ponderous 'make world' times post GCC 3.3...

[Harti Brandt]

On Wed, 13 Aug 2003, Steve Ames wrote:

SA>On Wed, Aug 13, 2003 at 05:51:49PM +0200, Harti Brandt wrote:
SA>> The problem is, that compiling just anything takes 3 time longer on an
SA>> actual kernel, than on one from June 1st. Same world, same kernel config,
SA>> same compiler (3.2.2).
SA>
SA>Not arguing that. I'm suggesting that perhaps one of the kernel debug
SA>options is pessimizing compiles across the board. It took me 12 hours
SA>to do a 'make world' when I had debugging options enabled in my kernel.
SA>I took them out and now I can 'make world' in 4.5 hours. I would expect
SA>that ratio to hold for all compiles and not just 'make world'.
SA>
SA>so.. running with a kernel with debug options causes my compiles to
SA>be almost 3 times as long. That's all I'm saying. What does your
SA>kernel config look like? Do you have all of the debug stuff turned
SA>on?

I have actually two machines and two kernels: one machine runs with
WITNESS and DIAGNOSTICS and the other without that. On both machines
the current kernel is three times slower than the old kernel. Of course
the debugging kernels are even slower than the non-debugging ones. But,
the three times ratio between old and new kernels is the same on both
machines.

harti
-- 
harti brandt,
http://www.fokus.fraunhofer.de/research/cc/cats/employees/hartmut.brandt/private
brandt at fokus.fraunhofer.de, harti at freebsd.org