Dear Sir's.
I find it difficult to see, how it help Astropulse, files take over 160 hours to complete. To secure a better flow, dereace size to 10 or 15 percent, witch allow older PC'es to operate those files.
I waste more than 160 hours of work, due to one single error in a file. The damage would have been less, if the size and time used, was smaller.
I wrote an e-mail to Matt, to hope for a better future for the Astropulse so more users may begin to search for silversticks, and not only the pure ones of gold. :-)
Yours truly
Leif.

---

You could always change your seti@home preferences on your account page. IF you dont want astropulse make sure only the multibean is checked.

Seti recommends a minimum of a 1.6GHZ CPU, you run a 1.7 and a 1.73. these barely meet the needs for Astropulse. You could try one of the optimized apps for astropulse. These are 3rd party apps that are able to crunch WU's much faster than the regular app.

As far as astropulse being cut into smaller Wu's that is very unlikely to happen.

---

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

You could always change your seti@home preferences on your account page. IF you dont want astropulse make sure only the multibean is checked.

Seti recommends a minimum of a 1.6GHZ CPU, you run a 1.7 and a 1.73. these barely meet the needs for Astropulse. You could try one of the optimized apps for astropulse. These are 3rd party apps that are able to crunch WU's much faster than the regular app.

As far as astropulse being cut into smaller Wu's that is very unlikely to happen.

. . . my 'little' P4 is a perfect example of what can be done [using Jason_G etc Lunatics] Thanks guys [JD included as well]

---

BOINC Wiki . . .

Science Status Page . . .

Hi, yes I will fully agree with this one, using the latest optimized application (v112r{Don't Panic}*); V5.03, you reduce the time from 25 till 45%, crunching an AP WU. (Measured on Intel Processors!)
On a stock Q6600, it takes 16-18 hours,(x4), on an OC'ed QX9650@3.7GHz, thet take 8-9 hours.
*Raistmer, Segur and Whale; optimized app.'s
Knights Who Say Ni, Ni ni
(on Intel proc's, with =>8MB L2 cache & fast >=1333MHz ,>=1600MHz DDR2, or faster DDR3)

---

Personally, I think longer WU files are a much needed item on the menu. I think they should at least be a week long like it used to take my AMD K6 166MHz machine to complete a standard SETI task back in 2002.

Imagine having one of those get corrupt on you - and all you got was a "1" added to your SETI total if it succeeded!

---

Personally, I think longer WU files are a much needed item on the menu. I think they should at least be a week long like it used to take my AMD K6 166MHz machine to complete a standard SETI task back in 2002.

Imagine having one of those get corrupt on you - and all you got was a "1" added to your SETI total if it succeeded!

. . . i believe - that there were a few individuals that crunched [present context] for approx 600 hrs and then got blitzed - i was reading about it in the Forums just recently [and i believe they were po'd]

> runs out of this Thread and proceeds to start dinner for 'is lady ;) 'ello Ozz

---

BOINC Wiki . . .

Science Status Page . . .

Dear Sir's.
I find it difficult to see, how it help Astropulse, files take over 160 hours to complete. To secure a better flow, dereace size to 10 or 15 percent, witch allow older PC'es to operate those files.
I waste more than 160 hours of work, due to one single error in a file. The damage would have been less, if the size and time used, was smaller.
I wrote an e-mail to Matt, to hope for a better future for the Astropulse so more users may begin to search for silversticks, and not only the pure ones of gold. :-)
Yours truly
Leif.

The point isn't better flow, it's to find evidence of intelligence in the universe.

Astropulse is a broadband search. You can't search for a broadband signal in narrowband data.

---

Personally, I think longer WU files are a much needed item on the menu. I think they should at least be a week long like it used to take my AMD K6 166MHz machine to complete a standard SETI task back in 2002.

Imagine having one of those get corrupt on you - and all you got was a "1" added to your SETI total if it succeeded!

. . . i believe - that there were a few individuals that crunched [present context] for approx 600 hrs and then got blitzed - i was reading about it in the Forums just recently [and i believe they were po'd]

> runs out of this Thread and proceeds to start dinner for 'is lady ;) 'ello Ozz

I understand what the complaint is, but essentially its the same complaint when you lost a workunit that crunched for a week and didn't get a simple "1" added to your account back in SETI Classic.

But from my own perspective/opinion, and again this goes into my belief that credits are meaningless, is that its not that big of a deal. I don't even check on my credits like others here do, so it may or may not have happened to me - but I don't care.

..and I think the longer workunits would help ease the server load, and I have often wondered why they don't make workunits longer to begin with.

I'm just exercising my right to disagree and show that there is a difference of opinion out there.

---

There will be increasing calls like this as PC crunching continues its inexorable upward rise. Many will not bother with the "new" PCs on an increasing basis as the power of existing PC will do what they want to do.

Increasingly therefore, over time, a kind of "disenfranchisement" will occur on a 2-4 year view, where an increasing number of lower spec machines will not take part in the inevitable increased WU size that is bound to happen. Already in Beta now is a test on a raised spec for AP WUs due to the hassles felt by lower spec machines. Yet in essense those machines in the real world are still handy beasties. They, as a genre, will also increase in power, just not fast enough to keep pace with the inevitable increase in WU size.

I can see a time in circa 2-4 years (ie "son of Astropulse"), when there will need to be a trickle system and interim credits along the lines of CPDN, if we are to keep on board what will be an increasing number of "lower spec" machines as time goes by. "Lower Spec" is only a relative term, that catagory will be increasingly powerful, and still useful to the Project in large numbers.

One size, as they say, dont fit all. The MBs are too small for long term, in circa two years PCs will spit those out in around a minute on CUDA (or "son of")- and each will have 10-20% deadtime per WU loading etc - they will have to increase in size. Then there will be the question whether maintaining two different sizes makes sense. In all this, the Science Rules of course, without questions, no debate, and the Science will be the start point of all deliberations.

However, the dynamics of exponential hardware capacity and power increases, versus fixed size of current WUs has much potential for biting the Project in the longer term - whatever the Science.

Regards
Zy

---