Prev: interrupting for overflow and loop termination
Next: "Livermore Loops" on x86 Linux
From: schouwla on 8 Feb 2006 03:26
All,


Does anyone know how much performance speedup I can expect by using 64
bit C++ / Windows XP 64 bit over the 32 bit versions?


Did anyone test this under Visual Studio 2005 or Intel C++ 8.1/9.0 ?

My application domain is montecarlo simulations but any test would be
great.


Regards
Lars Schouw

From: Christoph Breitkopf on 8 Feb 2006 05:49
schouwla(a)yahoo.com writes:

> Does anyone know how much performance speedup I can expect by using 64
> bit C++ / Windows XP 64 bit over the 32 bit versions?

I assume with 64-bit you are referring to x86-64, not Itanium.

> Did anyone test this under Visual Studio 2005 or Intel C++ 8.1/9.0 ?
>
> My application domain is montecarlo simulations but any test would be
> great.

Many (most?) applications will run a bit slower, because of larger pointers.

You will get speedups only if your application is memory-constrained
in 32-bit and would benefit from a larger unsegmented address-space,
or if inner loops can benefit from the additional registers available
in 64-bit mode. This also depends on the compiler, of course.

Regards,
Chris
From: Stephen SM WONG on 8 Feb 2006 12:15
But x86-64 has some 64-bit operations which operate on
64-bit registers, which will speed up your program if you
use 64-bit integers a lot.

My 2 cents.
Stephen Wong @ Hong Kong

On Wed, 8 Feb 2006, Christoph Breitkopf wrote:

> schouwla(a)yahoo.com writes:
>
>> Does anyone know how much performance speedup I can expect by using 64
>> bit C++ / Windows XP 64 bit over the 32 bit versions?
>
> I assume with 64-bit you are referring to x86-64, not Itanium.
>
>> Did anyone test this under Visual Studio 2005 or Intel C++ 8.1/9.0 ?
>>
>> My application domain is montecarlo simulations but any test would be
>> great.
>
> Many (most?) applications will run a bit slower, because of larger pointers.
>
> You will get speedups only if your application is memory-constrained
> in 32-bit and would benefit from a larger unsegmented address-space,
> or if inner loops can benefit from the additional registers available
> in 64-bit mode. This also depends on the compiler, of course.
>
> Regards,
> Chris
>
From: Brian Hurt on 12 Feb 2006 09:05
Christoph Breitkopf <chris(a)chr-breitkopf.de> writes:

>schouwla(a)yahoo.com writes:

>> Does anyone know how much performance speedup I can expect by using 64
>> bit C++ / Windows XP 64 bit over the 32 bit versions?

>I assume with 64-bit you are referring to x86-64, not Itanium.

>> Did anyone test this under Visual Studio 2005 or Intel C++ 8.1/9.0 ?
>>
>> My application domain is montecarlo simulations but any test would be
>> great.

>Many (most?) applications will run a bit slower, because of larger pointers.

On the x86, you get the advantage of 8 new registers in going to
64-bit. This generally increases the speed of most programs by more
than enough to overcome the decreased cache hit ratios 64 bits
induces, which means that 64-bit code is generally 10-15% faster than
the 32 bit code on the same hardware.

Brian

From: =?ISO-8859-1?Q?Niels_J=F8rgen_Kruse?= on 12 Feb 2006 09:30
Brian Hurt <bhurt(a)AUTO> wrote:

> On the x86, you get the advantage of 8 new registers in going to
> 64-bit. This generally increases the speed of most programs by more
> than enough to overcome the decreased cache hit ratios 64 bits
> induces, which means that 64-bit code is generally 10-15% faster than
> the 32 bit code on the same hardware.

There is a difference between AMD and Intel CPUs here. On Intel, the 8
registers subtract from the pool of rename registers, on AMD it is use'm
or lose'm.

--
Mvh./Regards, Niels J?rgen Kruse, Vanl?se, Denmark
 |  Next  |  Last
Pages: 1 2 3 4 5 6 7
Prev: interrupting for overflow and loop termination
Next: "Livermore Loops" on x86 Linux