of BRP apps" (set to 0.5 for 2 WUs/GPU).

For GPUGrid it doesn't make much sense to run multiple WUs/GPU, since they already utilize the GPU pretty high (>90% depending on GPU). Though I use app_config there to reserve a full CPU core for each WU on Kepler (by default the requested CPU reservation is much lower there, but on Kepler it seems the tasks require much more CPU resources). So there I use:

Quote:

---

<app_config>

<app>
<name>acemdlong</name>
<max_concurrent>1</max_concurrent>
<gpu_versions>
<gpu_usage>1</gpu_usage>
<cpu_usage>1</cpu_usage>
</gpu_versions>
</app>

<app>
<name>acemdshort</name>
<max_concurrent>1</max_concurrent>
<gpu_versions>
<gpu_usage>1</gpu_usage>
<cpu_usage>1</cpu_usage>
</gpu_versions>
</app>

<app>
<name>acemd2</name>
<max_concurrent>1</max_concurrent>
<gpu_versions>
<gpu_usage>1</gpu_usage>
<cpu_usage>1</cpu_usage>
</gpu_versions>
</app>

</app_config>

---

Quote:

---

Originally Posted by Otis11

Snow - How's POEM working? any apparent shortages? Is Docking@Home GPU based?

---

POEM appears to have a 100 WU in process limit - it is not stated on their site but I never get more than that and once I start reporting them in I fill back up to 100. Running 6 at a time via app_config on my 7950 because a single WU only uses 40% GPU ... each additional concurrent WU seems to take about another 10%. Six at a time takes about 26 minutes. Points are pretty good (if that matters) at 2,925.23 per WU but CPU usage is pretty heavy - they have it set to 1 GPU + 1 CPU but I found you really only need between .5 and .75 CPU for each GPU WU.
app_config ...

Code:

---

<app_config>

  <app>
      <!-- the name tag is the name of the application you want to control with this app_config file -->
      <name>poemcl</name>

      <!-- set max_concurrent to the total number of both CPU and GPU WUs to run at a time on your rig -->
      <max_concurrent>5</max_concurrent>

      <!-- gpu_version tag is where you take control of how BOINC allocates your resources for this app -->
      <!-- this line is unnecessary but adds control especially for those with two different cards in a rig -->
      <gpu_versions>

        <!-- gpu_usage calucation is 1 divided by the number of WUs you wish to run on any single card -->
        <!-- don't change this when adding a card as each will try to run this qty (if you have enough CPU) -->
        <gpu_usage>0.20</gpu_usage>

        <!-- cpu_usage is how much cpu resources BOINC will reserve to run each gpu WU-->
        <!-- calculation is the number of CPU cores/threads for all GPU work divided by the number of GPU WUs -->
        <cpu_usage>0.75</cpu_usage>

      </gpu_versions>

  </app>

</app_config>

---

Docking@Home is a CPU only project ... I just need 25k to make one of my stat goals (#2 on team XS for number of projects w/ 25K plus)

Quote:

---

Docking@Home is a project which uses Internet-connected computers to perform scientific calculations that aid in the creation of new and improved medicines. The project aims to help cure diseases such as Human Immunodeficiency Virus (HIV). Docking@Home is a collaboration between the University of Delaware, The Scripps Research Institute, and the University of California - Berkeley. It is part of the Dynamically Adaptive Protein-Ligand Docking System project and is supported by the National Science Foundation.

---

As team captain for GPUgrid, I'd just like to point out that unlike every other project XS is involved with, we are at #1 position there. So if you have an Nvidia card (500 series or better) come join our success & help keep us there :up:

Quote:

---

Originally Posted by stoneageman

As team captain for GPUgrid, I'd just like to point out that unlike every other project XS is involved with, we are at #1 position there. So if you have an Nvidia card (500 series or better) come join our success & help keep us there :up:

---

Are Noelia tasks finally working there? I had so many issues with those, that I didn't want to continue until it's fixed...

I've not seen a Noelia task for 3 weeks :)

Quote:

---

Originally Posted by stoneageman

As team captain for GPUgrid, I'd just like to point out that unlike every other project XS is involved with, we are at #1 position there. So if you have an Nvidia card (500 series or better) come join our success & help keep us there :up:

---

Stone man - has Nvidia resolved the one WU per GPU issue?

P.S. - nice avatar!

Quote:

---

Originally Posted by stoneageman

As team captain for GPUgrid, I'd just like to point out that unlike every other project XS is involved with, we are at #1 position there. So if you have an Nvidia card (500 series or better) come join our success & help keep us there :up:

---

As soon as I get some personal stats settled, I'll be back with a 670 + 660Ti combo :up:

Quote:

---

Originally Posted by Gandalf

Stone man - has Nvidia resolved the one WU per GPU issue?

---

you run long tasks with 100% utilization, should not be a problem there

will see if i have money to buy a couple of titans, most likely not

Computing considered - Titan should be the biggest-baddest kid in town at the moment?

Quote:

---

Originally Posted by Gandalf

Stone man - has Nvidia resolved the one WU per GPU issue?

P.S. - nice avatar!

---

From the testing done by others, overall there is nothing to be gained at present by running more than 1 wu per gpu.

Titan doesn't work for most projects yet. Seems that current CUDA 4 and 5 apps don't like it and they were not fixed yet. So for example you won't be able to run GPUGrid on Titans.
Though I think Einstein@Home BRP4 app (CUDA 3.1) is able to run on Titans, however the performance is not so great.

Here's a performance comparison of different GPUs for E@H: http://www.dskag.at/images/Research/...rmancelist.pdf

Quote:

---

Originally Posted by Mumak

you won't be able to run GPUGrid on Titans.

---

well then http://youtu.be/r-bA9FYB8HY?t=49s sums it up for me pretty well:P

I checked GPU usage when running MW@H tasks and it seems there might be some room left for running multiple WUs. So I changed it to run 2 WUs/GPU, more doesn't seem to make sense. The app name for app_config is "milkyway".

E@h seems to ask 0.5 cores to run well so only running 2 (wcg > everything after all:))

supercoolin is tearing up the milkway!!!!

My 2x7950 are doing ~560,000 pts/day there :)

Found another - anyone knowedgeable about Primegrid? Here's all I got:

Primegrid - Cryptography(Trying to find out various different primes in an attempt to improve encryption algorithms, as well as other algorithms)

I ran it for a little while. It's a small budget affair but the work has some important applications if you're into cryptography.