Gravitational Wave S6 Directed Search (CasA) ending
Submitted on 13 Aug 2014 12:41:12 UTC
We are nearing the end of the "S66CasA GW run". This means that only relatively few tasks of possibly remote frequency bands are available for this application, possibly resulting in an increased download volume per task. If you want to avoid this, you may opt-out of the "Gravitational Wave S6 Directed Search (CasA)" search in your Einstein@Home preferences on the web site.
BM
Language
Copyright © 2024 Einstein@Home. All rights reserved.
Comments
Gravitational Wave S6 Directed Search (CasA) ending
)
Good day!
Refuse I will not let it continue.
Is another gravitational wave
)
Is another gravitational wave search planned after the current one ends?
I guess it's time to start
)
I guess it's time to start running beta test applications again! Any idea on how much more work needs to be done on those too?
RE: Is another
)
Bernd as already answered that question on the Science forum:
S6CasA is actually part of a larger effort that we internally call "S6Directed", i.e. an analysis of "S6" data targeting single positions (hopefully GW sources) in the sky. We started with the most promising "Cassiopeia A", and will then simply move on to the next target (there are about a dozen on our list). We are also working on improving the analysis code for this, trying to make it more efficient and sensitive.
BM
RE: Is another
)
See here.
BM
BM
Nice, so we wont run out of
)
Nice, so we wont run out of work then :-)
RE: Nice, so we wont run
)
No, we certainly won't. However, the next GW run will take a bit longer to prepare. After S6CasA is finished we will probably have no GW work for a couple of weeks. We currently try to get the next Gamma-ray search (FGRP4) out the door to provide CPU work.
BM
BM
RE: RE: We are also
)
This is my first experience
)
This is my first experience with this sort of message. I signed onto BOINC at the end of July last year and chose your project as well as Rosetta from U of Washington. In just over a year my score is just below 2.5MM points, or whatever the unit of account is. But I do not care about that as I am fortunate enough to have a pretty powerful PC at home (Intel i7 with 4 cores and 8GB of Ram, and I allot 150GB of disk space to BOINC which means there is never a constraint) and am more than happy to allocate its unused cycles to something productive.
Any suggestions you may be able to provide as to how the number crunching I am contributing to goes would be of great interest to me. Anything you can e-mail to me or sent links to would be very much appreciated.
Finally, I am a dual US / UK citizen having spent 14 years working and living in London. I am now living and working in Dana Point, CA and hope to be here until the coroner confirms time of death.
Regards,
Brian
RE: If more "sensitive"
)
I am pretty sure we will include CasA again as one of the targets in the next directed search. We'll probably spend less computing time on it then, but that might be compensated by the improved efficiency / sensitivity of the search code.
We are all eagerly awaiting first data from the advanced LIGO detectors, which should be an order of magnitude more sensitive than "S6". If the data becomes available while the next directed S6 analysis is still running on E@H, we'll probably cut that short and focus on the new data.
BM
BM
Hallo Bernd! Will this new
)
Hallo Bernd!
Will this new "S7"(?) task than also available for crunching on GPU?
Kind regards and happy crunching
Martin
RE: We are nearing the end
)
Hopping more CPU tasks,
)
Hopping more CPU tasks, because also the FGRP3 search progress is finishing...And some expectatives to find some gravitational wave in the recently s6 series?
Grettings from Spain
FGRP4
)
FGRP4 starts:
http://einsteinathome.org/node/197685
I wish to opt-out of the
)
I wish to opt-out of the "Gravitational Wave S6 Directed Search (CasA)" search in your Einstein@Home preferences on the web site.
Just edit your preferences on
)
Just edit your preferences on the website to select those searches you want to run and remove the selection marker from those searches you don't want to run. It's all under your control.
Cheers,
Gary.
Gary, I've only been
)
Gary,
I've only been contributing my PCs available cycles for just over a year now and I wonder whether there is a way to assess my relative contribution to the project. I have a 4+ year old Windows 7 Pro PC with 8GB of ram and have allocated 150 MB of disk for Boinc projects. Needless to say, there has never been any real problem as BOINC manages contention reasonably well.
My CPU is an i7 with an AMD GPU. Normally there are 7 processes underway split between Einstein @ Home and unless Rosetta @ Home. So, just under 13 months into this gig and my point total is a bit over 2.5 Million. How does this stack up?
Thanks for your reply,
Brian
RE: ... I wonder whether
)
Hi Brian,
The fact that you are still contributing after more than a year, puts a high value on what you are doing.
You should not get too bogged down in 'relative' contribution to a project. It's hard to make comparisons with others since some are contributing to many different projects and some to only one or two. Some projects are quite miserly with their credit awards and some are so 'over-the-top' as to make comparisons quite useless. Some people choose to create 'farms' and many others just use the computer they have.
At the end of the day, every contribution is valued by the project staff. Projects like the ones you support couldn't exist without the many 'single machine' people, just like you, making their contribution.
Since your host is an i7, it has plenty of power to run the various CPU-based sub-projects. I notice you are also contributing to sub-projects that can use your GPU. This is all fine and there's no need to change anything. However, if the thinking behind your question was to work out ways to improve your work output, you could consider a GPU upgrade that would make a significant difference. If you are looking for suggestions, just ask. In a new thread would be best so as not to hijack this news thread.
