BTW: with "switching" Apps calling the switcher doesn't work right away, you'll need to call the actual App.
BM
This is the result with *linux-gnu_1
real 11m42.814s
user 11m34.615s
sys 0m3,040s
Tullio
Looks like a rather fast machine. Can you post the details?
Actually the time only makes sense to compare Apps e.g. on the same machine. Could you run this with the generic App (*linux-gnu_0), and possibly with the SSE2 App?
BM
The CPU is an AMD Opteron 1210 at 1.8 GHz running SuSE Linux 10.3. It was running other 2 projects on the second core with low priorities.
Here are the results with *linux-gnu_0
real 18m9921s
user 17m58,671s
sys 0m3.423s
There is a huge difference. Where can I find the SSE2 app?
Found it. Here are the results:
real 11m0,553s
user 10m52.413s
sys 0m3.072s
Question: can I rename this app to *linux-gnu_1? It is slightly faster (about 6 %).
Hey guys, I tested the "0" and "1" apps on my box (Core Duo 1600 MHz) and there seems to be quite a difference (the latter is about a third faster) if I read this correctly. Where can I find the "SSE 2" app to test it against those two?
Hey guys, I tested the "0" and "1" apps on my box (Core Duo 1600 MHz) and there seems to be quite a difference (the latter is about a third faster) if I read this correctly. Where can I find the "SSE 2" app to test it against those two?
Yonah ;-) although Dothans were nice CPUs aswell, especially those with loads of cache (my brother has one and I think it's a shame he has stopped doing BOINC on it due to heat issues). Thanks for the link, I'll run the reference WU with this app first and see if it significantly increases performance on my laptop, then I'll switch (or not).
Note that you better don't unzip the archive and run the script in the original BOINC/projects/einstein.phys.uwm.edu directory. In case you do have a workunit of the same frequency range you might mess up your skygrid file and thus your results (unzip should ask for confirmation before overwriting, but BOINC doesn't). At least better delete the skygrid file after running the test.
Mikie, you know what's interesting? I'm running Ubuntu aswell... makes me wonder if it's less the processor that doesn't profit from SSE 2 than a prob with the OS... Bernd, Bikeman, do you think it's worth playing around with a live CD to check? I should have Debian, Gentoo and BackTrack lying around... plus FreeBSD but I'm not sure if that agrees with my laptop (ATI chipset).
RE: RE: RE: Try
)
The CPU is an AMD Opteron 1210 at 1.8 GHz running SuSE Linux 10.3. It was running other 2 projects on the second core with low priorities.
Here are the results with *linux-gnu_0
real 18m9921s
user 17m58,671s
sys 0m3.423s
There is a huge difference. Where can I find the SSE2 app?
Found it. Here are the results:
real 11m0,553s
user 10m52.413s
sys 0m3.072s
Question: can I rename this app to *linux-gnu_1? It is slightly faster (about 6 %).
Hey guys, I tested the "0"
)
Hey guys, I tested the "0" and "1" apps on my box (Core Duo 1600 MHz) and there seems to be quite a difference (the latter is about a third faster) if I read this correctly. Where can I find the "SSE 2" app to test it against those two?
einstein_S5R3_4.49_i686-pc-linux-gnu_0:
1405.23 user
3.29 system
24:07.07 elapsed
einstein_S5R3_4.49_i686-pc-linux-gnu_1:
936.51 user
3.72 system
16:02.29 elapsed
RE: Hey guys, I tested the
)
Hi Annika!
It's here http://einstein.phys.uwm.edu/power_apps/einstein_S5R3_4.49_1_i686-pc-linux-gnu.gz
Just gunzip it and "chmod u+x" it , and that's it. If you want to replace the SSE app, note you'll have to rename it to
einstein_S5R3_4.49_i686-pc-linux-gnu_1
as well.
Should run on your Dothan as well.
CU
Bikeman
Yonah ;-) although Dothans
)
Yonah ;-) although Dothans were nice CPUs aswell, especially those with loads of cache (my brother has one and I think it's a shame he has stopped doing BOINC on it due to heat issues). Thanks for the link, I'll run the reference WU with this app first and see if it significantly increases performance on my laptop, then I'll switch (or not).
Well, that app doesn't seem
)
Well, that app doesn't seem to offer a better performance on my box:
935.27 user
6.88system
17:20.77elapsed
RE: Well, that app doesn't
)
It is faster on my Opteron 1210. 26k s instead of 30k s for a WU only partially crunched with it.
Tullio
Note that you better don't
)
Note that you better don't unzip the archive and run the script in the original BOINC/projects/einstein.phys.uwm.edu directory. In case you do have a workunit of the same frequency range you might mess up your skygrid file and thus your results (unzip should ask for confirmation before overwriting, but BOINC doesn't). At least better delete the skygrid file after running the test.
BM
BM
OK Bernd. I crunched my last
)
OK Bernd. I crunched my last WU 5/6 with the SSE2 optimized app in 23,932.788 s.Good work!
Tullio
Gives me 35.61 credits/h.
On a Kubuntu 8.04 64 bit
)
On a Kubuntu 8.04 64 bit setup with an AMD Turion X2 TL-56 I get the following results with the test workunit:
./einstein_S5R3_4.49_i686-pc-linux-gnu_0
real 19m27.247s
user 19m21.557s
sys 0m2.940s
./einstein_S5R3_4.49_i686-pc-linux-gnu_1(SSE)
real 12m27.322s
user 12m22.746s
sys 0m2.312s
./einstein_S5R3_4.49_1_i686-pc-linux-gnu(SSE2)
real 12m33.268s
user 12m28.435s
sys 0m2.484s
Looks like the SSE2 app has no performance gain on Turion processors.
Mikie, you know what's
)
Mikie, you know what's interesting? I'm running Ubuntu aswell... makes me wonder if it's less the processor that doesn't profit from SSE 2 than a prob with the OS... Bernd, Bikeman, do you think it's worth playing around with a live CD to check? I should have Debian, Gentoo and BackTrack lying around... plus FreeBSD but I'm not sure if that agrees with my laptop (ATI chipset).