Optomized S5 SSE3

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 560
Credit: 4,514,103
RAC: 218

RE: I have a Pentium D that

Message 38866 in response to message 38865

Quote:
I have a Pentium D that has SSE3 and several AMD machines. Let me know what I can do to help.

Ok. Here are the steps:

- download patch file
- stop boinc
- copy patch file ( and patcher ) to the project directory
- do patching ( run patcher with the data file as a parameter )
- start boinc
- do some wus
- give me a report about the results ( valid, invalid, zero credit, etc... )

Here is the first patch file: S5T0000.DAT

This patch doesn't have any improvement, but we can check the procedure.

Pepperammi
Pepperammi
Joined: 20 Feb 05
Posts: 131
Credit: 437,943
RAC: 0

Again you can use my computer

Message 38867 in response to message 38864

Again you can use my computer towards your efforts.
[edit]
deleted email-not needed, :)

Pepperammi
Pepperammi
Joined: 20 Feb 05
Posts: 131
Credit: 437,943
RAC: 0

RE: ... This patch doesn't

Message 38868 in response to message 38866

Quote:
...
This patch doesn't have any improvement, but we can check the procedure.


Are you sure?! Just gone throught it all and it seems it going 00.002%per/second instead of the 00.001%per/sec it was doing before. Estimated to completion is dropping at 2sec per sec too? nice for something that doesn't have any improvements :-)

Real test is for errors though as you say. Any way to check the patcher has done it's job?

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

Running now. Have a bunch of

Running now. Have a bunch of the short wu's, so I should be able to give you a report in a few hours.

Crunch3r
Crunch3r
Joined: 22 Jan 05
Posts: 90
Credit: 30,236,616
RAC: 0

Hi Akos, seems to be

Message 38870 in response to message 38869

Hi Akos,

seems to be working ;)
Don't know what you did but i noticed the same as pepperammi.

result

P.S.

I patched the einstein app. and restarted boinc at 98% done.
Result is valid though.

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 560
Credit: 4,514,103
RAC: 218

RE: RE: This patch

Message 38871 in response to message 38868

Quote:
Quote:
This patch doesn't have any improvement, but we can check the procedure.
Are you sure?! Just gone throught it all and it seems it going 00.002%per/second instead of the 00.001%per/sec it was doing before. Estimated to completion is dropping at 2sec per sec too? nice for something that doesn't have any improvements :-)

:-))

Quote:
Real test is for errors though as you say. Any way to check the patcher has done it's job?

S5T0000 has a 4956 byte length rebuilded SSE hot loop. We are testing the patcher, the recompiled hot loop and the whole test too.

edit: We will optimize this hot loop. The CPU spends the 99% of the time here.

Mats Nilsson
Mats Nilsson
Joined: 10 Dec 05
Posts: 94
Credit: 15,011,147
RAC: 0

Something is happening

Something is happening thanks, could it be true that is speeding up einstein, in that case great, is it possible too test this patch on my A64 that only have SSE2 capable.

Crunch3r
Crunch3r
Joined: 22 Jan 05
Posts: 90
Credit: 30,236,616
RAC: 0

RE: Something is happening

Message 38873 in response to message 38872

Quote:
Something is happening thanks, could it be true that is speeding up einstein, in that case great, is it possible too test this patch on my A64 that only have SSE2 capable.

It seems to run on all cpus capable of SSE.
Tested it on a dual P3 and it seems to be working ;)

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 560
Credit: 4,514,103
RAC: 218

S5T0001 - eliminated

S5T0001

- eliminated double jumps

Pepperammi
Pepperammi
Joined: 20 Feb 05
Posts: 131
Credit: 437,943
RAC: 0

RE: S5T0001 - eliminated

Message 38875 in response to message 38874

Quote:

S5T0001

- eliminated double jumps


so should we stop again and update now or wait?

Comment viewing options

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