diff options
| author | Rhys Perry <[email protected]> | 2020-06-15 14:30:34 +0100 |
|---|---|---|
| committer | Marge Bot <[email protected]> | 2020-06-15 18:24:22 +0000 |
| commit | 4784111abc113ae64ce1d597407e0c32d6a88160 (patch) | |
| tree | 8893a6492b9f32e60632e9685ae771a361f639fc /src/amd | |
| parent | dd233455679fe0f0ae441cc5ef3dd366132951e7 (diff) | |
aco: use 32-bit inline constants for 16-bit integer instructions
See https://reviews.llvm.org/D81841
Signed-off-by: Rhys Perry <[email protected]>
Reviewed-by: Daniel Schürmann <[email protected]>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5245>
Diffstat (limited to 'src/amd')
| -rw-r--r-- | src/amd/compiler/aco_opcodes.py | 20 |
1 files changed, 13 insertions, 7 deletions
diff --git a/src/amd/compiler/aco_opcodes.py b/src/amd/compiler/aco_opcodes.py index 1396b8c3af4..a0ecc9c5788 100644 --- a/src/amd/compiler/aco_opcodes.py +++ b/src/amd/compiler/aco_opcodes.py @@ -194,12 +194,21 @@ class Opcode(object): parts = name.replace('_e64', '').rsplit('_', 2) op_dtype = parts[-1] def_dtype = parts[-2] if len(parts) > 1 else parts[-1] - dtype_sizes = {'{}{}'.format(prefix, size) : size for prefix in 'biuf' for size in [64, 32, 24, 16]} - self.operand_size = dtype_sizes.get(op_dtype, 0) - self.definition_size = dtype_sizes.get(def_dtype, self.operand_size) + + def_dtype_sizes = {'{}{}'.format(prefix, size) : size for prefix in 'biuf' for size in [64, 32, 24, 16]} + op_dtype_sizes = {k:v for k, v in def_dtype_sizes.items()} + # inline constants are 32-bit for 16-bit integer/typeless instructions: https://reviews.llvm.org/D81841 + op_dtype_sizes['b16'] = 32 + op_dtype_sizes['i16'] = 32 + op_dtype_sizes['u16'] = 32 + + self.operand_size = op_dtype_sizes.get(op_dtype, 0) + self.definition_size = def_dtype_sizes.get(def_dtype, self.operand_size) # exceptions - if self.operand_size == 24: + if self.operand_size == 16 and op_dtype != 'f16': + self.operand_size = 16 + elif self.operand_size == 24: self.operand_size = 32 elif name in ['s_sext_i32_i8', 's_sext_i32_i16', 'v_msad_u8', 'v_cvt_pk_u16_u32', 'v_cvt_pk_i16_i32']: self.operand_size = 32 @@ -208,9 +217,6 @@ class Opcode(object): self.operand_size = 0 elif name in ['v_mad_u64_u32', 'v_mad_i64_i32']: self.operand_size = 0 - elif name.replace('_e64', '') in ['v_lshrrev_b16', 'v_ashrrev_i16', 'v_lshlrev_b16']: - # v_lshlrev_b16 tested on GFX10 with 1/2 PI inline constant - self.operand_size = 32 elif '_pk_' in name or name in ['v_lerp_u8', 'v_sad_u8', 'v_sad_u16', 'v_cvt_f32_ubyte0', 'v_cvt_f32_ubyte1', 'v_cvt_f32_ubyte2', 'v_cvt_f32_ubyte3']: |