A month ago I made optimized clients for Windows using an evaluation copy of Intel C++ compiler and fftw library, but as I mentioned in another thread, I received money from a person to buy the license...so anyone can use the binary the binaries legally. Thank you, Mr.X!! (He wants me not to mention his name.)

Links are as follows: I added seti-sse2.zip, which I hope works for Athlon64.

seti-plainvanilla.zip For P2 or earlier....but this uses MMX instructions.

seti-p3.zip for P3 (and Athlon XP?)

seti-sse2.zip for Athlon64(?) This requires special caution. No test has been done yet.

seti-p4-no-sse3.zip for P4 w/o SSE3. According to Intel, this is slightly different from the above one. So this doesn't work for Athlon64.

seti-p4-sse3.zip for P4 with SSE3

seti-pm.zip for Pentium M.

referenceunits.zip a work unit and a result unit for benchmark test.

Most of them are the ones I uploaded before, except for seti-sse2.zip. All of them are based on Apr-12 nightly build of boinc/seti and fftw-3.0. If you want to compile by yourself (you need Intel compiler and VC++7.1), sources are
seti-source-1.zip, seti-source-2.zip, and seti-source-3.zip.

CAUTION
These binaries are provided as is. No guarantees about proper operation can be given. Only you will be to blame if binaries cash your computer, damage data or hardware, kick your dog or insult your mother. (from Metod's page and I concur.)

I strongly recommend to make a backup copy of the whole boinc folder before installing these clients in case it crashes....it may lose your precious data due to incompatibility (or by my mistake?)

Happy crunching!! You can copy or distribute them for free :)

---

Luckiest in the world. WMD = Weapon of Mass Distraction.
Click this table.

You are the best! Respect!
Only one a problem - I can't download :(

---

You are the best! Respect!
Only one a problem - I can't download :(

Works fine for me, thanks Tetsuji

---

Raithmir's SPARC64/UltraSPARC Linux Builds
http://www.kulthea.net/boinc/

You are the best! Respect!
Only one a problem - I can't download :(

I know that problem :) I edited and fixed the links :) SORRY!!!!! and thanks!!!! But it isn't me that boost the client, but Intel compiler and fftw library :)

Good luck all!!!

---

Luckiest in the world. WMD = Weapon of Mass Distraction.
Click this table.

Thats really kool.
Thank you, Tetsuji Maverick Rai.
Thank you, Mr. X.
For all your works and efforts.

---

http://www.cyclistlog.com/

Yes, a really _big_ thankyou Tetsuji.
Brilliant.

Genki desu. Arigato gozaimasu.

(I hope those are appripriate Japanese Thank yous' - there seem to be so many ways for different situtions!)

---

Many thanks Maverick and also to the anonymous Mr "X" you are both champions!

---

I know that problem :) I edited and fixed the links :)

All OK!
Tnx!

---

I know that problem :) I edited and fixed the links :) SORRY!!!!! and thanks!!!! But it isn't me that boost the client, but Intel compiler and fftw library :)

Good luck all!!![/quote]

But without you both, there would be no optimised client. So please accept the admiration and gratitude of us all. Thank you Tetsuji and Mr. X.

Kindest regards and respect

Tony

---

I may be getting older but I refuse to grow up!

Thank you very much, Tetsuji and Mr. X! :)

Dig

Thanks everyone who made this possible. I downloaded two of the zips and I found a small problem. You still have the old app_info.xml file that does not tell boinc what to do in case it runs into a 4.09 seti work unit which is to use the 4.11 executable. Not going to be a problem for those of us that have been running optimized for the last month. However, if someone new downloads the files it may cause problems that we have seen before. I forget who posted the "correct" version of the app info file but I have a copy of it.

HokieForever

---

Horaaaaay!
:D

Hugs,

---

Thanks :)

Now also available at http://www.esea.dk/esea/optimized.asp

---

thank you TMR and Mr X!

happy crunching! (and faster)

Micah

---

S@h Berkeley's Staff Friends Club

You are welcome all :) and thanks Mr.X!

These days I've been wondering how to get more credits granted, and found a long "work_buf_min_days" in global_prefs.xml seems to work. The longest valid value sounds 10 days (as far as I read in this message board. I may be wrong.) Optimized clients calculates more with fewer instructions/clocks than the official ones. Benchmark tests in the core client isn't so optimized as seti client. So usually optimized clients claim less credits than the official ones.....so I guess many people are annoying...(and that it affects the other 3 hosts also)

Here is how it works. There are 4 clients crunching one work unit. And the granted credit is determined when 3 of them finish crunching. The 2nd highest (or lowest) claimed credit is granted. And the claimed credit of the 4th client is ignored. So if you are the 4th, you will (almost) always get "usual" or "ordinary" credits....

May the "4th" be with you!!!

sorry, I just wanted to write this ;) It works when there is only one optimized client among the first 3 clients....but to make it certain, you'd better be the 4th.

---

Luckiest in the world. WMD = Weapon of Mass Distraction.
Click this table.

Happy crunching!! You can copy or distribute them for free :)

Thanks a lot!

Of course again available on www.marisan.nl/seti/

Greetings,
Speedy67

---

Hi All,
I downloaded seti-p3-sse2.zip

I followed the directions
and Restarted BOINC all seems to be running with no problems
E@H is running because of the schedualer.
This was in the messages is this correct?
5/31/2005 12:56:09 PM|SETI@home|Found app_info.xml; using anonymous platform

Should I suspend E@H and Run S@H for a run?
What improvements should I see?

Thanks for all your hardwork!

Tim

---

Hi All,
I downloaded seti-p3-sse2.zip

I followed the directions
and Restarted BOINC all seems to be running with no problems
E@H is running because of the schedualer.
This was in the messages is this correct?
5/31/2005 12:56:09 PM|SETI@home|Found app_info.xml; using anonymous platform

It's normal :) See lower part of this page.

Should I suspend E@H and Run S@H for a run?
What improvements should I see?

Thanks for all your hardwork!

Tim

No you don't have to suspend E@H. Just let it go as it likes (core client will run S@H according to the preference values you have set.) Since optimized core client gives more benchmark scores, so your E@H will claim more credits than before. And the optimized Seti@Home will be faster. If you want to let SETI work more, then change the preference values as you like. If you want to test S@H sooner, you may suspend E@H for a while (but usually deadline of E@H is closer than that of S@H) If your E@H returns the result behind its deadline, your result will be rejected as "validation error."

And it's not so hard (compiling procedures have been established on Linux).....the harder part was to write how to build it in other threads :) But it's not so hard.

regards, and happy crunching, everyone..........

---

Luckiest in the world. WMD = Weapon of Mass Distraction.
Click this table.

And it's not so hard (compiling procedures have been established on Linux).....the harder part was to write how to build it in other threads :) But it's not so hard.

regards,[/quote]

Thanks Tetsuji,

I am going to try it on and amd machine next....

---

Thanks again to you, Tetsuji and thanks to Mr. X for the licensing!

---