[STREAM] Powermac G5 2.5MHz gcc4 results

Hello,

it looks like gcc4 now produces better results than gcc3.3, or that the
memory subsystem of the PCIe-PowerMacs is better than that of the older G5
generation - or a bit of both:

$ ./stream_o_apple
-------------------------------------------------------------
STREAM version $Revision: 5.6 $
-------------------------------------------------------------
This system uses 8 bytes per DOUBLE PRECISION word.
-------------------------------------------------------------
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
-------------------------------------------------------------
Printing one line per active thread....
-------------------------------------------------------------
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 11884 microseconds.
   (= 11884 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
-------------------------------------------------------------
WARNING -- The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
-------------------------------------------------------------
Function Rate (MB/s) Avg time Min time Max time
Copy: 2905.1456 0.0111 0.0110 0.0112
Scale: 2878.1706 0.0112 0.0111 0.0112
Add: 3304.1734 0.0146 0.0145 0.0147
Triad: 3301.2477 0.0146 0.0145 0.0146
-------------------------------------------------------------
Solution Validates
-------------------------------------------------------------

This was on a quad G5 PowerMac with two dual-core 970MP CPUs at 2.5MHz.

Compiled with Apple's version of gcc4 which is supplied with Xcode 2.4. The
gcc4.1 included in the Macports collection didn't give significantly better
results. Just like the -fast option provided by Apple's own gcc.

As soon as the new gcc4.2 with OpenMP support will be available, I will try
the multi-CPU test.

And I would love to run a binary compiled with the Absoft (ex-IBM) compiler,
but I don't have one... :-(

Sincerely,

Wolfgang Keller

--
My email-address is correct.
Do NOT remove ".nospam" to reply.

Received on Mon Sep 25 09:50:02 2006