OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

broadcom/compiler: produce better code for f2f16 with RTZ rounding

Browse Source 
Suggested by Georg Lehmann, this generates far less code and should
be more correct.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8090
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20721>
This commit is contained in:
Iago Toral Quiroga
2023-01-16 09:04:10 +01:00
parent a48642400b
commit 9bf525b4bd
1 changed files with 16 additions and 78 deletions
Download Patch File Download Diff File Expand all files Collapse all files

94
src/broadcom/compiler/nir_to_vir.c
View File

@@ -1299,88 +1299,26 @@ ntq_emit_cond_to_int(struct v3d_compile *c, enum v3d_qpu_cond cond)
static struct qreg
f2f16_rtz(struct v3d_compile *c, struct qreg f32)
{
/* The GPU doesn't provide a mechanism to modify the f32->f16 rounding
* method and seems to be using RTE by default, so we need to implement
* RTZ rounding in software :-(
*
* The implementation identifies the cases where RTZ applies and
* returns the correct result and for everything else, it just uses
* the default RTE conversion.
*/
static bool _first = true;
if (_first && V3D_DBG(PERF)) {
fprintf(stderr, "Shader uses round-toward-zero f32->f16 "
"conversion which is not supported in hardware.\n");
_first = false;
}
/* The GPU doesn't provide a mechanism to modify the f32->f16 rounding
* method and seems to be using RTE by default, so we need to implement
* RTZ rounding in software.
*/
struct qreg rf16 = vir_FMOV(c, f32);
vir_set_pack(c->defs[rf16.index], V3D_QPU_PACK_L);
struct qinst *inst;
struct qreg tmp;
struct qreg rf32 = vir_FMOV(c, rf16);
vir_set_unpack(c->defs[rf32.index], 0, V3D_QPU_UNPACK_L);
struct qreg result = vir_get_temp(c);
struct qreg f32_abs = vir_FMOV(c, f32);
vir_set_unpack(c->defs[f32_abs.index], 0, V3D_QPU_UNPACK_ABS);
struct qreg mantissa32 = vir_AND(c, f32, vir_uniform_ui(c, 0x007fffff));
struct qreg rf32_abs = vir_FMOV(c, rf32);
vir_set_unpack(c->defs[rf32_abs.index], 0, V3D_QPU_UNPACK_ABS);
/* Compute sign bit of result */
struct qreg sign = vir_AND(c, vir_SHR(c, f32, vir_uniform_ui(c, 16)),
vir_uniform_ui(c, 0x8000));
/* Check the cases were RTZ rounding is relevant based on exponent */
struct qreg exp32 = vir_AND(c, vir_SHR(c, f32, vir_uniform_ui(c, 23)),
vir_uniform_ui(c, 0xff));
struct qreg exp16 = vir_ADD(c, exp32, vir_uniform_ui(c, -127 + 15));
/* if (exp16 > 30) */
inst = vir_MIN_dest(c, vir_nop_reg(), exp16, vir_uniform_ui(c, 30));
vir_set_pf(c, inst, V3D_QPU_PF_PUSHC);
inst = vir_OR_dest(c, result, sign, vir_uniform_ui(c, 0x7bff));
vir_set_cond(inst, V3D_QPU_COND_IFA);
/* if (exp16 <= 30) */
inst = vir_OR_dest(c, result,
vir_OR(c, sign,
vir_SHL(c, exp16, vir_uniform_ui(c, 10))),
vir_SHR(c, mantissa32, vir_uniform_ui(c, 13)));
vir_set_cond(inst, V3D_QPU_COND_IFNA);
/* if (exp16 <= 0) */
inst = vir_MIN_dest(c, vir_nop_reg(), exp16, vir_uniform_ui(c, 0));
vir_set_pf(c, inst, V3D_QPU_PF_PUSHC);
tmp = vir_OR(c, mantissa32, vir_uniform_ui(c, 0x800000));
tmp = vir_SHR(c, tmp, vir_SUB(c, vir_uniform_ui(c, 14), exp16));
inst = vir_OR_dest(c, result, sign, tmp);
vir_set_cond(inst, V3D_QPU_COND_IFNA);
/* Cases where RTZ mode is not relevant: use default RTE conversion.
*
* The cases that are not affected by RTZ are:
*
* exp16 < - 10 || exp32 == 0 || exp32 == 0xff
*
* In V3D we can implement this condition as:
*
* !((exp16 >= -10) && !(exp32 == 0) && !(exp32 == 0xff)))
*/
/* exp16 >= -10 */
inst = vir_MIN_dest(c, vir_nop_reg(), exp16, vir_uniform_ui(c, -10));
vir_set_pf(c, inst, V3D_QPU_PF_PUSHC);
/* && !(exp32 == 0) */
inst = vir_MOV_dest(c, vir_nop_reg(), exp32);
vir_set_uf(c, inst, V3D_QPU_UF_ANDNZ);
/* && !(exp32 == 0xff) */
inst = vir_XOR_dest(c, vir_nop_reg(), exp32, vir_uniform_ui(c, 0xff));
vir_set_uf(c, inst, V3D_QPU_UF_ANDNZ);
/* Use regular RTE conversion if condition is False */
inst = vir_FMOV_dest(c, result, f32);
vir_set_pack(inst, V3D_QPU_PACK_L);
vir_set_cond(inst, V3D_QPU_COND_IFNA);
return vir_MOV(c, result);
vir_set_pf(c, vir_FCMP_dest(c, vir_nop_reg(), f32_abs, rf32_abs),
V3D_QPU_PF_PUSHN);
return vir_MOV(c, vir_SEL(c, V3D_QPU_COND_IFA,
vir_SUB(c, rf16, vir_uniform_ui(c, 1)), rf16));
}
/**