Message boards :
Number crunching :
Curious about Mac/intel and the Mac/PowerPC apps
Message board moderation
Author | Message |
---|---|
toffuuu Send message Joined: 30 Mar 00 Posts: 87 Credit: 1,887,469 RAC: 0 |
Can anyone here tell me if the regular SETI apps here for mac that SETI@Home @ berkeley makes, is Optimized in any way, to use SSE or SSE2 or SSE3? cause i dont see any indication of it in the stderr.txt file this is what i see.. thanks *edit* also i forgot to Mention if at all if AltiVec is used for the G4's and G5 as well in these "normal apps", but i have an intel mac so thats why im just curious.. setiathome_enhanced 5.28 XCode GCC 4.0.1 (Apple Computer, Inc. build 5367) i386 libboinc: 5.10.20 libboinc: 5.10.20 Work Unit Info: ............... WU true angle range is : 2.319404 Optimal function choices: ----------------------------------------------------- name ----------------------------------------------------- v_BaseLineSmooth (no other) v_vGetPowerSpectrumUnrolled 0.00024 0.00000 sse2_ChirpData_ak 0.02618 0.00000 v_vpfTranspose8x4ntw 0.01269 0.00000 BH SSE folding 0.00149 0.00000 Restarted at 80.92 percent. |
Gecko Send message Joined: 17 Nov 99 Posts: 454 Credit: 6,946,910 RAC: 47 |
Can anyone here tell me if the regular SETI apps here for mac that SETI@Home @ berkeley makes, is Optimized in any way, to use SSE or SSE2 or SSE3? cause i dont see any indication of it in the stderr.txt file this is what i see.. thanks The short answer to your question is "yes". Berkeley's current releases contain many "improvements" for specific optimizations that use SIMD instructions. |
toffuuu Send message Joined: 30 Mar 00 Posts: 87 Credit: 1,887,469 RAC: 0 |
I just realized i saw the SSE and SSE2 but what about SSE3??? I used to use Alex kahns Workers loved those, until I realized, Berkeley has a new version up so I just wanted to try, and may stick with it... |
Gecko Send message Joined: 17 Nov 99 Posts: 454 Credit: 6,946,910 RAC: 47 |
Alex still makes them & has released new optimized applications which are the quickest yet. See this thread: http://setiathome.berkeley.edu/forum_thread.php?id=31810&nowrap=true#656269 You will find both Core 2 and G4 aps. Have fun! |
Francois Piednoel Send message Joined: 14 Jun 00 Posts: 898 Credit: 5,969,361 RAC: 0 |
my favorit video about this: Bunny toasted and yes, it's true who? Skulltrail D5400XS |
toffuuu Send message Joined: 30 Mar 00 Posts: 87 Credit: 1,887,469 RAC: 0 |
Yes I know, but those are version 5.27, not 5.28 like here, and I use Intel Core Duo, not the core 2 one tho on my MBP, but I think Im gonna stick with the 5.28 for now, since i really dont see any difference between the 2, except the way Alex Kahns reports it back to seti@home so.. i dont know.. we'll see how my RAC goes for a few weeks I guess.. |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
There is an SSE3 version of chirping, and it was chosen on your result 646415864. On your host the difference in speed of the various chirping routines seems to be small enough that the brief and low priority testing doesn't give a consistent choice. Joe |
Adrian Taylor Send message Joined: 22 Apr 01 Posts: 95 Credit: 10,933,449 RAC: 0 |
alex does a brand new version for core duo. its definitely much faster than the stock Berkeley app and does exactly the same science, theres really no reason to run the stock app unless you want less science produced per hour :-) you can see from my 8-core xeon the boost the optimised app supplies, number one machine in the world for RAC he heh regards adream 63. (1) (b) "music" includes sounds wholly or predominantly characterised by the emission of a succession of repetitive beats |
Gecko Send message Joined: 17 Nov 99 Posts: 454 Credit: 6,946,910 RAC: 47 |
Sorry for confusion re: Core 2. Didn't realize you had Core Duo, but you no doubt saw the link w/ core Duo ap. As for version #, 5.28 is the current OSX MB version w/ revised credit multi of 2.85 Alex's V8 is also MB w/ the credit multi revised to 2.85. It does the the same science as stock Berkeley 5.28, correctly &......MUCH faster. Your choice. You can also edit the app_info.xml file w/ V8 to look like the following. This will then report as 5.28
|
toffuuu Send message Joined: 30 Mar 00 Posts: 87 Credit: 1,887,469 RAC: 0 |
I already know how to do all this if you look at my past task you'll see that i did for awhile... :-D Like this -> http://setiathome.berkeley.edu/result.php?resultid=644827313 and this one off of my G4 450mhz -> http://setiathome.berkeley.edu/result.php?resultid=645053720 |
Gecko Send message Joined: 17 Nov 99 Posts: 454 Credit: 6,946,910 RAC: 47 |
I already know how to do all this if you look at my past task you'll see that i did for awhile... :-D Ok. Trying to help, but I guess I don't understand the root of your remaining concern. There is certainly nothing wrong w/ running stock 5.28 from Berkeley. V8 is substantially quicker however and is quite a different animal under the hood. Run em' both, compare & decide from there. Cheers! |
toffuuu Send message Joined: 30 Mar 00 Posts: 87 Credit: 1,887,469 RAC: 0 |
I already know how to do all this if you look at my past task you'll see that i did for awhile... :-D Well I was just wondering if the 5.28 was far improved agianst the 5.27 V8 Optimised one that Kahn made, and as you said it isnt and I agree.. what I am now am gonna do is is not reqest anymore work and let 5.28 finish all the rest of the Units i have on here, then go back to v8.... which my G4 is still using.. lol |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.