Bruce, a question about An Optimized Application

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9,352,143
RAC: 0

RE: RE: Thanks for your

Message 23661 in response to message 23637

Quote:
Quote:
Thanks for your continued efforts. Seems to be running fine on a K6-2, and hopefully I'll be back in around 100 Ksecs. or so with a positive report! :-)

Probably this code isn't the best for K6-2, but I hope that it will be faster than original.

Roger that, but even though it may not be optimum for the K6 family, it's looking at this point like it's still going to be about twice as fast as stock Albert! Sweet! :-)

Alinator

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 247
Credit: 219,266,664
RAC: 0

RE: ...it's looking at this

Message 23662 in response to message 23661

Quote:
...it's looking at this point like it's still going to be about twice as fast as stock Albert! Sweet! :-)Alinator


This "unofficial" Beta test of ours seems to be going quite well! :)

One hopes the official powers that be will release this as an official release so that the entire project can benefit from this wonderful optimization!

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157,718
RAC: 0

RE: RE: ...it's looking

Message 23663 in response to message 23662

Quote:
Quote:
...it's looking at this point like it's still going to be about twice as fast as stock Albert! Sweet! :-)Alinator

This "unofficial" Beta test of ours seems to be going quite well! :)

One hopes the official powers that be will release this as an official release so that the entire project can benefit from this wonderful optimization!

The feeling I get is that it will happen rather quickly, though there may be some delay while they try to massage the code for our folks on Linux and older, non-Altivec Macs. Shouldn't forget the Solaris, either.
I expect there will be a scramble to upgrade the servers, too, to handle the vastly increased throughput.

microcraft
"The arc of history is long, but it bends toward justice" - MLK

RandyC
RandyC
Joined: 18 Jan 05
Posts: 3,007
Credit: 111,139,797
RAC: 1

RE: RE: RE: ...it's

Message 23664 in response to message 23663

Quote:
Quote:
Quote:
...it's looking at this point like it's still going to be about twice as fast as stock Albert! Sweet! :-)Alinator

This "unofficial" Beta test of ours seems to be going quite well! :)

One hopes the official powers that be will release this as an official release so that the entire project can benefit from this wonderful optimization!

The feeling I get is that it will happen rather quickly, though there may be some delay while they try to massage the code for our folks on Linux and older, non-Altivec Macs. Shouldn't forget the Solaris, either.
I expect there will be a scramble to upgrade the servers, too, to handle the vastly increased throughput.

Hope they increase the daily quota too. I've got one system that runs out of WUs (the small ones) in less than 12 hrs. (That's all it's been getting since I switched to the optimized app.)

Seti Classic Final Total: 11446 WU.

Nightbird
Nightbird
Joined: 17 Feb 05
Posts: 79
Credit: 561,723
RAC: 0

A36 runs perfectly on my

A36 runs perfectly on my Barton 3000+.
No crash, wus validated OK
It's time now to install A36 on a Barton 2500+ and an Athlon64 3200+

Definitively, a superb application
Thanks to you and kudos Akosf.

[

Terry
Terry
Joined: 9 Feb 05
Posts: 55
Credit: 1,781,475
RAC: 0

RE: Hope they increase the

Message 23666 in response to message 23664

Quote:

Hope they increase the daily quota too. I've got one system that runs out of WUs (the small ones) in less than 12 hrs. (That's all it's been getting since I switched to the optimized app.)

I just got a batch that's running around 14 minutes each. At that rate, I'll be out in less than 8 hours. I don't run anything but einstein now, so if they don't increase the quota, my cpus will get some well deserved rest.

Terry

Melf-san
Melf-san
Joined: 30 Nov 05
Posts: 10
Credit: 3,382,540
RAC: 1,330

Currently testing the new app

Currently testing the new app and its really nice :-)
Going like 1 minute = 1 % on my Athlon 64 3000+ :D

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 81

I examined the linux code

I examined the linux code closely. It's horrible.
The 'blind flying' type optimization is hopeless. I'm sorry...

Probably a better compiler would be able to increase its performance significantly.

RandyC
RandyC
Joined: 18 Jan 05
Posts: 3,007
Credit: 111,139,797
RAC: 1

RE: I examined the linux

Message 23669 in response to message 23668

Quote:

I examined the linux code closely. It's horrible.
The 'blind flying' type optimization is hopeless. I'm sorry...

Probably a better compiler would be able to increase its performance significantly.

Thanks for looking anyway. Nobody can do everything (and do a good job of it, that is).

Seti Classic Final Total: 11446 WU.

DanNeely
DanNeely
Joined: 4 Sep 05
Posts: 1,323
Credit: 1,797,270,291
RAC: 819,464

RE: The feeling I get is

Message 23670 in response to message 23663

Quote:
The feeling I get is that it will happen rather quickly, though there may be some delay while they try to massage the code for our folks on Linux and older, non-Altivec Macs. Shouldn't forget the Solaris, either.
I expect there will be a scramble to upgrade the servers, too, to handle the vastly increased throughput.

IT depends. Judging from comments Bruce's made earlier the gains from hand optimization vs letting the compiler do all the work are very minimal outside of altivec macs. Which means that first he'll have to convert aksof's enhancements (improvment's in the fundamental algorithm) back into C(++?) and build executables for all the various platforms before releasing anything.

Comment viewing options

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