Einstein@home News Q1 2011

[derk]

Team Admin News

Welcome to the first edition of the OcUK Einstein@home News
Any suggestions for additions/alterations to the news are most welcome.

The following inactive memberships (zero credit) have been terminated

Leigh Gilbert has been removed
Barry Bott has been removed
Åmbrosius has been removed
sefese has been removed
nick name nick has been removed
abb has been removed
Capnqweel has been removed
madman has been removed
Daniel R. Gigot has been removed
chriswebuk has been removed
DATAjammer has been removed
net_finding has been removed
SilentBobOCUK has been removed
jim Tenzin has been removed
Andrew Johnson has been removed
Jim has been removed
rizo has been removed
Gary has been removed
Phil_Duncan has been removed
mattcreese has been removed
kategomm has been removed
wigan meat pie has been removed
Twiss has been removed
Niko has been removed
Brumratt has been removed
Aaron has been removed
simianbenzoate has been removed
Iain has been removed
hikaru452 has been removed
Daniel has been removed
edd has been removed
David has been removed
Rodz37 has been removed
Adele has been removed
mhapti has been removed
skutterpunk has been removed
Ben Brooks has been removed
Michael has been removed
creed has been removed
not-the-infidel has been removed
John Stephen Reid has been removed
ca9phoenix has been removed

28/01/2011

quadcorele11 has been removed
GB097690 has been removed
Ronz has been removed
Nemesis has been removed
Graham Anderson has been removed
sorcerer has been removed
Con has been removed
THEHEAP'S has been removed
Princeofdevils has been removed
Dunk has been removed

For your information all the above had mostly joined years ago and not returned any results but you are welcome back again as an active participant.
​

Project News

New Einstein@Home Radio Pulsar Search and NVIDIA GPU Code

Einstein@Home is beginning a new round of searching for radio pulsars in short-orbital-period binary systems.
This is accompanied by the release of a new application (called BRP3). The new application is particularly efficient on NVIDIA Graphics Processor Cards (up to
a factor of 20 faster than the CPU-only application). In addition, when running on an NVIDIA GPU card, this new application makes very little use of the CPU
(typically around 20% CPU use when the GPU is devoted to Einstein@Home).

The NVIDIA GPU application is initially available for Windows and Linux only. We hope to have a Macintosh version available soon. Due to limitations in the
NVIDIA drivers, the Linux version still makes heavy use of the CPU. This will be fixed in Spring 2011, when a new version of the NVIDIA Driver is released.
Many thanks to NVIDIA technical support for their assistance!

Because we have exhausted the backlog of data from Arecibo Observatory, this new application is being shipped with data from the Parkes Multibeam
Pulsar Survey (from the Parkes Radio Telescope in Australia). In the next weeks we expect to also start using this new application on fresh Arecibo data taken
with the latest 'Mock Spectrometer' back-end.

Bruce Allen
Director, Einstein@Home 20 Jan 2011 15:57:08 UTC


Older News - Einstein@home users discover rare star

http://www.bbc.co.uk/news/science-environment-10959590​

Signiture image's

Sig's for forum usage should you require one, if you need any help setting it up all you need do is ask

Useful Links


nVidia Drivers http://www.nvidia.co.uk/Download/index.aspx?lang=en-uk

PSU Calculater http://www.coolermaster.outervision.com/advance.jsp

Server Status http://einstein.phys.uwm.edu/server_status.html
​

 

[BaBa]

Thanks for the news Mr Halz

Anyone running Nvidia gpu's the new cuda app is a LOT better than the old one it runs the gpu harder and has less cpu input

 

[TheTross]

Ooh, Einstein news! Cheers halz

I can also say that the new app is more efficient than the older one but there is currently an issue with validation in that, in some instances, tasks completed on different hardware (i.e. CPU vs GPU) will not validate against each other, even though both tasks are actually valid. As far as I know though, the Einstein admin are aware of the problem and are looking into it.

 

[kolala]

Well done Mr Halz, reminded me that my einstein progress has been lacking recently. I've fired it up again. All we need now is E@H tiffys

 

[Meridian]

Ah - my main backup project when SETI is playing silly beggars. Sadly it remains that though.


M

 

[derk]

Well done Mr Halz, reminded me that my einstein progress has been lacking recently. I've fired it up again. All we need now is E@H tiffys

I think I will issue Tiffys direct for the time being......gotta make them first mind

Ah - my main backup project when SETI is playing silly beggars. Sadly it remains that though.

M

Using Einstein as backup is good enough for me, your doing pretty well at it though for a backup project.

 

[derk]

OK, I have done the newsletter, made the sigs/wallpaper and I have just emailed approximately 270 members.
All the emails were directed to this post so come on peeps, it wouldnt take many more members to raise our output to the highest in the UK

WE NEED YOU​

 

[Biffa]

It just doesn't grab my attention this project. Dunno why, just went back for a look and a read of the forums and just couldn't get excited.

 

[sculptor]

It just doesn't grab my attention this project. Dunno why, just went back for a look and a read of the forums and just couldn't get excited.

