> Are all these values posted here taken from einstein_4.44 on Windows, or are
> there some Linux ones between them? On our systems (mainly Intel) the speedup
> under Linux (gcc) was larger than that of Windows (MSVC), but we've never ever
> seen a slowdown. The two (Windows) machines I'm sitting in front of (2,6GHz P4
> and 1,8GHz Centrino) consistently show a speedup of about 20%. On Linux
> Systems we've seen up to 80%.
>
> We are working to get the Mac graphics running, the next Mac app will have the
> faster "science code", too.
>
> BM
All mine were Windows. Bounced back and forth on the AMD 64 with Suse 64 and Win 64, but am too busy to mess with it much right now, so just left it on Win 64.
Glad to hear about the upcoming Mac improvements. I'll probably be adding another Mac in the next month.
Hi @ all, all is running superb!
No problems here, I'm realy happy about this. On my AMD 2800 XP+ @ 2300 MHz, no changing of the calculations times during the different versions :).
Hmm, just went and checked my results log and I wasn't hallucinating, The 2.66P4 laptop was doing 50-some minute units under the old core (Windows XP SP1) and now is 70-some minute units. For reference it's a Compaq nx9010. The host ID for the machine is 984 if the team would like to check the results (could it have been exitting early on an unreported error?)
No, the results of that host look all ok. However, the cpu hours per WU dropped from 1.8 to 1.3 since using einstein_4.44. Maybe your machine is busy with something else? Maybe with somethig it shouldn't do?
I'm curious as to how it can show 1.8 cpu hours but list the running time as 50-some minutes. I didn't actually check the clock hours with the reported times perhaps it was merely mis-reporting the run time? The current run-time seems in line with those others are reporting, it's the old results that seem to be unusual.
Hm, I can see that the receiving times of 4.35 WUs are about 105 mins apart, while now are about 75 mins. So this would figure, too.
I know that the estimated time is calculated based on the benchmarks (and others) and these are wrong in the 4.13 client for Windows. Maybe you just had a lucky day when your initital benchmark was ran and it now had been ran again?
Sorry, I'm only guessing, but I really don't know what happened.
> Are all these values posted
)
> Are all these values posted here taken from einstein_4.44 on Windows, or are
> there some Linux ones between them? On our systems (mainly Intel) the speedup
> under Linux (gcc) was larger than that of Windows (MSVC), but we've never ever
> seen a slowdown. The two (Windows) machines I'm sitting in front of (2,6GHz P4
> and 1,8GHz Centrino) consistently show a speedup of about 20%. On Linux
> Systems we've seen up to 80%.
>
> We are working to get the Mac graphics running, the next Mac app will have the
> faster "science code", too.
>
> BM
All mine were Windows. Bounced back and forth on the AMD 64 with Suse 64 and Win 64, but am too busy to mess with it much right now, so just left it on Win 64.
Glad to hear about the upcoming Mac improvements. I'll probably be adding another Mac in the next month.
Team MacNN - The best Macintosh team ever.
On my boxes, I see : P4 with
)
On my boxes, I see : P4 with HT 30%-40% faster, on P4 without HT 20% faster with 4.44 compared against 4.35
Supporting BOINC, a great concept !
Thanks a lot for the
)
Thanks a lot for the feedback. Anyone else noticed a slowdown like senator2?
BM
BM
Hi @ all, all is running
)
Hi @ all, all is running superb!
No problems here, I'm realy happy about this. On my AMD 2800 XP+ @ 2300 MHz, no changing of the calculations times during the different versions :).
Greetings from Germany
Basti
Join Ad Astra
Hmm, just went and checked
)
Hmm, just went and checked my results log and I wasn't hallucinating, The 2.66P4 laptop was doing 50-some minute units under the old core (Windows XP SP1) and now is 70-some minute units. For reference it's a Compaq nx9010. The host ID for the machine is 984 if the team would like to check the results (could it have been exitting early on an unreported error?)
No, the results of that host
)
No, the results of that host look all ok. However, the cpu hours per WU dropped from 1.8 to 1.3 since using einstein_4.44. Maybe your machine is busy with something else? Maybe with somethig it shouldn't do?
BM
BM
I'm curious as to how it
)
I'm curious as to how it can show 1.8 cpu hours but list the running time as 50-some minutes. I didn't actually check the clock hours with the reported times perhaps it was merely mis-reporting the run time? The current run-time seems in line with those others are reporting, it's the old results that seem to be unusual.
Hm, I can see that the
)
Hm, I can see that the receiving times of 4.35 WUs are about 105 mins apart, while now are about 75 mins. So this would figure, too.
I know that the estimated time is calculated based on the benchmarks (and others) and these are wrong in the 4.13 client for Windows. Maybe you just had a lucky day when your initital benchmark was ran and it now had been ran again?
Sorry, I'm only guessing, but I really don't know what happened.
BM
BM
Sounds reasonable to me
)
Sounds reasonable to me