Very Computer

by **Chris Esse** » Sat, 13 Sep 1997 04:00:00

> Idealy I'm looking for the differance in floating point calculation speed
> for:

> Intel Pentium 200 (none MMX)
> Intel Pentium 200 MMX
> Intel Pentium Pro 200 (256k cache)
> AMD K6 200
> AMD K6 233

> Any others that people feel would be good for the job also appreciated.

Don't forget the DEC Alpha 533...should do the trick with a spreadsheet
or two...will cut down on kernel compiles no doubt too. :-)

by **M Hodgso** » Sat, 13 Sep 1997 04:00:00

OK, I'm new to the list, but have had a quick scan through the AMD K6
issue. It seems to me that providing you get a new and fixed chip, then
the AMD provides a great investment, offering good performance at a
reasonable price....

The thing is that I am looking to construct a machine dedicated to one
task. Like another poster I am wanting to peform chemistry simulations
which will be very floating point intensive... Graphics etc. are not at
all important. Whilst I appreciate that Intel chips have far better
floating point handeling, they are also more expensive. Has anyone out
there benchmarked the speed differance ?

Idealy I'm looking for the differance in floating point calculation speed
for:

Intel Pentium 200 (none MMX)
Intel Pentium 200 MMX
Intel Pentium Pro 200 (256k cache)
AMD K6 200
AMD K6 233

Any others that people feel would be good for the job also appreciated.

Please send replies to:

Thanks in advance
-Michael

by **Christopher Brow** » Sat, 13 Sep 1997 04:00:00

Quote:>> Idealy I'm looking for the differance in floating point calculation speed
>> for:

>> Intel Pentium 200 (none MMX)
>> Intel Pentium 200 MMX
>> Intel Pentium Pro 200 (256k cache)
>> AMD K6 200
>> AMD K6 233

>> Any others that people feel would be good for the job also appreciated.

>Don't forget the DEC Alpha 533...should do the trick with a spreadsheet
>or two...will cut down on kernel compiles no doubt too. :-)

Not to be too picky, but spreadsheets and kernel compiles are not
generally very sensitive to the speed of the FPU. Kernel compiles are
*far* more sensitive to speed of integer operations, and spreadsheet
"kernels" typically do more work checking pointers to cells than they
do FP work. (And to be perhaps *too* picky, there aren't at this
point any "industrial-strength" native Alpha spreadsheet packages...
But I digress...)

All that being said, performance of even *relatively* FP intensive
applications may not get hurt badly by the slower Cyrix/AMD FPUs if
one spends the money saved on RAM. You may lose on the FP side, but
yet pick performance back up on the other side because there's an
extra 32MB of RAM in which to buffer calculated values...
--

take you today? A: Confutatis maledictis, flammis acribus *is...

by **Mark Hea** » Tue, 16 Sep 1997 04:00:00

: Idealy I'm looking for the differance in floating point calculation speed
: for:

: Intel Pentium Pro 200 (256k cache)
: AMD K6 200

According to Dave Parsons (Am I correct?) Who has benchmarked the FPU of
both the above chips with industrial strength apps. Claims that the 80521
(pro) is about twice as fast than the equivalently clocked K6. Plus its
very easy to get a 80521 to run at 240Mhz. Or 233Mhz At least.

How fast can you compress mp3s? :-)

--
-- mark heath - Netspace Online Systems. http://www.netspace.net.au/
Obnoxious Usenet Habits #28. Followup every post in a newsgroup ranking
them on a scale from 1to 10.
:wq

Very Computer

K6 vs Others Benchmarks

K6 vs Others Benchmarks

K6 vs Others Benchmarks

K6 vs Others Benchmarks

K6 vs Others Benchmarks