tor callgrinds

[Christopher Layne]

On Fri, Feb 16, 2007 at 02:00:00PM -0800, Christopher Layne wrote:
> Thought you guys might find this interesting. I did a couple of callgrind
> runs on 2 different tor builds, 1 using -Os and the other using -O3. The

So did a bit more research on spec'ing which cost models are default in
callgrind and now have it logging jumps, asm instructions, and l1/l2/dram
performance counters in the simulator.  If anyone is interested on the
machine specifically it's a 2.1 ghz Celeron-D (Prescott) running under
Linux 2.6.20. I've rebuilt openssl, libz, and libevent with cranked up
optimization/debug on, so more interesting things to look at.

Currently doing another long grind.

Preemptively, from just a short run, I do notice a couple of places where
zlib itself could be optimized to provide better performance/throughput,
but no reason to right home yet.

I'm also interested on the impact of -Os vs -O3 overall, and on a CPU
such as this one with a rather small cache - seeing that these days core
throughput is being more cache limited than cycle limited.

-cl