Has anyone has gotten BOINC 5.2.13 or 5.2.14 to compile with optimizations for linux? I'm using gcc 4.0.2_20050901-3 to compile from the stable cvs source and when running ./configure I get:

configure: creating ./config.status
config.status: creating
.infig.status: error: cannot find input file:

I did everything the same with 5.2.14 as I did with 5.2.8

./autosetup

export CFLAGS="-march=pentium3 -O3 -fomit-frame-pointer -funroll-loops -fforce-addr -ffast-math -ftracer"

export CXXFLAGS=$CFLAGS

./configure --disable-server

make clean

make

The link is to an image of boincmgr running in KDE on SuSE 10.0. It was compiled here with the same version of gcc on the same computer. It uses wxGTK-compat.
My BOINC 5.2.8

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

Has anyone has gotten BOINC 5.2.13 or 5.2.14 to compile with optimizations for linux?

Altho I havn't gotten the compiles to work myself either (still trying to learn tho), I have been using the boinc and seti executables from http://naparst.name/ and they are fast.

He has a seti client compiled with SSE but I don't have a p3 machine around anymore to try it on.

Hope this helps at least a little.

---

I see from the Boinc alpha/dev mail list that TMR and Lars have had similar problems.

tony

Altho I havn't gotten the compiles to work myself either (still trying to learn tho), I have been using the boinc and seti executables from http://naparst.name/ and they are fast.

He has a seti client compiled with SSE but I don't have a p3 machine around anymore to try it on.

Hope this helps at least a little.

I tried that build but after the first launch it was unable to connect to the host. Blanking the password didn't help.

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

I see from the Boinc alpha/dev mail list that TMR and Lars have had similar problems.

tony

How would I go about subscribing? I have a feeling many of the problems I've run into will be dealt with there.

[edit] I just found the subscription page. [/edit]

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

I see from the Boinc alpha/dev mail list that TMR and Lars have had similar problems.

tony

How would I go about subscribing? I have a feeling many of the problems I've run into will be dealt with there.

[edit] I just found the subscription page. [/edit]

you can subscribe to the Alpha mail list here, and the boinc Development mail list here

tony

I see from the Boinc alpha/dev mail list that TMR and Lars have had similar problems.

tony

How would I go about subscribing? I have a feeling many of the problems I've run into will be dealt with there.

[edit] I just found the subscription page. [/edit]

you can subscribe to the Alpha mail list here, and the boinc Development mail list here

tony

Thanks tony. I'm subscribed to both now. I did the dev subscription and was going through the december digest. It seems there's a fix for the cvs in the works. That's a relief. I like to run the most recent version of BOINC and the boincmgr I compile looks so much better than the one berkeley provides.

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

I would like to thank tony for pointing me to the dev list, to Tetsuji for posting the report of the problem, and Egon for the solution. "dos2unix configure.ac" took care of the problem and BOINC 5.2.14 compiled fine and is now up and running.

Now to find a newer set of flags to use. I'm still getting linux benchmark results significantly lower than the win2k results.

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

Hello,

First of all i want to say that i didn't want to hijack this thread.
But as it deals with an optimized 5.2.14 for linux i thought that i don't have to start a new thread about the same topic.

Thanks

Here we go:

As it's been a long time since i've updated my Linux Boinc clients i thought i should do so now, because it has been requested.

I've merged P3 and AthlonXP(SSE) into one Boinc version and the same happend to P4 and AMD64(SSE2).

Changes:
- the clients are statically linked no dependencies needed.
- enabled "return_results_immediately"

You can download them at my page.

P.S.
If you don't see them try to refresh your browser

---

Join BOINC United now!

What flags did you use to compile? What compiler? Which source package?

[edit] Too slow. You beat me. How did you enable "return_results_immediately"?[/edit]

---

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)

What flags did you use to compile? What compiler? Which source package?

Used: ICC + latest CVS Source.

---

Join BOINC United now!

You can download them at my page.

Uhm, and the point of optimizing a client that failed alpha-testing is...?

Did you know David Anderson declared 5.2.14 as "flunked" in Alpha?

David Anderson to boinc_alpha
More options Dec 9 (1 day ago)

Testers:
5.2.14 didn't pass - no need for further testing.
We'll have a new version to test in about 1 week.

We got no test results for WinNT, Mac OS X or Linux,
modem-connected computers, or computers with SOCKS proxies.
If you have one of these platforms and will be able
to test with it on a consistent basis, let me know
(reply directly to me).
Otherwise we'll get some more testers.

Thanks

-- David

I thought you should know before to much effort was applied

tony

---

You can download them at my page.

Uhm, and the point of optimizing a client that failed alpha-testing is...?

Just for fun. Nobody is forced to use this one and i've been asked to do so.
And most important, after i posted this optimized version, berkeley can just release 5.2.15....

Result:
Everyone is happy


EDIT

If this one is "flunked" Alpha then why is it in the stable CVS ???

EDIT

---

Join BOINC United now!

You can download them at my page.

Uhm, and the point of optimizing a client that failed alpha-testing is...?

Just for fun. Nobody is forced to use this one and i've been ased to do so.
And most important, after i posted this optimized version, berkeley can just release 5.2.15....

Result:
Everyone is happy

@Crunch3r

just saw that you are getting in time into the 6 Digit Club, over one Million

Congratz to this

Greetings from Germany NRW
Ulli

You can download them at my page.

Uhm, and the point of optimizing a client that failed alpha-testing is...?

Just for fun. Nobody is forced to use this one and i've been ased to do so.
And most important, after i posted this optimized version, berkeley can just release 5.2.15....

Result:
Everyone is happy

@Crunch3r

just saw that you are getting in time into the 6 Digit Club, over one Million

Congratz to this

Greetings from Germany NRW
Ulli

At this point i should be way over 6 digits... just a simply math here... 7 days * 4500 = 31000 credits pending...

Your Congrats have to go to the optimized client wich increased my RAC from 3700 to 4500 ;)

---

Join BOINC United now!