I noticed that Release mode compilation is Non-Blitz by default. Why is this? After some thinking it seems that with Blitz even the compiler optimizations could be better because the compiler sees more functions together, and can make better inlining decisions.
Hard reason: MSC8 crashes in BLITZ / release (perhaps because of too long files fed into the optimizer)
Another reason: BLITZ tends to force much more .obj to be linked into target - even with MSC which performs function level linking, or with MINGW / ULD which does roughly the same, BLITZed release targets are significantly longer.
Members
Pages
Help
Register
Login
Home
