does anybody have the source code for SETI?

[Ameesh]

hey i would really like one so i can compile a version for my p4 and p3's im sure we are missing out on a grip of optimizations that can be done.

 

[BurntKooshie]

We'd all love to optimize it....Awhile back, I believe Intel, and AMD had some people work on optimizing the clients even more (but they don't want to reveal the source code, otherwise people might end up hacking it even more easily than they do now). The result was massively delayed, but it was the 3.0 client - blazing fast....too fast, infact. So they used the opportunity to add more searching features into the client, and thus the client now takes longer (again).

 

[Ameesh]

dammit! we cant run the 3.0 client any more?

 

[Ameesh]

oh the other thing is that the winnt file is called 386cmdlin etc so does that mean its only using the 386 instruction set?

 

[MrAtheist]

<< oh the other thing is that the winnt file is called 386cmdlin etc so does that mean its only using the 386 instruction set? >>



No, it means it uses instructions from the 386 and upward instruction-sets. I think

 

[Ameesh]

doesnt that seem backwards? all x86 instruction sets are backwards comatible not forwards compatible.

 

[JHutch]

Ameesh,

It is a fairly trivial thing to code a client to run different code paths based on what instruction set is detected. Therefore, it can run on a 386 and higher, while having the ability to use things like MMX, SSE, SSE2, 3DNOW!, etc, if it finds the capability is there. (NOTE: I don't know how much, if any, of those instructions are actually used in the SETI client.)

JHutch

 

[ElFenix]

386 simply means it runs the IA-32 instruction set, introduced with the 386.

 

[Assimilator1]

BK
I believe Intel, and AMD had some people work on optimizing the clients even more .........The result was massively delayed, but it was the 3.0 client

I'm not saying your wrong ,but are you sure about that? ,I've not heard that before


Ameesh
All clients prior to v3.03 were blocked on the 1st Feb.
(BTW on some older PII's CLi v2.4 was faster than v3.0 ,the cut off point was about 350 MHz)

 

[MGallik]

In another year or so, we'll all have 20Ghz cpu's
and be knocking down WU's in 30 minutes anyway.

 

[Assimilator1]

And long before that they'll run out of bandwidth &amp; introduce v4 with x2 WU times!

Hopefully they will bring on that Southern hemisphere telescope soon &amp; add a seperate server

 

[BadThad]

In the interest of science, they would never reveal the source code. Plus, there are some freaking idiots out there that would hack it to produce massive wu's to bolster their stats, even though the results are invalid...it has already happened. :|