>>The cstream.c uses a different timing routine and
>>works fine (although it gives very different performance on our SGI
>>Indy than stream_d for two of the tests).
>Better or worse?
I looked more carefully at it, and noticed that the stream_d program
has a default N of 1000000 while the cstream has a default N of
(1023*1024). This caused the tests for cstream that had two
right-hand-side components to conflict in the cache lowering
performance significantly. Since the fortran version had a default N
of 2000000, I used the decimal N for the two programs and got
This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:04 CDT