Posts by HAL9000

1) (Message 5377)
Posted 26 Apr 2017 by HAL9000
Post:
Basically Credits are simply Kudos for your participation in this or other projects, worth absolutely nothing in real terms, lol. The main aim of this and other projects is for you to GIVE your spare computer/android device time in the cause of science, we do it to aid the projects in the search for answers in science.

Also once you get enough credits you can trade them in for the BOINC toaster.
2) (Message 4931)
Posted 29 May 2016 by HAL9000
Post:
wow, this looks very good.. incredible performance, perfect consumption and good pricing..

GTX 660Ti does 2460 SP GFLOPS pulling 150watts max
GTX 770 does 3213 SP GFLOPS pulling 230watts max
GTX 780 does 3977 SP GFLOPS pulling 250watts max
GTX 970 does 3493 SP GFLOPS pulling 145watts max
GTX 980 does 4612 SP GFLOPS pulling 165watts max
GTX 1070 will do 6500 GFLOPS SP pulling 150 watts and MSRP of $379.
GTX 1080 will do 8227 GFLOPS SP pulling 180 watts and MSRP of $599.

You can replace a 660Ti, 770 and 780 with 2x GTX 1070 and gain 34% more GFLOPS and use less than 50% of the power. :)


Those rating are single precision (FP32) and are useless here, since Asteroids uses double precision(FP64)...

AFAIK, a gtx 1080s FP64 performance is 1/32 FP32 peak flops...
A GTX 660TIs FP64 performance is 1/24 FP32...


In checking https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units it looks like:
Fermi DP = SP/8
Kepler DP = SP/24
Kepler Titans DP = SP/3
Maxwell DP = SP/32
Pascal DP = SP/32

The 10 series looks to be the first to have a DP GFLOPs/Watt ratio greater than 1.
It's a shame there isn't an app for Radeon GPUs.
GTX 660Ti 0.68 DP FLOPs/Watt
GTX 770   0.58 DP FLOPs/Watt
GTX 780   0.66 DP FLOPs/Watt
GTX 970   0.75 DP FLOPs/Watt
GTX 980   0.87 DP FLOPs/Watt
GTX 1070  1.21 DP FLOPs/Watt
GTX 1080  1.03 DP FLOPs/Watt
R9 390X   2.69 DP FLOPs/Watt
3) (Message 4924)
Posted 26 May 2016 by HAL9000
Post:
wow, this looks very good.. incredible performance, perfect consumption and good pricing..

GTX 660Ti does 2460 SP GFLOPS pulling 150watts max
GTX 770 does 3213 SP GFLOPS pulling 230watts max
GTX 780 does 3977 SP GFLOPS pulling 250watts max
GTX 970 does 3493 SP GFLOPS pulling 145watts max
GTX 980 does 4612 SP GFLOPS pulling 165watts max
GTX 1070 will do 6500 GFLOPS SP pulling 150 watts and MSRP of $379.
GTX 1080 will do 8227 GFLOPS SP pulling 180 watts and MSRP of $599.

You can replace a 660Ti, 770 and 780 with 2x GTX 1070 and gain 34% more GFLOPS and use less than 50% of the power. :)

Do you happen to have any run time numbers for any of those GPUs? I found my EVGA 750 Ti FTW ran tasks at a rate of ~90 min.
4) (Message 4587)
Posted 5 Aug 2015 by HAL9000
Post:
I'm sure these questions have been answered before, but the search feature doesn't seem to be finding the answers very well, so hopefully someone can enlighten me.

Why are there so many WUs that error out with "Cancelled by Server" status? I've had about a dozen in the last week. Yeah, I know that's not really a lot considering, but I am a curious people...

I'm assuming that the WUs crunch with a buddy cruncher, which is possibly why I have over 200 tasks in the Validation Pending cue, but I would like to confirm that, if possible, as I've got some WUs that have been pending validation since 19 July...

Thanks for any assistance, and I apologize for bringing this up for quite possibly the billionth time.

Astroid@home uses a minimum quorum of 2 for its work. So if one of those two misses their deadline then a new task is generated & is sent to a 3rd host for processing. If the host that missed their deadline returns the work after the 3rd host has received the new task, but before their return their result. Then the server cancels the third task.

Personally I'm of the opinion that if you miss the deadline then the server should tell you "sorry I'm not accepting that". Rather than shafting the 3rd person.
5) (Message 4572)
Posted 17 Jul 2015 by HAL9000
Post:
Thanks for answering. Yes it is running Android. The only project that sends it WU's at the moment is Collatz. So, if they can manage it why can't other projects?


If you really want it that bad, learn to code. ;) I don't say to be spiteful, but that's really the best way to get it (unless you know someone who you can get to do it).

Even SETI@home with multiple applications still has to balance the working on the clients. In fact, some of the clients you use now evolved from 3rd party optimized clients. It still doesn't support every possible configuration out there.

There's two things that need to happen:

1. You have to have at least one of the device in the configuration you want to make a client for.

2. There has to be someone who knows how to compile and test the client on that configuration.


Even if you donate a device to the project, someone still has to create it and often times that is done by volunteers, such as students. That person can't focus just on a small segment. Also, if that person leaves, there is nobody to maintain the software.

SETI is under Berkeley's banner, but it doesn't mean that it gets the full force of Berkeley's funds. Most of these projects are under the banner of an educational institution. Some consider themselves lucky just getting hosting services from their institution...

Indeed much of the code development for the science apps on SETI@home has been done by the lunatics group. An effort was started to test Android x86 apps on SETI@home Beta. However few people have devices for testing & recently time has been directed to other apps that are more widely used.
6) (Message 3753)
Posted 8 Nov 2014 by HAL9000
Post:
It is probably related to the 2nd server that was recently added. A hiccup or two when messing around with the BOINC back end is not out of the question.