diff options
| author | Samuel Iglesias Gonsálvez <[email protected]> | 2019-02-12 15:43:10 +0100 |
|---|---|---|
| committer | Andres Gomez <[email protected]> | 2019-09-17 23:39:18 +0300 |
| commit | 7580707345b7df0f262935c30b43bde16d297e39 (patch) | |
| tree | 9652120d350202eebe9253f0bc9c829d0b057da6 /src/intel | |
| parent | 0ac07c7ca7207f3f1388c0450b456ecc578d9c5b (diff) | |
nir: mind rounding mode on fadd, fsub, fmul and fma opcodes
According to Vulkan spec, the new execution modes affect only
correctly rounded SPIR-V instructions, which includes fadd, fsub and
fmul.
v2:
- Fix fmul, fsub and fadd round-to-zero definitions, they should use
auxiliary functions to calculate the proper value because Mesa uses
round-to-nearest-even rounding mode by default (Connor).
v3:
- Do an actual fused multiply-add at ffma (Connor).
v4:
- Simplify fadd and fmul for bit sizes < 64 (Connor).
- Do not use double ffma for 32 bits float (Connor).
Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Signed-off-by: Andres Gomez <[email protected]>
Reviewed-by: Connor Abbott <[email protected]> [v3]
Diffstat (limited to 'src/intel')
0 files changed, 0 insertions, 0 deletions