diff options
Diffstat (limited to '')
| -rw-r--r-- | gfx/qcms/src/transform_sse2.rs | 159 |
1 files changed, 159 insertions, 0 deletions
diff --git a/gfx/qcms/src/transform_sse2.rs b/gfx/qcms/src/transform_sse2.rs new file mode 100644 index 0000000000..f6bccaadc3 --- /dev/null +++ b/gfx/qcms/src/transform_sse2.rs @@ -0,0 +1,159 @@ +use crate::transform::{qcms_transform, Format, BGRA, CLAMPMAXVAL, FLOATSCALE, RGB, RGBA}; +#[cfg(target_arch = "x86")] +pub use std::arch::x86::{ + __m128, __m128i, _mm_add_ps, _mm_cvtps_epi32, _mm_load_ps, _mm_load_ss, _mm_max_ps, _mm_min_ps, + _mm_mul_ps, _mm_set1_ps, _mm_setzero_ps, _mm_shuffle_ps, _mm_store_si128, +}; +#[cfg(target_arch = "x86_64")] +pub use std::arch::x86_64::{ + __m128, __m128i, _mm_add_ps, _mm_cvtps_epi32, _mm_load_ps, _mm_load_ss, _mm_max_ps, _mm_min_ps, + _mm_mul_ps, _mm_set1_ps, _mm_setzero_ps, _mm_shuffle_ps, _mm_store_si128, +}; + +#[repr(align(16))] +struct Output([u32; 4]); + +unsafe extern "C" fn qcms_transform_data_template_lut_sse2<F: Format>( + transform: &qcms_transform, + mut src: *const u8, + mut dest: *mut u8, + mut length: usize, +) { + let mat: *const [f32; 4] = (*transform).matrix.as_ptr(); + let mut input: Output = std::mem::zeroed(); + /* share input and output locations to save having to keep the + * locations in separate registers */ + let output: *const u32 = &mut input as *mut Output as *mut u32; + /* deref *transform now to avoid it in loop */ + let igtbl_r: *const f32 = (*transform).input_gamma_table_r.as_ref().unwrap().as_ptr(); + let igtbl_g: *const f32 = (*transform).input_gamma_table_g.as_ref().unwrap().as_ptr(); + let igtbl_b: *const f32 = (*transform).input_gamma_table_b.as_ref().unwrap().as_ptr(); + /* deref *transform now to avoid it in loop */ + let otdata_r: *const u8 = (*transform) + .output_table_r + .as_deref() + .unwrap() + .data + .as_ptr(); + let otdata_g: *const u8 = (*transform) + .output_table_g + .as_deref() + .unwrap() + .data + .as_ptr(); + let otdata_b: *const u8 = (*transform) + .output_table_b + .as_deref() + .unwrap() + .data + .as_ptr(); + /* input matrix values never change */ + let mat0: __m128 = _mm_load_ps((*mat.offset(0isize)).as_ptr()); + let mat1: __m128 = _mm_load_ps((*mat.offset(1isize)).as_ptr()); + let mat2: __m128 = _mm_load_ps((*mat.offset(2isize)).as_ptr()); + /* these values don't change, either */ + let max: __m128 = _mm_set1_ps(CLAMPMAXVAL); + let min: __m128 = _mm_setzero_ps(); + let scale: __m128 = _mm_set1_ps(FLOATSCALE); + let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32; + /* working variables */ + let mut vec_r: __m128; + let mut vec_g: __m128; + let mut vec_b: __m128; + let mut result: __m128; + let mut alpha: u8 = 0; + /* CYA */ + if length == 0 { + return; + } + /* one pixel is handled outside of the loop */ + length -= 1; + /* setup for transforming 1st pixel */ + vec_r = _mm_load_ss(&*igtbl_r.offset(*src.add(F::kRIndex) as isize)); + vec_g = _mm_load_ss(&*igtbl_g.offset(*src.add(F::kGIndex) as isize)); + vec_b = _mm_load_ss(&*igtbl_b.offset(*src.add(F::kBIndex) as isize)); + if