Implement basic RGB op fusing in skcms.
Change-Id: I8f5ce4cba60bfd644c61e46bca2303b0a846c1c3
Reviewed-on: https://skia-review.googlesource.com/c/skcms/+/771528
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
diff --git a/skcms.cc b/skcms.cc
index 7786e41..0e2770d 100644
--- a/skcms.cc
+++ b/skcms.cc
@@ -2350,26 +2350,31 @@
M(gamma_g) \
M(gamma_b) \
M(gamma_a) \
+ M(gamma_rgb) \
\
M(tf_r) \
M(tf_g) \
M(tf_b) \
M(tf_a) \
+ M(tf_rgb) \
\
M(pq_r) \
M(pq_g) \
M(pq_b) \
M(pq_a) \
+ M(pq_rgb) \
\
M(hlg_r) \
M(hlg_g) \
M(hlg_b) \
M(hlg_a) \
+ M(hlg_rgb) \
\
M(hlginv_r) \
M(hlginv_g) \
M(hlginv_b) \
M(hlginv_a) \
+ M(hlginv_rgb) \
\
M(table_r) \
M(table_g) \
@@ -2616,7 +2621,35 @@
if (ops[position].arg) {
++position;
}
+
+ // Identify separate R/G/B functions which can be fused into a single op.
+ // (We do this check inside the loop in order to allow R+G+B+A to be fused into RGB+A.)
+ if (index == 2 && position == 3) {
+ struct FusableOps {
+ Op r, g, b, rgb;
+ };
+ static constexpr FusableOps kFusableOps[] = {
+ {Op_gamma_r, Op_gamma_g, Op_gamma_b, Op_gamma_rgb},
+ {Op_tf_r, Op_tf_g, Op_tf_b, Op_tf_rgb},
+ {Op_pq_r, Op_pq_g, Op_pq_b, Op_pq_rgb},
+ {Op_hlg_r, Op_hlg_g, Op_hlg_b, Op_hlg_rgb},
+ {Op_hlginv_r, Op_hlginv_g, Op_hlginv_b, Op_hlginv_rgb},
+ };
+ for (const FusableOps& fusableOp : kFusableOps) {
+ if (ops[0].op == fusableOp.r &&
+ ops[1].op == fusableOp.g &&
+ ops[2].op == fusableOp.b &&
+ (0 == memcmp(ops[0].arg, ops[1].arg, sizeof(skcms_TransferFunction))) &&
+ (0 == memcmp(ops[0].arg, ops[2].arg, sizeof(skcms_TransferFunction)))) {
+
+ ops[0].op = fusableOp.rgb;
+ position = 1;
+ break;
+ }
+ }
+ }
}
+
return position;
}
@@ -2712,9 +2745,14 @@
*contexts++ = c;
};
- auto add_op_list = [&](const OpAndArg* oa, int count) {
- for (const OpAndArg* end = oa + count; oa != end; ++oa) {
- add_op_ctx(oa->op, oa->arg);
+ auto add_curve_ops = [&](const skcms_Curve* curves, int numChannels) {
+ OpAndArg oa[4];
+ assert(numChannels <= ARRAY_COUNT(oa));
+
+ int numOps = select_curve_ops(curves, numChannels, oa);
+
+ for (int i = 0; i < numOps; ++i) {
+ add_op_ctx(oa[i].op, oa[i].arg);
}
};
@@ -2749,9 +2787,7 @@
case skcms_PixelFormat_RGBA_8888_sRGB >> 1:
add_op(Op_load_8888);
- add_op_ctx(Op_tf_r, skcms_sRGB_TransferFunction());
- add_op_ctx(Op_tf_g, skcms_sRGB_TransferFunction());
- add_op_ctx(Op_tf_b, skcms_sRGB_TransferFunction());
+ add_op_ctx(Op_tf_rgb, skcms_sRGB_TransferFunction());
