x64: Different results with simd depending on optimization level #3336
Description
This input file:
(module
(type (;0;) (func (result v128)))
(func (;0;) (type 0) (result v128)
v128.const i32x4 0x733c3e67 0x3c3e6776 0x3e677673 0x6776733c
i64x2.abs
i64x2.bitmask
i8x16.splat
v128.const i32x4 0x733c3e67 0x3c3e6776 0x3e677673 0x6776733c
i64x2.ge_s
f32x4.floor
v128.not
i16x8.extadd_pairwise_i8x16_u)
(export "x" (func 0)))currently yields:
$ cargo run -q testcase0.wat --wasm-features all --invoke x --opt-level 0
warning: using `--invoke` with a function that returns values is experimental and may break in the future
0
$ cargo -q run testcase0.wat --wasm-features all --invoke x --opt-level 2
warning: using `--invoke` with a function that returns values is experimental and may break in the future
2648111804832205941349218734877573630