Encode [u|s]widen_low for x86

This commit is contained in:
Andrew Brown
2020-07-07 16:13:50 -07:00
parent fafef7db77
commit c8ddf8a34c
5 changed files with 103 additions and 10 deletions

View File

@@ -1669,6 +1669,7 @@ fn define_simd(
let ssub_sat = shared.by_name("ssub_sat");
let store = shared.by_name("store");
let store_complex = shared.by_name("store_complex");
let swiden_low = shared.by_name("swiden_low");
let uadd_sat = shared.by_name("uadd_sat");
let uload8x8 = shared.by_name("uload8x8");
let uload8x8_complex = shared.by_name("uload8x8_complex");
@@ -1678,6 +1679,7 @@ fn define_simd(
let uload32x2_complex = shared.by_name("uload32x2_complex");
let snarrow = shared.by_name("snarrow");
let unarrow = shared.by_name("unarrow");
let uwiden_low = shared.by_name("uwiden_low");
let ushr_imm = shared.by_name("ushr_imm");
let usub_sat = shared.by_name("usub_sat");
let vconst = shared.by_name("vconst");
@@ -1915,6 +1917,16 @@ fn define_simd(
let unarrow = unarrow.bind(vector(*ty, sse_vector_size));
e.enc_both_inferred_maybe_isap(unarrow, rec_fa.opcodes(*opcodes), *isap);
}
for (ty, swiden_opcode, uwiden_opcode) in &[
(I8, &PMOVSXBW[..], &PMOVZXBW[..]),
(I16, &PMOVSXWD[..], &PMOVZXWD[..]),
] {
let isap = Some(use_sse41_simd);
let swiden_low = swiden_low.bind(vector(*ty, sse_vector_size));
e.enc_both_inferred_maybe_isap(swiden_low, rec_furm.opcodes(*swiden_opcode), isap);
let uwiden_low = uwiden_low.bind(vector(*ty, sse_vector_size));
e.enc_both_inferred_maybe_isap(uwiden_low, rec_furm.opcodes(*uwiden_opcode), isap);
}
for ty in &[I8, I16, I32, I64] {
e.enc_both_inferred_maybe_isap(
x86_palignr.bind(vector(*ty, sse_vector_size)),

View File

@@ -477,7 +477,7 @@ pub static PMOVSXBW: [u8; 4] = [0x66, 0x0f, 0x38, 0x20];
pub static PMOVSXWD: [u8; 4] = [0x66, 0x0f, 0x38, 0x23];
/// Sign extend 2 packed 32-bit integers in the low 8 bytes of xmm2/m64 to 2 packed 64-bit
/// integers in xmm1.
/// integers in xmm1 (SSE4.1).
pub static PMOVSXDQ: [u8; 4] = [0x66, 0x0f, 0x38, 0x25];
/// Zero extend 8 packed 8-bit integers in the low 8 bytes of xmm2/m64 to 8 packed 16-bit
@@ -489,7 +489,7 @@ pub static PMOVZXBW: [u8; 4] = [0x66, 0x0f, 0x38, 0x30];
pub static PMOVZXWD: [u8; 4] = [0x66, 0x0f, 0x38, 0x33];
/// Zero extend 2 packed 32-bit integers in the low 8 bytes of xmm2/m64 to 2 packed 64-bit
/// integers in xmm1.
/// integers in xmm1 (SSE4.1).
pub static PMOVZXDQ: [u8; 4] = [0x66, 0x0f, 0x38, 0x35];
/// Multiply the packed signed word integers in xmm1 and xmm2/m128, and store the low 16 bits of