Why alternatives failed? #54
-
|
You've mentioned you analyzed MIR jit, libfirm and other alternatives. Why did they fail for you? Especially interesting reasoning about MIR. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
You may find some comments in my presentation. See https://www.researchgate.net/publication/374470404_IR_JIT_Framework_a_base_for_the_next_generation_JIT_for_PHP (slide 10).
MIR is an interesting project and I like its ideas, but at time I analysed it I wasn't satisfied with generated code quality, absence of tooling (disassembler, gdb, perf), missing of x86 32-bit support. Finally MIR is actually slow (see vnmakarov/mir#84) Do you use MIR? |
Beta Was this translation helpful? Give feedback.
You may find some comments in my presentation. See https://www.researchgate.net/publication/374470404_IR_JIT_Framework_a_base_for_the_next_generation_JIT_for_PHP (slide 10).
MIR is an interesting project and I like its ideas, but at time I analysed it I wasn't satisfied with generated code quality, absence of tooling (disassembler, gdb, perf), missing of x86 32-bit support. Finally MIR is actually slow (see vnmakarov/mir#84)
Do you use MIR?