From: Chris Pemberton on
On 8/12/2010 4:30 AM, Scot T. Martin wrote:
> This issue is more subtle than outlined in the last email because I ran
> Benchmark[] on yet another computer that also has Windows 7 64-bit and got
> very good results. Here is a full comparison --- any insights on the slow
> one?
>
> **** FASTEST ****
> Intel(R) Core(TM) i5 CPU 650 @3.20GHz, 8.00 GB
> === System Information ===
> Machine Name: smartin-t1500pc
> System: Microsoft Windows (64-bit)
> Date: August 11, 2010
> Mathematica Version: 7.0.0
> Benchmark Result: 4.71
>
> smartin-t1500pc
> Microsoft Windows (64-bit) 18.3 1.05 0.36 0.52 1.22 1.67
> 0.22 0.50 1.81 2.11 1.08 1.84 1.34 1.11 1.59
> 1.89
>
>
> **** NEXT FASTEST ****
> Intel(R) Core(TM)2 Duo CPUE8500(a)3.16 Ghz, 2.75 GB RAM
> === System Information ===
> Machine Name: martin2
> System: Microsoft Windows (32-bit)
> Date: August 11, 2010
> Mathematica Version: 7.0.0
> Benchmark Result: 4.12
>
>
> martin2
> Microsoft Windows (32-bit) 20.9 1.20 0.59 0.83 1.25 1.89 0.44
> 0.77 2.61 1.83 1.42 2.05 1.45 1.25 1.42 1.91
>
>
>
> *** WAY TOO SLOW (new system) ***
> Intel(R) Xeon(R) CPU,W3680(a)3.33 GHz, 8.00 GB install, Windows 7 professional
> 64 bit
> === System Information ===
> Machine Name: profmartinoffic
> System: Microsoft Windows (64-bit)
> Date: August 11, 2010
> Mathematica Version: 7.0.0
> Benchmark Result: 1.13
>
> profmartinoffic
> Microsoft Windows (64-bit) 76.2 1.87 0.33 0.39 26.10 1.83 0.23
> 0.48 1.89 1.70 0.84 1.72 1.05 1.15 23.80 12.80
>
>
>
> original email:
>
> Folks, just installed Mathematica on new Windows 7 64-bit machine and
> Benchmark[] is much slower than on 32-bit Windows XP.
> MathematicaMark7 Rating of new system is 1.13 compared to 4.12 of old
> system.
>
> Any explanations for this? Details are below. Are many of Mathematica's
> routines still in 32-bit and thus slower on 64-bit
> machine? Is this an issue related to Windows XP compared to Windows 7
> vis-a-vis Mathematica?
>
> Needless to say, new system is much better in terms of CPU, memory, etc.,
> so this result is very surprising to me. I'd like to
> know what I can adjust to get a blazing Benchmark!
>
>
>
Perhaps it is not Mathematica that is working poorly; perhaps something
is mis-configured with your machine? You can test your machine with
free trial software from passmark: (www.passmark.com) and then compare
your results against thousands of other machines. Might be useful if a
problem exists independent of Mathematica.