Boinc Client 4.25
4 Mar 2005 12:00:35 UTC
Topic 188249
(moderation:
Boinc Clients 4.25 are available >> http://boinc.berkeley.edu/download.php?dev=1
upgrade from 4.24 to 4.25 was no problem. everythink looks fine, but i think it works a little bit "slower".
Benchmark results (athlon 2900+, winxp sp2):
2005-03-04 12:51:54 [---] Number of CPUs: 1
2005-03-04 12:51:54 [---] 1965 double precision MIPS (Whetstone) per CPU
2005-03-04 12:51:54 [---] 3334 integer MIPS (Dhrystone) per CPU
greetz from old europe
Boinc Client 4.25
)
> Boinc Clients 4.25 are available >>
> http://boinc.berkeley.edu/download.php?dev=1
>
> upgrade from 4.24 to 4.25 was no problem. everythink looks fine, but i think
> it works a little bit "slower".
>
> Benchmark results (athlon 2900+, winxp sp2):
> 2005-03-04 12:51:54 [---] Number of CPUs: 1
> 2005-03-04 12:51:54 [---] 1965 double precision MIPS (Whetstone) per CPU
> 2005-03-04 12:51:54 [---] 3334 integer MIPS (Dhrystone) per CPU
>
> greetz from old europe
>
hm, my benchmarks are much slower... i used to have 1800 / 4300 and now i have 1800 / 3000
instead of claiming about 85 credits (what i almost everytime got granted in the end) it now only claims 70 credits....
The benchmarks will vary
)
The benchmarks will vary depending on what all your system may have running at the time.
"everythink looks fine, but i
)
"everythink looks fine, but i think it works a little bit "slower""
>>The reason that 4.25 is a bit slower on some operations is the stack trace code that has been put in to track down a crash that someone was reporting.
I thought it was going to be
)
I thought it was going to be slower because the Intel compiler for windows was "cheating" the benchmarks and relising that some sections of code were not doing anything and were simply "optimizing them out". The new benchmark code has apparently got code that actually gets run instead of skiped over...
At least that's the feeling I've been getting around the forums.
@MetalWarrior If you are
)
@MetalWarrior
If you are comparing to BOINC 4.1x then Biogenesis has given the answer. The 4.2x series has a corrected benchmark for windows (the were inflated in 4.1x).
@Biogenesis
They use the Micro$oft Visual Studio .NET (Visual C++ 7.0) not Intel. When they were using version 6, the compiler wasn't striping the benchmark code bare.
Professor Desty Nova
Researching Karma the Hard Way
Yes, the 4.25 client looks
)
Yes, the 4.25 client looks like a goer from the alpha side.
The release is probally scheduled for monday at this stage.
73 de Peter VK3AVE
> Yes, the 4.25 client looks
)
> Yes, the 4.25 client looks like a goer from the alpha side.
>
> The release is probally scheduled for monday at this stage.
>
>
>
And what will the new client bring? Any new enhancements over 4.19?
Hi You will have more
)
Hi
You will have more control over your projects.
You can suspend a result or an project and continue when you whant it.
You can abort a result when the deadline is over and so on.
greetz from Germany
Mike
Does anybody have any
)
Does anybody have any experience about installing 4.25 over 4.19? Any success stories? Does 4.19 have to be uninstalled first? Are WUs lost?
> Does anybody have any
)
> Does anybody have any experience about installing 4.25 over 4.19? Any success
> stories? Does 4.19 have to be uninstalled first? Are WUs lost?
==========
Yes you have to un-install v4.19 first before installing v4.25, the Installer will even tell you that if you try to install it over v4.19 ... I've been running v4.25 for about a week now with no problems on all my Computers ...