> There are some separate issues here: > > 1) Big slowdown with introduction of linkonce > processing > 2) Performance of ld vs gld before the big > slowdown > 3) Unknown performance of ld group processing > when facing > super large C++ objects.
I don't understand what (3) refers to -- maybe it doesn't matter. The link task I sent you is the worst one we have. I've now tried GCC 4.3.2 with the snv_104 linker, and it doesn't make much difference (if any) -- the link time is still in the 17 - 18 minute range. Maybe the changes that you thought were in 4.3 are actually in 4.4? BTW, it's easy enough for me to switch back to the snv_86 linker. So there's no particular rush here. -- Scott -- This message posted from opensolaris.org