F::kAIndex != 0xff { + alpha = *src.add(F::kAIndex) + } + src = src.offset(components as isize); + let mut i: u32 = 0; + while (i as usize) < length { + /* position values from gamma tables */ + vec_r = _mm_shuffle_ps(vec_r, vec_r, 0); + vec_g = _mm_shuffle_ps(vec_g, vec_g, 0); + vec_b = _mm_shuffle_ps(vec_b, vec_b, 0); + /* gamma * matrix */ + vec_r = _mm_mul_ps(vec_r, mat0); + vec_g = _mm_mul_ps(vec_g, mat1); + vec_b = _mm_mul_ps(vec_b, mat2); + /* store alpha for this pixel; load alpha for next */ + if F::kAIndex != 0xff { + *dest.add(F::kAIndex) = alpha; + alpha = *src.add(F::kAIndex) + } + /* crunch, crunch, crunch */ + vec_r = _mm_add_ps(vec_r, _mm_add_ps(vec_g, vec_b)); + vec_r = _mm_max_ps(min, vec_r); + vec_r = _mm_min_ps(max, vec_r); + result = _mm_mul_ps(vec_r, scale); + /* store calc'd output tables indices */ + _mm_store_si128(output as *mut __m128i, _mm_cvtps_epi32(result)); + /* load gamma values for next loop while store completes */ + vec_r = _mm_load_ss(&*igtbl_r.offset(*src.add(F::kRIndex) as isize)); + vec_g = _mm_load_ss(&*igtbl_g.offset(*src.add(F::kGIndex) as isize)); + vec_b = _mm_load_ss(&*igtbl_b.offset(*src.add(F::kBIndex) as isize)); + src = src.offset(components as isize); + /* use calc'd indices to output RGB values */ + *dest.add(F::kRIndex) = *otdata_r.offset(*output.offset(0isize) as isize); + *dest.add(F::kGIndex) = *otdata_g.offset(*output.offset(1isize) as isize); + *dest.add(F::kBIndex) = *otdata_b.offset(*output.offset(2isize) as isize); + dest = dest.offset(components as isize); + i += 1 + } + /* handle final (maybe only) pixel */ + vec_r = _mm_shuffle_ps(vec_r, vec_r, 0); + vec_g = _mm_shuffle_ps(vec_g, vec_g, 0); + vec_b = _mm_shuffle_ps(vec_b, vec_b, 0); + vec_r = _mm_mul_ps(vec_r, mat0); + vec_g = _mm_mul_ps(vec_g, mat1); + vec_b = _mm_mul_ps(vec_b, mat2); + if F::kAIndex != 0xff { + *dest.add(F::kAIndex) = alpha + } + vec_r = _mm_add_ps(vec_r, _mm_add_ps(vec_g, vec_b)); + vec_r = _mm_max_ps(min, vec_r); + vec_r = _mm_min_ps(max, vec_r); + result = _mm_mul_ps(vec_r, scale); + _mm_store_si128(output as *mut __m128i, _mm_cvtps_epi32(result)); + *dest.add(F::kRIndex) = *otdata_r.offset(*output.offset(0isize) as isize); + *dest.add(F::kGIndex) = *otdata_g.offset(*output.offset(1isize) as isize); + *dest.add(F::kBIndex) = *otdata_b.offset(*output.offset(2isize) as isize); +} +#[no_mangle] +pub unsafe fn qcms_transform_data_rgb_out_lut_sse2( + transform: &qcms_transform, + src: *const u8, + dest: *mut u8, + length: usize, +) { + qcms_transform_data_template_lut_sse2::<RGB>(transform, src, dest, length); +} +#[no_mangle] +pub unsafe fn qcms_transform_data_rgba_out_lut_sse2( + transform: &qcms_transform, + src: *const u8, + dest: *mut u8, + length: usize, +) { + qcms_transform_data_template_lut_sse2::<RGBA>(transform, src, dest, length); +} + +#[no_mangle] +pub unsafe fn qcms_transform_data_bgra_out_lut_sse2( + transform: &qcms_transform, + src: *const u8, + dest: *mut u8, + length: usize, +) { + qcms_transform_data_template_lut_sse2::<BGRA>(transform, src, dest, length); +} |