[compilerdev]

See my answer above, it's because the build was LTCG and in the best case at most 4 cores were used during optimization/code geneation/linking.

[glandium]

That's not what I'm talking about, though. Compiling large C++ code requires much more memory than compiling C, which, on a TR and its NUMA architecture, can make significant differences. And we do know that Linux tends to work better on NUMA than Windows, so comparing a C compile on Linux to a C++ compile on Windows, even if it didn't have the LTCG setup problem, would still be apples to oranges.