cfallin opened PR #10 from cranelift-backend-transition
to main
:
This RFC proposes to switch the default Cranelift compiler backend for the
cranelift-codegen
crate (i.e., use of Cranelift as a library) and for Wasmtime from the old x86-64 backend to the new x86-64 backend. Details on how we should decide, what steps we should take to build confidence, ensure backwards compatibility, and gradually switch over are proposed.
bjorn3 submitted PR Review.
bjorn3 created PR Review Comment:
#1140](bjorn3/rustc_codegen_cranelift#1140)) have already switched to using the new backend by default (Lucet) or as only (cg\_clif) option.
bjorn3 submitted PR Review.
bjorn3 created PR Review Comment:
incompleteness because it is not a regression overall.
bjorn3 submitted PR Review.
bjorn3 created PR Review Comment:
Maybe saltwater by @jyn514? It has been a couple of months since the last change through it seems.
bjorn3 submitted PR Review.
jyn514 submitted PR Review.
jyn514 created PR Review Comment:
I don't really work much on it any more, it uses like a 6 month old version on cranelift. I wouldn't make decisions based on it.
jyn514 edited PR Review Comment.
joshtriplett submitted PR Review.
joshtriplett created PR Review Comment:
This transition's rationale largely derives from the rationale for the
cfallin updated PR #10 from cranelift-backend-transition
to main
.
cfallin submitted PR Review.
cfallin created PR Review Comment:
Thanks!
cfallin submitted PR Review.
cfallin created PR Review Comment:
Thanks!
cfallin submitted PR Review.
cfallin created PR Review Comment:
Fixed, thanks!
bnjbvr submitted PR Review.
bnjbvr submitted PR Review.
bnjbvr created PR Review Comment:
In order to help guiding the answer to this question, how about another question: from the performance analysis that we've made on the new backend, it would be interesting to get a rough idea of the speedups we can reach, and with which associated cost? What I mean is, if we think we can get e.g. a 10% speedup easily (famous last words, but imagine in e.g. one or two months of work), then it makes it easier to accept a degradation in the short-term.
jedisct1 submitted PR Review.
pchickey submitted PR Review.
abrown submitted PR Review.
jlb6740 submitted PR Review.
bnjbvr submitted PR Review.
cfallin merged PR #10.
cfallin edited PR #10 from cranelift-backend-transition
to main
:
This RFC proposes to switch the default Cranelift compiler backend for the
cranelift-codegen
crate (i.e., use of Cranelift as a library) and for Wasmtime from the old x86-64 backend to the new x86-64 backend. Details on how we should decide, what steps we should take to build confidence, ensure backwards compatibility, and gradually switch over are proposed.
Last updated: Dec 23 2024 at 12:05 UTC