break;
}
if (srcFmt == skcms_PixelFormat_RGB_hhh_Norm ||
@@ -2796,21 +2832,14 @@
if (srcProfile->has_A2B) {
if (srcProfile->A2B.input_channels) {
- OpAndArg oa[4];
- assert(srcProfile->A2B.input_channels <= ARRAY_COUNT(oa));
- int numOps = select_curve_ops(srcProfile->A2B.input_curves,
- (int)srcProfile->A2B.input_channels,
- oa);
- add_op_list(oa, numOps);
-
+ add_curve_ops(srcProfile->A2B.input_curves,
+ (int)srcProfile->A2B.input_channels);
add_op(Op_clamp);
add_op_ctx(Op_clut_A2B, &srcProfile->A2B);
}
if (srcProfile->A2B.matrix_channels == 3) {
- OpAndArg oa[3];
- int numOps = select_curve_ops(srcProfile->A2B.matrix_curves, /*numChannels=*/3, oa);
- add_op_list(oa, numOps);
+ add_curve_ops(srcProfile->A2B.matrix_curves, /*numChannels=*/3);
static const skcms_Matrix3x4 I = {{
{1,0,0,0},
@@ -2823,9 +2852,7 @@
}
if (srcProfile->A2B.output_channels == 3) {
- OpAndArg oa[3];
- int numOps = select_curve_ops(srcProfile->A2B.output_curves, /*numChannels=*/3, oa);
- add_op_list(oa, numOps);
+ add_curve_ops(srcProfile->A2B.output_curves, /*numChannels=*/3);
}
if (srcProfile->pcs == skcms_Signature_Lab) {
@@ -2833,9 +2860,7 @@
}
} else if (srcProfile->has_trc && srcProfile->has_toXYZD50) {
- OpAndArg oa[3];
- int numOps = select_curve_ops(srcProfile->trc, /*numChannels=*/3, oa);
- add_op_list(oa, numOps);
+ add_curve_ops(srcProfile->trc, /*numChannels=*/3);
} else {
return false;
}
@@ -2854,9 +2879,7 @@
}
if (dstProfile->B2A.input_channels == 3) {
- OpAndArg oa[3];
- int numOps = select_curve_ops(dstProfile->B2A.input_curves, /*numChannels=*/3, oa);
- add_op_list(oa, numOps);
+ add_curve_ops(dstProfile->B2A.input_curves, /*numChannels=*/3);
}
if (dstProfile->B2A.matrix_channels == 3) {
@@ -2869,21 +2892,15 @@
add_op_ctx(Op_matrix_3x4, &dstProfile->B2A.matrix);
}
- OpAndArg oa[3];
- int numOps = select_curve_ops(dstProfile->B2A.matrix_curves, /*numChannels=*/3, oa);
- add_op_list(oa, numOps);
+ add_curve_ops(dstProfile->B2A.matrix_curves, /*numChannels=*/3);
}
if (dstProfile->B2A.output_channels) {
add_op(Op_clamp);
add_op_ctx(Op_clut_B2A, &dstProfile->B2A);
- OpAndArg oa[4];
- assert(dstProfile->B2A.output_channels <= ARRAY_COUNT(oa));
- int numOps = select_curve_ops(dstProfile->B2A.output_curves,
- (int)dstProfile->B2A.output_channels,
- oa);
- add_op_list(oa, numOps);
+ add_curve_ops(dstProfile->B2A.output_curves,
+ (int)dstProfile->B2A.output_channels);
}
} else {
// This is a TRC destination.
