Eric has announced that the testing of SETI@home v7 will commence shortly on SETI Beta.

Does anyone have the inside track on what features the new application will have? Will it, for example, analyse the data in a new or better fashion? And, once released, will it keep the Lunatics crew busy creating a whole new generation of optimised apps?

---

Good to know.
So i´ll turn beta back on soon.

---

With each crime and every kindness we birth our future.

.. i´m running it like hell for DAYS .. greetings! ;-)

Lunatics already has access to the app.

---

Lunatics already has access to the app.

And has contributed some refinements to the development of what will become the stock app.

Eric has announced that the testing of SETI@home v7 will commence shortly on SETI Beta.

Does anyone have the inside track on what features the new application will have? Will it, for example, analyse the data in a new or better fashion? And, once released, will it keep the Lunatics crew busy creating a whole new generation of optimised apps?

Sources are available as a branch. An Autocorrelation search is being added, analyzing the chirped data at full 9765.625 Hz. bandwidth (102.4 microsecond delay resolution).

                                                                 Joe

What does that mean in layman's terms? Are there any other large advancements to the science and/or app in this major release?

a change from v6 to v7 would denote a major change

the Correlation would be a factor in that...nearly 7 times the current bandwidth searched i believe.

---

S@h BETA - News - Stopped SETI@home work generation. - Message 40183

(...)
The only major change is the addition of a new signal type which is detected using an autocorrelation analysis. Some SETI researchers have theorized that ET would send a signal waveform, and then send a copy of the same waveform, but slightly delayed (but before the first waveform had ended). This type of waveform is easy to detect with an autocorrelation analysis, and could contain a data stream rather than just being a narrow band signal that doesn't contain data.


But, we will see 6.03, 6.08 - 6.10 CUDA still here at S@h?
Or replace S@h V7 all 6.x apps?

The S@h V7 app will be only for CPU? Maybe one day also CUDA?

---

And will ATi see some official application support from the project? And will it run on machines that run BOINC as a service on Windows Vista/7?

But, we will see 6.03, 6.08 - 6.10 CUDA still here at S@h?
Or replace S@h V7 all 6.x apps?

The S@h V7 app will be only for CPU? Maybe one day also CUDA?

Good question. I think we are looking initially at a replacement for 6.03/CPU. Then somebody will have to adapt the code for stock cuda.

As far as I understood the original plan was to have it running side by side with current cuda and optimized applications.

I suppose Jason and Raistmer can integrate the new code fairly easily into their current work, but getting the whole bunch of AK_v8 apps updated...

---

Carola
-------
I'm multilingual - I can misunderstand people in several languages!

I would like to humbly request AVX support in v7, if possible. We are seeing more and more people who are getting the new sandy bridge CPUs, and it's a pretty big performance jump if we use those AVX instructions vs regular apps. It's something that will need to be put into the future apps anyway, so now is as good of a time as any. :)

We had a thread dedicated to AVX a while back, but it kinda fizzled out. I'm running Win 7 x64 SP1, which has AVX support (as do most newer linux distributions). SP1 goes public within the next month, and we're sure to get more people asking about it.


Thanks for your consideration and a HUGE thanks to all the programmers who put in the time and effort to keep things up-to-date with the project. You guys really don't get the thanks you deserve.

---

-baron_iv
Proud member of:
GPU Users Group

I would like to humbly request AVX support in v7, if possible. We are seeing more and more people who are getting the new sandy bridge CPUs, and it's a pretty big performance jump if we use those AVX instructions vs regular apps. It's something that will need to be put into the future apps anyway, so now is as good of a time as any. :)

The developers are certainly aware of AVX, and it will come during the lifetime of v7 - but please don't ask too much on day one.

The first priority will be to ensure that the applications run on current, mainstream hardware, and produce accurate results which validate against the stock applications.

Then, as Carola says, it will be necessary to upgrade the CUDA applications, to support what has become a significant bloc within the user base.

