Reply To: gcc 4.4 (4.4.2) compile error with mode=release related to -finline-limit=20000 compile option [SOLVED]

[Anonymous]

Benchmark) I’m not aware of anyone, anyplace, anytime actually running that code – only the benchmarking server. There’s just some pathing we’re missing or something. It failed for me on the zinc ion residue type, of all things…

National Lab) I feel your pain. My lab has access to a BlueGene we can’t use, because the admins won’t update the XLC compiler. It’s not really a bug we can fix – the code it can’t handle is our implementation of owning pointers, which are so widespread in the code that “fixing it” would be forking Rosetta wholesale. The Baker lab has access to a different BlueGene that works fine, grumble grumble. Most of the stuff my lab does requires enough memory that BlueGenes wouldn’t work anyway.