@@ -2913,8 +2930,8 @@
oa[index].op != Op_table_g &&
oa[index].op != Op_table_b &&
oa[index].op != Op_table_a);
+ add_op_ctx(oa[index].op, oa[index].arg);
}
- add_op_list(oa, numOps);
}
}
@@ -2966,9 +2983,7 @@
case skcms_PixelFormat_RGBA_ffff >> 1: add_op(Op_store_ffff); break;
case skcms_PixelFormat_RGBA_8888_sRGB >> 1:
- add_op_ctx(Op_tf_r, skcms_sRGB_Inverse_TransferFunction());
- add_op_ctx(Op_tf_g, skcms_sRGB_Inverse_TransferFunction());
- add_op_ctx(Op_tf_b, skcms_sRGB_Inverse_TransferFunction());
+ add_op_ctx(Op_tf_rgb, skcms_sRGB_Inverse_TransferFunction());
add_op(Op_store_8888);
break;
}
diff --git a/src/Transform_inl.h b/src/Transform_inl.h
index a2b19f1..6f64b08 100644
--- a/src/Transform_inl.h
+++ b/src/Transform_inl.h
@@ -1131,26 +1131,56 @@
STAGE(gamma_b, const skcms_TransferFunction* tf) { b = apply_gamma(tf, b); }
STAGE(gamma_a, const skcms_TransferFunction* tf) { a = apply_gamma(tf, a); }
+STAGE(gamma_rgb, const skcms_TransferFunction* tf) {
+ r = apply_gamma(tf, r);
+ g = apply_gamma(tf, g);
+ b = apply_gamma(tf, b);
+}
+
STAGE(tf_r, const skcms_TransferFunction* tf) { r = apply_tf(tf, r); }
STAGE(tf_g, const skcms_TransferFunction* tf) { g = apply_tf(tf, g); }
STAGE(tf_b, const skcms_TransferFunction* tf) { b = apply_tf(tf, b); }
STAGE(tf_a, const skcms_TransferFunction* tf) { a = apply_tf(tf, a); }
+STAGE(tf_rgb, const skcms_TransferFunction* tf) {
+ r = apply_tf(tf, r);
+ g = apply_tf(tf, g);
+ b = apply_tf(tf, b);
+}
+
STAGE(pq_r, const skcms_TransferFunction* tf) { r = apply_pq(tf, r); }
STAGE(pq_g, const skcms_TransferFunction* tf) { g = apply_pq(tf, g); }
STAGE(pq_b, const skcms_TransferFunction* tf) { b = apply_pq(tf, b); }
STAGE(pq_a, const skcms_TransferFunction* tf) { a = apply_pq(tf, a); }
+STAGE(pq_rgb, const skcms_TransferFunction* tf) {
+ r = apply_pq(tf, r);
+ g = apply_pq(tf, g);
+ b = apply_pq(tf, b);
+}
+
STAGE(hlg_r, const skcms_TransferFunction* tf) { r = apply_hlg(tf, r); }
STAGE(hlg_g, const skcms_TransferFunction* tf) { g = apply_hlg(tf, g); }
STAGE(hlg_b, const skcms_TransferFunction* tf) { b = apply_hlg(tf, b); }
STAGE(hlg_a, const skcms_TransferFunction* tf) { a = apply_hlg(tf, a); }
+STAGE(hlg_rgb, const skcms_TransferFunction* tf) {
+ r = apply_hlg(tf, r);
+ g = apply_hlg(tf, g);
+ b = apply_hlg(tf, b);
+}
+
STAGE(hlginv_r, const skcms_TransferFunction* tf) { r = apply_hlginv(tf, r); }
STAGE(hlginv_g, const skcms_TransferFunction* tf) { g = apply_hlginv(tf, g); }
STAGE(hlginv_b, const skcms_TransferFunction* tf) { b = apply_hlginv(tf, b); }
STAGE(hlginv_a, const skcms_TransferFunction* tf) { a = apply_hlginv(tf, a); }
+STAGE(hlginv_rgb, const skcms_TransferFunction* tf) {
+ r = apply_hlginv(tf, r);
+ g = apply_hlginv(tf, g);
+ b = apply_hlginv(tf, b);
+}
+
STAGE(table_r, const skcms_Curve* curve) { r = table(curve, r); }
STAGE(table_g, const skcms_Curve* curve) { g = table(curve, g); }
STAGE(table_b, const skcms_Curve* curve) { b = table(curve, b); }
