HINT for LinuxPPC

HINT for LinuxPPC

[David DeHaven]

Has anyone compiled and run the Hint benchmarks for LinuxPPC? I need to 
run some benchmarks on different processor and cache configurations.

-DrD-

[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: HINT for LinuxPPC

[Nicholas Coult]

On Thu, 8 Jul 1999, David DeHaven wrote:

> 
> 
> Has anyone compiled and run the Hint benchmarks for LinuxPPC? I need to 
> run some benchmarks on different processor and cache configurations.
> 
> -DrD-
> 

Yes, I ran the Hint benchmark on my LinuxPPC machine.  My hint benchmark
page compares the results to a PII/300 running Linux:

http://www.ima.umn.edu/~coult/hint.html

Dr. Nicholas Coult, Research Associate
Institute for Mathematics and Its Applications, University of Minnesota
Home Page: www.ima.umn.edu/~coult              Office: 555 Vincent Hall
Email: coult@ima.umn.edu      Phone: (612) 624-2527 Fax: (612) 624-4163


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: HINT for LinuxPPC

[David Edelsohn]

>>>>> Nicholas Coult writes:

Nicholas> Yes, I ran the Hint benchmark on my LinuxPPC machine.  My hint benchmark
Nicholas> page compares the results to a PII/300 running Linux:

Nicholas> http://www.ima.umn.edu/~coult/hint.html

	egcs-1.1.2 does not actually know about the PPC750, so -mcpu=750
did not actually help.  gcc-2.95 knows about PPC750 scheduling, makes
better use of indexed addressing, and utilizes the CTR register for loops
which should improve performance.  Other PowerPC improvement are coming
down the pike as soon as they can be cleaned up and added to GCC.

	Also, once the new p2 backend contributed to GCC by Cygnus is
installed and stable, that should provide a better test of PII
performance. 

David

[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]