I have compiled mono with the LLVM back-end and run some tests.
Clearly there remains work to be done, but there were some cases where I
obtained executions speeds equal to an equivalent optimized C version. This
is really exciting for someone like me.
It is my belief / understanding that
Hey,
1) Is the LLVM back-end intended to be a long term solution to code
optimization for performance sensitive code in mono?
2) How confident are you that the issues described can be resolved without
kludge?
3) Would success with the LLVM back-end necessarily curtail efforts to
improve
Hello,
1) Is the LLVM back-end intended to be a long term solution to code
optimization for performance sensitive code in mono?
So far it looks very promising, and if things continue to evolve nicely in this
direction, it would make sense.
3) Would success with the LLVM back-end necessarily