Cheers,
Gary.
RE: Gary, I've only been
)
Many of us started that way Brian and as Gary mentioned you could update your video card to a GPU for our project (and most likely a PSU upgrade)
Once you get started here running the GPU tasks you will start thinking of ways to add more to your host farm.
I am getting close to starting my 10th year here and for most of those years I just had a few CPU hosts here way back to when I got my first dual-core host.
And after a couple years of reading about GPU tasks and members here doing those tasks got me hooked and soon after getting my first GeForce 550Ti *and new PSU) I had to do that with all 7 of my hosts and trying to figure out the tricks.
(one thing I started doing was using the same newer Power supply to power 2 separate GPU cards in 2 separate hosts so I didn't have to buy and install the new power supplies for each computer)
And the one you have here has enough RAM so you can look for a good deal on the video card you want and those PSU's are easy to find with plenty of power (and lifetime warranty) at a fair price.
It would be almost impossible to have a farm with a RAC like Gary has here
Of course it will always work better at the Einstein Project
(and yeah a different thread)
RE: Hallo Bernd! Will this
)
Of course!
I can assure You it will. It's only a matter of time.
CUDA version will be released first I think, because it is pretty simple to write and debug.
I cannot really estimate it, because I have not seen *the code* (in the new app) which can and shall be parallelized, yet.
I can only guess it will definitely take some time - just like with any GPU app on any project.
(probably some 3 to 6 months for the CUDA ver. , and I guess around 10 months for the openCL version).
[/b]
BRISINGR-II: PRIME X370-PRO,AMD'Zen 1800X 3.7/4.1, 2x8 DDR4 G.Sk.3602@2400,Asus STRIX GTX1070 DirectCUIII 8GB,*2017-04-08
BRISINGR: nb ASUS G73-JH,i7 1.73,4x2 DDR3-1333 CL7,ATi5870M 1GB,*2011-02-24
RE: Will this new "S7"(?)
)
Well, "S7" is (at least) three Einstein@Home "runs" into the future, I have to admit that my crystal ball is getting very blurry focusing in that distance.
The S6Bucket follow-up (internally named "S6BucketFUP") run will probably done with an only slightly modified application code that is currently used for S6CasA. The modifications won't affect the core calculation code, so until further notice I assume that the current S6CasA OpenCl code will remain usable during that run.
For the planned S6Directed run then we are developing a new, faster core calculation engine in cooperation with Christope Chocquet, to avoid design decisions that would hinder a GPU port. According to current plan Christophe will also make a GPU (OpenCL) port of the application, but not before we are confident that the CPU code does what we expect. So most likely the S6Directed run will start with CPU app versions only, and we'll have GPU app versions a little later.
In general I would expect this for every new run that uses a new application code - get it working scientifically reliable with CPU apps first, then make GPU versions later.
BM
BM
RE: RE: Will this new
)
It would be great to have OpenCL app for S7 run quicker than for S6 run. My Radeon R280X can process 4 at the same time, and calculate them on avarage ~15 times faster (from 10-20 times faster, up to 50+ times faster when compared to old machines) for a net increase in electrical efficiency of between 4-10 times. This saves us cost and speeds up calculations.
Nice run for the S6 (CasA)
)
Nice run for the S6 (CasA) which is ending, can't wait for the replacement.
RE: Well, "S7" is (at
)
EEEk......I just made the mistake of using my blurry crystal ball to see that it has been costing me $60 per month to run just my computers here......so I tossed my crystal ball up in the air and.......ok that made a mess in my yard
RE: RE: Will this new
)
Thanks for the updates. I am looking forward to the S6BucketFUP data and applications. I recently switched all my active GPUs to S6CasA and the application has been running stable.
I would like to ask you to
)
I would like to ask you to make new applications consume smaller amount of memory than they use these days. I have machines with 512Mb of RAM with internal VGA card that uses 8-32 Mb of these. And I already can not run S6 there. This was the last Einstein app that was able to run on them. Of cause I will try to increase amount of memory for these rigs. But currently they are not available for me.
RE: I would like to ask you
)
This is certainly not the direction we will go for Gravitational Wave search apps. We try to have at least one "low profile" application running, but not necessarily and in the future rather unlikely the GW app.
The new FGRP4 (Gamma-ray) app should have a (slightly) smaller footprint than the previous FGRP3, until this is out officially I'd recommend to try BRP4 (Radio Pulsar).
BM
BM
Wonderful Scott ;') Lets
)
Wonderful Scott ;')
Lets grus the Bays! Oh yes this is FUN!
WTF? Why do we get the same
)
WTF? Why do we get the same bloody messages marked as NEW time after time? ??
Let's keep this conversation
)
Let's keep this conversation civilized, please . . .
Thanks for merging my computation total on my 1 computer running BOINC.
Using both CPU (I7-2600) and 2 GPU's (HD5870).
Messages can be marked new if
)
Messages can be marked new if a new message was added that seemed inappropriate and was hidden. The message is still new, but you do not see it.
Also, after so many months, BOINC automatically marks some messages as unread. This is BOINC's issue, not Einstein's. Go to the Development boards and as why there.