557: Update loader-kernel feature name, fix compilation track state r=syrusakbary a=bjfish
- feature name `loader:kernel` -> `loader-kernel` (no colons)
- add track state `false` to loader kernel default, @losfair let me know if this default is incorrect
- cargo fmt
Co-authored-by: Brandon Fish <brandon.j.fish@gmail.com>
552: Allow disabling state tracking for faster startup. r=bjfish a=losfair
~10% compilation speedup when `--no-track-state` is enabled.
Co-authored-by: losfair <zhy20000919@hotmail.com>
551: Try a new list of optimization passes. r=nlewycky a=nlewycky
A few notes:
a) the inliner doesn't help because all the calls are indirect and not even opt -O2 can figure out which functions they're actually calling.
b) aggressive instruction combining is not a super-set of the instruction combiner. Instcombine is made up of a large number (probably 10,000s) of patterns, and some particularly slow ones were taken out and moved to the aggressive instruction combiner. Aggressive instcombine *only* runs that handful of optimizations, which fired zero times on our example wasm files.
c) NewGVN is not ready for production, it has asserts that fire when building sqlite or cowsay. This is why sqlite didn't build with the llvm backend.
d) Scalar-replacement-of-aggregates (sroa) is a strict superset of promote-memory-to-registers (mem2reg), and you probably want sroa because it's usually faster. It also fires 10,000s more times than mem2reg on lua.wasm.
e) Aggressive-dead-code-elimination was only deleting as much regular dead-code-elimination, but is slower because it depends on a postdominator tree (PDT) analysis that. Other passes don't need PDT so we'll have to build it for just this one pass (as opposed to regular dominator-tree which is reused by many passes). I've replaced this with bit-tracking dead-code-elimination which deletes more code than dce/adce.
Co-authored-by: Nick Lewycky <nicholas@mxc.ca>
Co-authored-by: nlewycky <nicholas@mxc.ca>
Fix typo in function name. Use two fcmp instructions instead of unpacking the bits of the IEEE float and using integer arithmetic to determine details about its value.
A few notes:
a) the inliner doesn't help because all the calls are indirect and not even opt -O2 can figure out which functions they're actually calling.
b) aggressive instruction combining is not a super-set of the instruction combiner. Instcombine is made up of a large number (probably 10,000s) of patterns, and some particularly slow ones were taken out and moved to the aggressive instruction combiner. Aggressive instcombine *only* runs that handful of optimizations, which fired zero times on our example wasm files.
c) NewGVN is not ready for production, it has asserts that fire when building sqlite or cowsay. This is why sqlite didn't build with the llvm backend.
d) Scalar-replacement-of-aggregates (sroa) is a strict superset of promote-memory-to-registers (mem2reg), and you probably want sroa because it's usually faster. It also fires 10,000s more times than mem2reg on lua.wasm.
e) Aggressive-dead-code-elimination was only deleting as much regular dead-code-elimination, but is slower because it depends on a postdominator tree (PDT) analysis that. Other passes don't need PDT so we'll have to build it for just this one pass (as opposed to regular dominator-tree which is reused by many passes). I've replaced this with bit-tracking dead-code-elimination which deletes more code than dce/adce.
543: update version numbers to 0.5.5 r=MarkMcCaskey a=MarkMcCaskey
544: Use bitcast instead of alloca+load+ptrcast+store sequence. r=MarkMcCaskey a=nlewycky
Co-authored-by: Mark McCaskey <mark@wasmer.io>
Co-authored-by: Mark McCaskey <markmccaskey@users.noreply.github.com>
Co-authored-by: Nick Lewycky <nick@wasmer.io>
Co-authored-by: nlewycky <nick@wasmer.io>