And remember, the developers - all too few of them - are currently still trying to apply the finer details of the last big technology leap - Fermi - to a SETI application which started life with no thought of parallel processing. Unless we can recruit programmers who already have some knowledge of the Sandy Bridge processor - which will be a challenge in itself - progress is likely to be slow and frustrating. Sorry, but that's the way it is.

I pray that when CUDA/Open CL/ATI GPU comes that it also comes with support for OSX not just Win/Linux. I have a hard time with the fact the project has no GPU support for OSX.

---

I would like to humbly request AVX support in v7, if possible. We are seeing more and more people who are getting the new sandy bridge CPUs, and it's a pretty big performance jump if we use those AVX instructions vs regular apps. It's something that will need to be put into the future apps anyway, so now is as good of a time as any. :)

The developers are certainly aware of AVX, and it will come during the lifetime of v7 - but please don't ask too much on day one.

The first priority will be to ensure that the applications run on current, mainstream hardware, and produce accurate results which validate against the stock applications.

Then, as Carola says, it will be necessary to upgrade the CUDA applications, to support what has become a significant bloc within the user base.

And remember, the developers - all too few of them - are currently still trying to apply the finer details of the last big technology leap - Fermi - to a SETI application which started life with no thought of parallel processing. Unless we can recruit programmers who already have some knowledge of the Sandy Bridge processor - which will be a challenge in itself - progress is likely to be slow and frustrating. Sorry, but that's the way it is.

Fair enough. I just signed up to help test the new apps on several of my computers, hopefully I can do something to help.

It's a shame that the companies who make this hardware don't support projects like this. I mean a guy who knew what he was doing could probably whip up a working AVX app in no time at all. The same is true with nvidia and AMD/ATI with Fermi and APP (or whatever they're calling it this week).

I didn't mean to push you guys, I know you're doing the best you can with what you've got, and I do appreciate it very much, as I'm sure many others do.

---

-baron_iv
Proud member of:
GPU Users Group

Eric has announced that the testing of SETI@home v7 will commence shortly on SETI Beta.

Does anyone have the inside track on what features the new application will have? Will it, for example, analyse the data in a new or better fashion? And, once released, will it keep the Lunatics crew busy creating a whole new generation of optimised apps?

Sources are available as a branch. An Autocorrelation search is being added, analyzing the chirped data at full 9765.625 Hz. bandwidth (102.4 microsecond delay resolution).

                                                                 Joe

It's good to see so much interest in v7. I see Eric has also proposed a Astropulse Alpha on his Beta Applications page.

Michael

This is a great idea but we have 10 years of work without this search done. I'm wondering if a separate app should be created to parallel the current work instead.

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

This is a great idea but we have 10 years of work without this search done. I'm wondering if a separate app should be created to parallel the current work instead.

If I read Eric's message (40171) now again a few times at the S@h BETA forum.. I guess we will see S@h Enhanced and S@h V7 simultaneously.

Why we should cancel S@h Enhanced and use only S@h V7 if a few people think, ET could/would use (also) this way?
The work we did with S@h Enhanced was nonsense?

Maybe the usage/name of S@h V7 isn't a well idea. It confused us.
Maybe an other name would be a better idea? Maybe S@h Autocor (for autocorrelation analysis), or something other?

---

from what I read a few days ago they are not sending any more 6.XX work then start fresh with the v7 so that there won't be to much confusion as to whats coming out

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

from what I read a few days ago they are not sending any more 6.XX work then start fresh with the v7 so that there won't be to much confusion as to whats coming out

I have no knowledge about the BETA server.
Maybe they can generate/serve only S@h Enhanced and AP simultaneously. A 3rd WU type isn't possible?

S@h Enhanced is well tested currently. So maybe not needed to send out currently Enhanced WUs at BETA?

Ohh well.. one well meant message of Eric at BETA and a lot of guessing and confusing.

;-)