HWR_SkyDomeVertex: tiny optimization

save a tiny bit of float division, use float sin and cos functions, handle implicit float conversions
This commit is contained in:
Alug committed 2025-11-13 16:45:57 +01:00
1 parent aa985823ba
commit dfca24b370
1 file changed
+20 -16
+20 -16
View File
@@ -5067,42 +5067,46 @@ static void HWR_ProjectPrecipitationSprite(precipmobj_t *thing)
static gl_sky_t gl_sky;
#define DEG2RADGL(a) ((a * M_PIl) / 180.0f)
static void HWR_SkyDomeVertex(gl_sky_t *sky, gl_skyvertex_t *vbo, int r, int c, signed char yflip, float delta, boolean foglayer)
{
static constexpr float radians = (float)(M_PIl / 180.0f);
static constexpr float scale = 10000.0f;
static constexpr float maxSideAngle = DEG2RADGL(60.0f);
static constexpr float maxSideAngle = 60.0f * radians;
float topAngle = DEG2RADGL(c / (float)sky->columns * 360.0f);
float topAngle = ((float)c / (float)sky->columns * 360.0f) * radians;
float sideAngle = (maxSideAngle * (float)(sky->rows - r) / (float)sky->rows);
float height = (float)(sin(sideAngle));
float realRadius = (scale * (float)cos(sideAngle));
float x = (realRadius * (float)cos(topAngle));
float height = sinf(sideAngle);
float realRadius = (scale * cosf(sideAngle));
float x = (realRadius * cosf(topAngle));
float y = (!yflip) ? scale * height : -scale * height;
float z = (realRadius * (float)sin(topAngle));
float timesRepeat = (4 * (256.0f / sky->width));
if (std::fpclassify(timesRepeat) == FP_ZERO)
timesRepeat = 1.0f;
float z = (realRadius * sinf(topAngle));
float timesRepeat;
if (!foglayer)
{
timesRepeat = (4.0f * (256.0f / (float)sky->width));
if (std::fpclassify(timesRepeat) == FP_ZERO)
timesRepeat = 1.0f;
vbo->r = 255;
vbo->g = 255;
vbo->b = 255;
vbo->a = (r == 0 ? 0 : 255);
// And the texture coordinates.
vbo->u = (-timesRepeat * c / (float)sky->columns);
if (!yflip) // Flipped Y is for the lower hemisphere.
vbo->v = (r / (float)sky->rows) + 0.5f;
vbo->u = (-timesRepeat * (float)c / (float)sky->columns);
if (!yflip) // Flipped Y is for the lower hemisphere.
vbo->v = ((float)r / (float)sky->rows) + 0.5f;
else
vbo->v = 1.0f + ((sky->rows - r) / (float)sky->rows) + 0.5f;
vbo->v = 1.0f + ((float)(sky->rows - r) / (float)sky->rows) + 0.5f;
}
if (r != 4)
{
y += 300.0f;
}
// And finally the vertex.
vbo->x = x;