FWIW, the new stock "power" app seems to be inline with the mac app now...
if you look at the top host wich is a mac it completes a wu in about 14k sec to 16k sec.
Since "Re allowing" einstein to run on hostid=825074, the very first two wus yielded "computation errors". I just installed 4.36 this morning(it was running 4.26), so I'll wait to see what the next batch does. I checked and all 4 files are in the einstein directory. (even read the app_info.xml to be sure. The machine is an "Mobile" AMD64 3700 laptop with 1G of ram, and winxp home.
I forgot that I have boinc installed as a service on the ONLY windows machine I've got running ATM (the wifes laptop). In my defense,,,it was 3:30 this morning when I did it via VNC.
Well i've run my first couple of w/u through using the 4.36 optimized app, seems to be a lot quicker than original estimate (24-25 hrs).
I wonder would a SSSE3 optimized version improve times any more? I run a SSSE3 optimized s@h app and its around 25-40% faster (depends on angle ranges) than the standard app.
Well i've run my first couple of w/u through using the 4.36 optimized app, seems to be a lot quicker than original estimate (24-25 hrs).
I wonder would a SSSE3 optimized version improve times any more? I run a SSSE3 optimized s@h app and its around 25-40% faster (depends on angle ranges) than the standard app.
Yes, the initial estimate (before the first result is submitted) is just based on the BOINC benchmark and a rough estimate of the number of operations that the app communicates to BOINC, or in other words, ist's just an informed guess. BOINC will "learn" as you crunch more results to make better estimates.
I wonder would a SSSE3 optimized version improve times any more? I run a SSSE3 optimized s@h app and its around 25-40% faster (depends on angle ranges) than the standard app.
I wonder would a SSSE3 optimized version improve times any more? I run a SSSE3 optimized s@h app and its around 25-40% faster (depends on angle ranges) than the standard app.
FWIW, the new stock "power"
)
FWIW, the new stock "power" app seems to be inline with the mac app now...
if you look at the top host wich is a mac it completes a wu in about 14k sec to 16k sec.
ran some samples today on host running windows http://einsteinathome.org/host/1101157/tasks
HTH
Since "Re allowing" einstein
)
Since "Re allowing" einstein to run on hostid=825074, the very first two wus yielded "computation errors". I just installed 4.36 this morning(it was running 4.26), so I'll wait to see what the next batch does. I checked and all 4 files are in the einstein directory. (even read the app_info.xml to be sure. The machine is an "Mobile" AMD64 3700 laptop with 1G of ram, and winxp home.
Hmmm,,,,How would my kid say
)
Hmmm,,,,How would my kid say this?????
Oops, My bad
I forgot that I have boinc installed as a service on the ONLY windows machine I've got running ATM (the wifes laptop). In my defense,,,it was 3:30 this morning when I did it via VNC.
soooooo
nevermind.
I have 4.36 install and just
)
I have 4.36 install and just finished a work unit.
At the bottom of the result page it say version 4.26 but in the text towards the top, it says 4.36.
Anything I did wrong and need to change?
CLICK TO HELP BUILD
RE: I have 4.36 install and
)
You are fine. And 4.36 ran that unit. Once you are running results which downloaded after you had installed 4.36 the labels will catch up.
RE: RE: I have 4.36
)
Thanks.
CLICK TO HELP BUILD
Well i've run my first couple
)
Well i've run my first couple of w/u through using the 4.36 optimized app, seems to be a lot quicker than original estimate (24-25 hrs).
I wonder would a SSSE3 optimized version improve times any more? I run a SSSE3 optimized s@h app and its around 25-40% faster (depends on angle ranges) than the standard app.
BOINC blog
RE: Well i've run my first
)
Yes, the initial estimate (before the first result is submitted) is just based on the BOINC benchmark and a rough estimate of the number of operations that the app communicates to BOINC, or in other words, ist's just an informed guess. BOINC will "learn" as you crunch more results to make better estimates.
RE: I wonder would a SSSE3
)
See Message # 81998 ..
Regards
RE: RE: I wonder would a
)
Yep, I predicted that question would be asked... LOL