On Dec 15, 2006, at 17:53 , Ron wrote:
At 09:50 AM 12/15/2006, Greg Smith wrote:
On Fri, 15 Dec 2006, Merlin Moncure wrote:
The slower is probably due to the unroll loops switch which can
actually hurt code due to the larger footprint (less cache
coherency).
The cache issues are so important with current processors that I'd
suggest throwing -Os (optimize for size) into the mix people
test. That one may stack usefully with -O2, but probably not with
-O3 (3 includes optimizations that increase code size).
-Os
Optimize for size. -Os enables all -O2 optimizations that do not
typically increase code size. It also performs further
optimizations designed to reduce code size.
So far I have been compiling PostgreSQL and running my pgbench script
manually, but this makes me want to modify my script to run pgbench
automatically using all possible permutations of a set of compiler
flags.
Last I tried GCC to produce 32-bit code on this Opteron system,
though, it complained about the lack of a compiler; can I persuade it
to generate 32-bit code (or 32-bit pointers for that matter) without
going the cross-compilation route?
Alexander.
