Jump to content
34 posts in this topic

Recommended Posts

I wrote a small Benchmark called "CoreBench" my goal is to finally demonstrate real speed of real-applications (worehouses/shops management) on MacOS and on Windows

For now i've collected a small amount of results from some friends...

Now what i'm asking is to write on this topic, the final score of Corebench with some system infos

 

like:

4150 Intel I7 2.6Ghz MacOS 10.6.7

 

or (better) to link the app-screenshot maded by CoreBench...

 

CoreBench.screenshot.1.jpg

 

the download link for latest Corebench is: http://www.wcn.it/CoreBench.html

 

(P.S. i dont' need result from MultiCore app... just CoreBench)

 

I'm very interested on scores from 10.6.2 (I think the MacOS speed 'is diminished after this release)

a comparative test on the same machine (10.6.2 and 10.6>2) would be great

 

Thanks in advance for any help/test

Updated build is 0.60-0.62

CoreBench.MacOS.vs.Windows.BenchMark.jpg

http://www.wcn.it/CoreBench.ReactOS.faster.than.Windows.Benchmark.jpg

Thanks in advance for any help/test

 

 

Hi.... these are my benchs results from both lion 11459e & win7 64bit sp1), my system info is in rig.

Cheers

 

 

UPDATE: oh i've lost to say that my q9300 intel quad processor is overclocked @3.446 mhz instead of standard 2.500mhz with 4 gb of ddr2 ram@919mhz... byeeeee

post-449896-1306764291_thumb.jpg

post-449896-1306764297_thumb.jpg

Hi.... these are my benchs results from both lion 11459e & win7 64bit sp1), my system info is in rig.

Cheers

 

Thanks very much! i will add your scores to "comparison screen" this evening.

P.S. OMG time to add "Lion" to corebench :mellow:

"I'm very interested on scores from 10.6.2 (I think the MacOS speed 'is diminished after this release)

a comparative test on the same machine (10.6.2 and 10.6>2) would be great"

 

Hhmm, do you really think that? Even OS X vs Win 7 , same CPU shows near same results, so the bench may be very hardwareclose and not uses much OS X speed diffs .

What kind of CPU code is gettings computed ? I ask because my 10.6.8, C2D 2.8 GHz / 3 MB Cache ( low end CPU today) runs not much (4159 points) slower than faster clocked CPUs?! Or misunderstood reailpoints as benchmark indicator (higher=faster) ?

THANKS

Bildschirmfoto_2011_05_30_um_19.44.51.jpg

Hhmm, do you really think that? Even OS X vs Win 7 , same CPU shows near same results, so the bench may be very hardwareclose and not uses much OS X speed diffs .

You're right, seems that a very small GCC app can do the same things on Windows or MacOS on the same time.

This morning I did not have enough data yet to confirm it.

now... i think that a small app (compiled using GCC/MINGW) that just do a Arrays swaps.. string search, simple floating point operations (like a small shop management app) run on Win or MacOS on the same way...

(i think also that a Microsoft Visual C app or a native Objective C app runs much faster on her native platform).. but this is not the scope of this small Benchmark.

 

corebench.results.MacOS.vs.Windows.jpg

an updated screenshot of today results...

3. Mac OS x 10.6.7 turbo x42

4 Win 7 turbo x42

really impressive

seems that the turbo mode can really speed-up (small i think) apps!

anyone else have this "options" on his MB?

( a lot of results... now i must add a scroll function on comparison window ;) for the 0.49 build )

 

Core2Duo E7300@2.66Ghz

OS X 10.6.7 (kernel in 64-bit mode)

umh... CoreBench is a 32Bit app... i don't know if a 32app run a bit slowly on 64bit mode... can you launch MacOS on i386 mode for a MacOS64 vs MacOS32 for 32Bit Apps? thx!

umh... CoreBench is a 32Bit app... i don't know if a 32app run a bit slowly on 64bit mode... can you launch MacOS on i386 mode for a MacOS64 vs MacOS32 for 32Bit Apps? thx!

I've updated my first post with the scores from OS X with the kernel in 32-bit mode. There's no real difference in the scores.

because there is almost no diference between my old pentium t3200 @ 2 GHZ and i5 or i7 :D

:) you are right, and i think your P4 3.2 works very fast compared to a brand new 2011-AtomDualcore

the point is that this a comparison benchmark "MacOS vs Windows" .. that simulate a software for Wharehouse/shop management ... i don't ask the benchmark to do a massive multicore zip compression (using 4/8 4.2Ghz cores? for compressing what?) or a massive fibonacci sequence analisys (for what again?) but... just to execute a REAL app, for a small SHOP... that has a BEGIN .. and and END... not something like: How many items my x4 4.2Ghz can rename on 5 seconds ( 2 Millions? maybe 4?... for what? no one need to rename 1 millions of items on a warehouse in a couple of seconds.)

 

a way to retrieve a speed measurement like other benchmarks is to apply the CoreBench result to an "hyperbole" where 5000 points corresponds to +infinite

 

well... using this approach ... the i7 930 scores probably change to an p4 3.2 * 10 score ( i try to guess)

But on real (work) world.. when the i7 is to give the his maximum power (using caching... multithreading.... and any sort of magical thing...) ... the app... just END... ( job is done :o )

 

Well .. this is what i think about my benchmark..

 

Startas, thanks for giving me the opportunity to express my "logic" :) and i promise you that for tomorrow the "Heavy Apps" vs "Small Apps" scores will be added to CoreBench.

 

P.S. an I7 is x4 with HT ( something like ... 4+4 cores ) .. and athe P4 3.2 is an X1 with HT ( something like 1+1 cores ) CoreBench do just a "single thread test"... 1/8 of the maximum i7 power ... and 50% of the P4 maximum power ( rough estimate ).

 

I've updated my first post with the scores from OS X with the kernel in 32-bit mode. There's no real difference in the scores.

We have finally established that a "small GCC 32-app" accomplish his work on the same time using a 64bit Kernel or a 32bit Kernel :) Thank you blackosx!

×
×
  • Create New...