• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

New MultiBeam (MB) SETI Client Released

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Me too, waiting for the cache to empty on my P4 so I can load the new client. Ill leave my c2d system on regular client for now, until some optimizations come out.
 
The MB app is backwards compatible, it can run older WUs, so there's no reason to stop getting work
 
Yeah I know that, tried deleting the app_info file and for some reason it still continued to crunch with the old client and didn't download the new one, :shrug:.
And since I only had 2 WUs in cache I figured out that I should wait for that 2 ones to finish.

Anyway, tried messing with some files and ended up trashing those 2 WUs :bang head :bang head, well thats what redundancy is for @ Berkeley :shrug:

Im running multi beam client now, will post computation times after it finishes some WUs.
 
Chicken said:
we're pretty much ready to release, it just depends on what multiplier finally becomes the new stock one.

On Beta, it used to be 3.81, on the main project, 3.35 (because MB WUs were thought to give less cr/h than enhanced ones).

Turns out, with some added chicken flavour, the new stock app is so much faster than before that the stock multiplier was decreased instead of increased - namely, to 2.8.

To contrast -

5.15 (stock app til now) - 100% cr/h
5.27 (new stock app?) - 83.6% cr/h
Beta MB apps - 113.7% cr/h

The forums are already erupting with indignant power crunchers.

Let's face it, they're the only ones who lose - people who use the stock app will not see their cr/h change much, they'll just process more WUs. Chances are, they'll never even notice it.

The real difference is that from now on the optimized apps will bring less gains over the stock ones. This is something our egos will just have to get to terms with ;o)

My RAC may plummet, but the science done still remains valid.

And THAT'S what counts.

Regards,
Simon.
 
So units that have been crunched with the chicken app will receive a fixed % fewer cobblestones than the same work unit with the stock app to make up for the faster crunching that the Chicken app can do?
 
Not quite.

Those of us who historically used chicken have experienced higher than normal production.

The new client is already somewhat optimized thus we will not see the same huge gains that we are used to.

Thus it will appear that our RAC is dropping.

It's like daylight savings time, the more you talk about it the less anybody understands it.
 
Yea - AMD is still the "generic" one even though the last I saw over 30% of all SETI machines are AMD - a much bigger chunk than Linux. The 939s and up all have SSE3 but apparently it's not quite the same as Intel's. Maybe there's no benefit from it or ... :shrug:


But I'm still happy the KWSN supply apps for us - you gotta' give them kudos even if you disagree with the approach ...
 
QuietIce said:
Yea - AMD is still the "generic" one even though the last I saw over 30% of all SETI machines are AMD - a much bigger chunk than Linux. The 939s and up all have SSE3 but apparently it's not quite the same as Intel's. Maybe there's no benefit from it or ... :shrug:


But I'm still happy the KWSN supply apps for us - you gotta' give them kudos even if you disagree with the approach ...

The barrier is a licensing problem with the tools used to produce the optimized client. It is not licensed to produce distributable products that run AMD SSE3. (That what I get out of the discussions, but I may be slightly misunderstanding the finer details)
 
Back