KonstantinSeurer
/
mesa
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

compiler/spirv: implement 16-bit atan 

v2:
 - use nir_fadd_imm and nir_fmul_imm helpers (Jason)
 - rebased on top of new sized boolean opcodes
 - use nir_b2f helper

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
This commit is contained in:
Iago Toral Quiroga 2018-04-18 09:36:41 +02:00
parent df118535ca
commit 1c8de08ec9
1 changed files with 11 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
src/compiler/spirv/vtn_glsl450.c
View File

@ -237,8 +237,10 @@ build_fsum(nir_builder *b, nir_ssa_def **xs, int terms)
static nir_ssa_def *
build_atan(nir_builder *b, nir_ssa_def *y_over_x)
{
const uint32_t bit_size = y_over_x->bit_size;
nir_ssa_def *abs_y_over_x = nir_fabs(b, y_over_x);
nir_ssa_def *one = nir_imm_float(b, 1.0f);
nir_ssa_def *one = nir_imm_floatN_t(b, 1.0f, bit_size);
/*
* range-reduction, first step:
@ -265,12 +267,12 @@ build_atan(nir_builder *b, nir_ssa_def *y_over_x)
nir_ssa_def *x_11 = nir_fmul(b, x_9, x_2);
nir_ssa_def *polynomial_terms[] = {
nir_fmul(b, x, nir_imm_float(b, 0.9999793128310355f)),
nir_fmul(b, x_3, nir_imm_float(b, -0.3326756418091246f)),
nir_fmul(b, x_5, nir_imm_float(b, 0.1938924977115610f)),
nir_fmul(b, x_7, nir_imm_float(b, -0.1173503194786851f)),
nir_fmul(b, x_9, nir_imm_float(b, 0.0536813784310406f)),
nir_fmul(b, x_11, nir_imm_float(b, -0.0121323213173444f)),
nir_fmul_imm(b, x, 0.9999793128310355f),
nir_fmul_imm(b, x_3, -0.3326756418091246f),
nir_fmul_imm(b, x_5, 0.1938924977115610f),
nir_fmul_imm(b, x_7, -0.1173503194786851f),
nir_fmul_imm(b, x_9, 0.0536813784310406f),
nir_fmul_imm(b, x_11, -0.0121323213173444f),
};
nir_ssa_def *tmp =
@ -278,11 +280,8 @@ build_atan(nir_builder *b, nir_ssa_def *y_over_x)
/* range-reduction fixup */
tmp = nir_fadd(b, tmp,
nir_fmul(b,
nir_b2f32(b, nir_flt(b, one, abs_y_over_x)),
nir_fadd(b, nir_fmul(b, tmp,
nir_imm_float(b, -2.0f)),
nir_imm_float(b, M_PI_2f))));
nir_fmul(b, nir_b2f(b, nir_flt(b, one, abs_y_over_x), bit_size),
nir_fadd_imm(b, nir_fmul_imm(b, tmp, -2.0f), M_PI_2f)));
/* sign fixup */
return nir_fmul(b, tmp, nir_fsign(b, y_over_x));