AQUA

tullio
tullio
Joined: 22 Jan 05
Posts: 2,087
Credit: 50,598,154
RAC: 22,208
Topic 194107

I've joined this project since I am interested in quantum computing. The AQUA app uses very little memory and runs well on my Linux box. Unfortunately, it does not do any checkpointing and so, once it grabs your CPU, it won't let it until it finished, I crunched a WU for 16 CPU hours straight, then I had to put No new tasks in order to allow my other 6 projects to run.
Tullio

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4,049
Credit: 224,526,366
RAC: 23,502

AQUA

Might be wrong project board?

BM

BM

tullio
tullio
Joined: 22 Jan 05
Posts: 2,087
Credit: 50,598,154
RAC: 22,208

RE: Might be wrong project

Message 89302 in response to message 89301

Quote:

Might be wrong project board?

BM


Sure, but I wanted to spread word about this not so BOINC project, which is sponsored by a private firm.

tullio
tullio
Joined: 22 Jan 05
Posts: 2,087
Credit: 50,598,154
RAC: 22,208

I've signaled this behavior

I've signaled this behavior of AQUA to dr. Anderson. I post here his reply because it may interest others:

Making applications restartable can be quite a bit of work.
I'm sure AQUA will do this eventually.

When an application (like AQUA) doesn't checkpoint,
BOINC will run it to completion.
However, after this happens BOINC won't get more work from AQUA for a while,
so that the work done for the various projects balances out in the long run.

David

mikey
mikey
Joined: 22 Jan 05
Posts: 8,428
Credit: 642,316,340
RAC: 149,668

RE: I've signaled this

Message 89304 in response to message 89303

Quote:

I've signaled this behavior of AQUA to dr. Anderson. I post here his reply because it may interest others:

Making applications restartable can be quite a bit of work.
I'm sure AQUA will do this eventually.

When an application (like AQUA) doesn't checkpoint,
BOINC will run it to completion.
However, after this happens BOINC won't get more work from AQUA for a while,
so that the work done for the various projects balances out in the long run.
David

When Boinc does this it is called Long Term Debt. Over time all your applications will follow your settings for each project to use whatever percentage you set. However over the short term it will be out of whack, as you are noticing. Long term can be in the months if you run several projects on the same pc.

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

Though the project is

Though the project is sponsored and run essentially for the benefit of a commercial company the point of the Aqua project is to prove that quantum computing can occur and that the quantum CPU that the firm is making is in fact doing quantum computing.

In that this will make a, ahem, quantum leap in computing ability in some instances I would lean towards supporting the project regardless.

As to the no checkpointing ... I set my switch interval to 720 minutes to try to complete tasks in their entirety (unless the total run time is over 12 hours) to avoid having 30-40 partially started tasks ...

Sadly, BOINC manager still manages somehow to mess things up and worse has a bad habit of starting a new task for a project when there is one waiting that has only minutes to go before it is done. I have long asked for a preference setting to have BOINC Manager *NOT* switch tasks if the task to stop has an estimated end time less than some limit ... (24 hours? :))

Anyway ... just because the project is run for and by a commercial company does not make it completely evil ... but ... YMMV and a logical choice not to support those projects ...

tullio
tullio
Joined: 22 Jan 05
Posts: 2,087
Credit: 50,598,154
RAC: 22,208

I have been interested in

I have been interested in quantum computing since 1996 when I coauthored a paper on it with Giuseppe Castagnoli on the Italian version of MIT's Technology Review. I started the AQUA project also for this reason and was surprised to find it does not do any checkpointing. But I am still doing it, since luckily, their WUs often converge to a solution in a short time.
However the adiabatic approach to quantum computing is not the only one. See this article in the Physics on line magazine of the American Physical Society:
Physics.
In its bibliography you'll find the name of my friend Castagnoli and also the name of prof. Franco Rasetti of the Torino Polytechnic who wrote several papers with Castagnoli. So D-Wave is not my only source of information on this subject.
Tullio

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

RE: I have been interested

Message 89307 in response to message 89306

Quote:
I have been interested in quantum computing since 1996 when I coauthored a paper on it with Giuseppe Castagnoli on the Italian version of MIT's Technology Review. I started the AQUA project also for this reason and was surprised to find it does not do any checkpointing. But I am still doing it, since luckily, their WUs often converge to a solution in a short time.
However the adiabatic approach to quantum computing is not the only one. See this article in the Physics on line magazine of the American Physical Society:
Physics.
In its bibliography you'll find the name of my friend Castagnoli and also the name of prof. Franco Rasetti of the Torino Polytechnic who wrote several papers with Castagnoli. So D-Wave is not my only source of information on this subject.
Tullio

And I wish I was that smart ... :)

Every time I read about quantum mechanics it makes my head hurt ... of course, I always loved the two slit experiment ... until that one guy came up with that wrinkle that seems to make trouble for the particle theory ... I read about it in Scientific American a couple years back and I know that there is some stuff on the web about it, Afshar ... that is his name ...

Though I have seen some posts that claim that he proves nothing and his experiment questions nothing ... sadly I can't follow the logic of the nay sayers so I don't know who is right ...

tullio
tullio
Joined: 22 Jan 05
Posts: 2,087
Credit: 50,598,154
RAC: 22,208

Quantum mechanics has always

Quantum mechanics has always divided physicists. Some of its pioneers like Einstein, Max Planck, Louis de Broglie and even Schroedinger have never accepted the Copenaghen interpretation put forward by Niels Bohr. Then Einstein about 1935 gave it a shock with the Einstein-Podolsky-Rosen gedanken experiment, which he considered a paradox. Alas, it is not, and the entanglement concept deriving from it is used now in quantum cryptography, especially in quantum key distribution, which was performed on a distance of 150 km. Whatever one can think of quantum mechanics, it works.
Tullio

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.