Serious performance problems

[Paolo Carlini]

Hi,

Colin Fox wrote:

> This is alarming, to say the least. Is this a known issue?

I believe it is, and Loren James Rittle worked out a fix (should be
http://gcc.gnu.org/ml/libstdc++/2001-07/msg00019.html ) which is *already*
on the mainline code (Gcc3.1 experimental).

Yesterday I tried a snapshot of 3.1 (
http://www.netwinder.org/build/gcc.html ) exactly to test the improvements
and, indeed, it looks like 3.1 is now much *faster* than the old Gcc
compilers!

Cheers,
Paolo.