AMD vs. Intel Floating Point

Kofi Anim-Appiah

unread,

Jun 15, 2001, 5:59:51 AM6/15/01

to

Has anyone experienced any floating-point problems on systems using any of
the AMD processors (Duron, Athlon, Thunderbird)? I'm considering buying a
new processor and would like to know if the AMDs are as accurate with
floating point math as the Pentiums

Thanks.

Kofi

Jens-Peer Kuska

unread,

Jun 19, 2001, 5:55:14 AM6/19/01

to

Hi,

a foolish question. An Athlon is typical faster in
floting point math than an Intel CPU. Only the
Itanium has better floting point performance.

Regards
Jens

Bartek Garda

unread,

Jun 20, 2001, 4:50:24 AM6/20/01

to

I work on Athlon 800Mhz with Mathematica4.0 and never faced any problems!

Użytkownik "Jens-Peer Kuska" <ku...@informatik.uni-leipzig.de> napisał w
wiadomości news:9gn7i2$ere$1...@smc.vnet.net...

Orestis Vantzos

unread,

Jun 20, 2001, 4:56:32 AM6/20/01

to

In what sense is it foolish? Seti@home for instance, which relies heavily on
floating point operations, does work slower on AMD chips...
Orestis
"Jens-Peer Kuska" <ku...@informatik.uni-leipzig.de> wrote in message
news:9gn7i2$ere$1...@smc.vnet.net...

Jens-Peer Kuska

unread,

Jun 21, 2001, 2:00:44 AM6/21/01

to

Hi,

I have never seen an benchmark where a Intel CPU of the same
clock speed beats an Athlon.

You may look at:

http://fampm201.tu-graz.ac.at/karl/timings40.html

and see that the fastest 5 (five !) entries
are Athlon CPU's. Since an Athlon has one floating point
pipeline more than an Intel CPU it is foolish to ask

"Work three workers more than two ?"

I would realy like to see why Seti@home is slower on an
Athlon -- but it is definetly *not* the floating point
performance.

BTW since when where *screen saver* used as floating point
benchmarks ?

Regards
Jens

Morfeas79a

unread,

Jun 21, 2001, 2:07:50 AM6/21/01

to

Kofi
as it is well known the AMD processors up untill the model of K6-3D
have serious problems in their floating point operations - this can be
observed by running programs with great CPU load like SETI@home, the
time for a AMD computer to finish one work unit is about twice as big
as this in an Intel computer running on the same MHz. The problem has
been solved in later models. Of course all this is not known to
Mr.Kuska who thinks that 90% of the questions sent in this newsgroup
are of trivial nature or in anycase foolish.

Regards
Jim

Orestis Vantzos

unread,

Jun 22, 2001, 2:20:16 AM6/22/01

to

I am not a hardware spec, I write what I know to be true. Seti@home runs
better on Intel processors and I have seen that with my own eyes. Why I
consider it a quasi-valid benchmark for floating-point operations? Well, all
it really does is certain mathematical transformations (Fast Fourier among
them)...it has to depend on floating -point operations. Could the difference
be in the math-coprocessor? Errors in AMD floating-point operations that
force them to repeat a portion of the operations? I can't really tell,
hardware is not my field...
Orestis
PS. Jens loosen up...

"Jens-Peer Kuska" <ku...@informatik.uni-leipzig.de> wrote in message

news:9gs2ic$k3p$1...@smc.vnet.net...

Jens-Peer Kuska

unread,

Jun 22, 2001, 2:25:21 AM6/22/01

to

Hi,

a) the Mathematica speed comparsion from

http://fampm201.tu-graz.ac.at/karl/timings40.html

is posted regular in this news group
b) on the www-site of *this* news group

http://smc.vnet.net/mathgroup.html

the second head line is a link to various speed
comparsions found at

http://smc.vnet.net/mathbench.html

and it is quite natural to assume, that a poster to a
news-group has visited the newsgroup hompage and
is able to read and understand the headings on a page
that begins with:

-----------------------------------------------------------
>Designed by S. Christensen.
>
>MathGroup
>
>The Email Group for Mathematica Users
>
>Comparison of Mathematica on Various Computers
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>MathGroup is now linked to the moderated newsgroup
>
>comp.soft-sys.math.mathematica
>
> on the Internet. Contact your local system administrator
> to find out how to read this new group.
-----------------------------------------------------------

*and* it is quite natural to assume that a poster to the news group
has read the group rules (on the same page) one of it say:

>PLEASE SEARCH THE ARCHIVES BEFORE YOU ASK WHAT MIGHT BE A COMMON QUESTION.
>See the links above for this.

It must be also sayed, that the Mathematica speed depends in the
most (symbolic) applications not on the floating point power
of the CPU. The most actions performed by Mathematica are pointer
operations with it's internal data structures. I would assume that
80-90 % of Mathematica's CPU load are pure interger operations.
High precision calculations, symbolic operations, operations with
integers, rationals ... all that don't use the floating point hardware.

It depends shaply on the application how much floating point operations
are used. But when a Mathematica function has such a huge floating
point
load it is always better to write a MathLink program.

Regards
Jens

Dimitris Skipis

unread,

Jun 22, 2001, 2:26:22 AM6/22/01

to

> BTW since when where *screen saver* used as floating point
> benchmarks ?
>
> Regards
> Jens

A foolish question.
Of course SETI@home -which is not an ordinary screensaver, but a distributed
program to aid the SETI project through fine data analysis- is a major
floating point benchmark
since it uses 100% of your processor doing FFT, searching for
pulses and triplets etc.
If you can recall, when the G4 was introduced by Apple, it was publicly
tested on SETI@home.
For any further questions on the SETI@home program and project, visit
www.setiathome.ssl.berkely.edu

Regards
Jim