Ah comon Mr. Biffa, Im only running a dual core @ 3.6GHz and Im doing pretty well, shirley you could spare a couple of cores.
Apart from Climate Pred its the only DC project I know of that has had a positive outcome from everyones contribution (not knocking the other projects).
Tom

 

[Biffa]

Really? Whats the point of finding pulsars?

 

[derk]

Really? Whats the point of finding pulsars?

The study of pulsars has resulted in many applications in physics and astronomy. Striking examples include the confirmation of the existence of gravitational radiation as predicted by general relativity and the first detection of an extrasolar planetary system.

The discovery of pulsars allowed astronomers to study an object never observed before, the neutron star. This kind of object is the only place where the behavior of matter at nuclear density can be observed (though not directly). Also, millisecond pulsars have allowed a test of general relativity in conditions of an intense gravitational field.

In order for the effects of general relativity to be measurable with current instruments, pulsars with orbital periods less than about 10 years would need to be discovered, such pulsars would orbit at distances inside 0.01 pc from Sgr A. Searches are currently underway, at present, five pulsars are known to lie within 100 pc from Sgr A

Source: http://en.wikipedia.org/wiki/Pulsar

Ah comon Mr. Biffa, Im only running a dual core @ 3.6GHz and Im doing pretty well, shirley you could spare a couple of cores.

Calm down, there's no need to call him Shirley

 

[Spleenus]

OOh fantastic, good to see some stats OCuk based. Started doing Einstein again a couple of months ago after my interest in physics and astrophysics was re-kindled.

 

[BaBa]

Ive kind of jumped in with both feet, hope to be heading for uk No.1 for rac and maybe top 10 in the world

 

[derk]

OOh fantastic, good to see some stats OCuk based. Started doing Einstein again a couple of months ago after my interest in physics and astrophysics was re-kindled.

Thanks Spleenus...........and you never know - it could be you

http://www.bbc.co.uk/news/science-environment-10959590

Ive kind of jumped in with both feet, hope to be heading for uk No.1 for rac and maybe top 10 in the world

I wouldnt expect any less from you, your contribution to the team is simply amazing S'ah

 

[Spleenus]

I've noticed there's an Nvidia client now available. Does anyone have any comparable stats of how long to complete work units compared to cpu? I looked on the main einstein@home and it mentions it's something like upto 20x faster than the CPU application. Wonder if we'll see some nvidia farms!

 

[derk]

I've noticed there's an Nvidia client now available.

Did you not read my news post , ok so I added it later

Does anyone have any comparable stats of how long to complete work units compared to cpu? I looked on the main einstein@home and it mentions it's something like upto 20x faster than the CPU application.

Havent seen any so I am going to try it myself..............in a big way

 

[TheTross]

@Spleenus:

On my GTS250 (shaders @ ~1900) it takes around 70 minutes to do one of the new BRP3 WUs. Briefly looking through my results, one of the WUs was validated against one that was done on an i7 920, which took 10.5 hours. (Granted though that that particular chip can potentially do eight at a time, so eight in 10.5 hours doesn't sound too bad...)

Remember that the current BRP3 app doesn't use 100% of the GPU, so you can create an app_info file to allow more than one to run on a card simultaneously. This obviously slows the crunch times down per WU but gives a higher output overall, so that might well be worth giving a try. (Point in case: I can do three in about 160 minutes if done this way, as opposed to three in ~210 minutes if run singly.)

 

[BaBa]

Ive been running one of my 460's on an i7 860 rig clocked at 3.5 and it was doing 53-54mins per wu.

When i moved the card into an AMD 1055t rig also clocked at 3.5 times are 70-75 mins per wu.

Im thinking its motherboard memory bandwidth holding it back as the AMD rig is on ddr2 and the intel ddr3 but i would like to know your thoughts

It seems the new cuda app is still very cpu bound but lots better than the old one also dont think its well optimized for fermi.

TT what motherboard is your GTS250 in and how much memory does the card have?

 

[TheTross]

My setup:

Asus P5Q-E
Q9550 @ 3.4
4GB DDR2 (although only 3.25GB seen by 32-bit XP Pro)
GTS250 1GB Green Edition

Note that the stock shader clock for the GTS250 Green is 1512MHz, but it clocks to just above 1900 before becoming unstable, so it's at just under 1900 to give it a bit of headroom.

It might also be worth mentioning that I got CPU use of about 10% per CUDA app instance, so with three BRP3 tasks running I lost about the equivalent of one core.

 

[BaBa]

Thanks TT ,ive just culled all my g92 cards including two gts250's but they are all 512mb so cant run multiple units on einstein

I think my 460's should get 15.5k each my 560ti's(1000mhz gigabyte) should get 18.5k each these are in my amd rigs and my 260.216 should get 16k but this is in my i7 860 rig.

Comared to this your gts250 is getting 13.5k, dont think thats bad for a g92 card

I think this might force me to upgrade to i7's because a 260 should not be doing better ppd than a 460