sum 46 apps/blur/test.cpp __m128i a, b, c, sum, avg; sum 56 apps/blur/test.cpp sum = _mm_add_epi16(_mm_add_epi16(a, b), c); sum 57 apps/blur/test.cpp avg = _mm_mulhi_epi16(sum, one_third); sum 69 apps/blur/test.cpp sum = _mm_add_epi16(_mm_add_epi16(a, b), c); sum 70 apps/blur/test.cpp avg = _mm_mulhi_epi16(sum, one_third); sum 94 apps/fft/complex.h return ComplexExpr(Halide::sum(re(z), s + "_re"), sum 95 apps/fft/complex.h Halide::sum(im(z), s + "_im")); sum 248 apps/fft/fft.cpp X(A({n}, args)) = sum(expj((sign*2*kPi*k*n)/N) * x(A({k}, args))); sum 17 apps/hexagon_benchmarks/conv3x3_generator.cpp Expr sum = cast(accumulator_type, 0); sum 20 apps/hexagon_benchmarks/conv3x3_generator.cpp sum += cast<int16_t>(bounded_input(x+j, y+i)) * cast<int16_t>(mask(j+1, i+1)); sum 23 apps/hexagon_benchmarks/conv3x3_generator.cpp output(x, y) = cast<uint8_t>(clamp(sum >> 4, 0, 255)); sum 113 apps/hexagon_benchmarks/process.h int16_t sum = 0; sum 116 apps/hexagon_benchmarks/process.h sum += static_cast<int16_t>(u8_in(clamp(x+rx, 0, W-1), clamp(y+ry, 0, H-1))) sum 120 apps/hexagon_benchmarks/process.h sum = sum >> 4; sum 121 apps/hexagon_benchmarks/process.h sum = clamp<int16_t>(sum, 0, 255); sum 123 apps/hexagon_benchmarks/process.h if (sum != out_xy) { sum 124 apps/hexagon_benchmarks/process.h printf("Conv3x3a16: Mismatch at %d %d : %d != %d\n", x, y, out_xy, sum); sum 474 apps/hexagon_benchmarks/process.h int32_t sum = 0; sum 477 apps/hexagon_benchmarks/process.h sum += static_cast<int16_t>(u8_in(clamp(x+rx, 0, W-1), clamp(y+ry, 0, H-1))) sum 481 apps/hexagon_benchmarks/process.h sum = sum >> 4; sum 482 apps/hexagon_benchmarks/process.h sum = clamp(sum, 0, 255); sum 484 apps/hexagon_benchmarks/process.h if (sum != out_xy) { sum 485 apps/hexagon_benchmarks/process.h printf("Conv3x3a32: Mismatch at %d %d : %d != %d\n", x, y, out_xy, sum); sum 89 apps/linear_algebra/benchmarks/eigen_benchmarks.cpp L1Benchmark(asum, type_name<T>(), result = x.array().abs().sum()); sum 138 apps/linear_algebra/src/blas_l1_generators.cpp result(0) = sum(dot(lanes)); sum 139 apps/linear_algebra/src/blas_l1_generators.cpp result(0) += sum(x_(tail) * y_(tail)); sum 146 apps/linear_algebra/src/blas_l1_generators.cpp result(0) = sum(x_(k) * y_(k)); sum 205 apps/linear_algebra/src/blas_l1_generators.cpp result(0) = sum(norm(lanes)); sum 206 apps/linear_algebra/src/blas_l1_generators.cpp result(0) += sum(abs(x_(tail))); sum 213 apps/linear_algebra/src/blas_l1_generators.cpp result(0) = sum(abs(x_(k))); sum 141 apps/resize/resize.cpp kernelx(x, k) = kx(x, k) / sum(kx(x, domx)); sum 142 apps/resize/resize.cpp kernely(y, k) = ky(y, k) / sum(ky(y, domy)); sum 148 apps/resize/resize.cpp resized_x(x, y, c) = sum(kernelx(x, domx) * cast<float>(clamped(domx + beginx, y, c))); sum 149 apps/resize/resize.cpp resized_y(x, y, c) = sum(kernely(y, domy) * resized_x(x, domy + beginy, c)); sum 16 python_bindings/python/InlineReductions.cpp return h::sum(e, name); sum 20 python_bindings/python/InlineReductions.cpp return h::sum(r, e, name); sum 105 src/InlineReductions.cpp return sum(RDom(), e, name); sum 36 src/InlineReductions.h EXPORT Expr sum(Expr, const std::string &s = "sum"); sum 52 src/InlineReductions.h EXPORT Expr sum(RDom, Expr, const std::string &s = "sum"); sum 20 test/correctness/bool_compute_root_vectorize.cpp int32_t result = evaluate_may_gpu<int32_t>(sum(selector(range.x, range.y))); sum 41 test/correctness/cascaded_filters.cpp float err = evaluate_may_gpu<float>(sum(abs(result(RDom(result)) - 1.0f))); sum 67 test/correctness/convolution.cpp blur2(x, y) = sum(tent(r.x, r.y) * input(x + r.x - 1, y + r.y - 1)); sum 39 test/correctness/convolution_multiple_kernels.cpp blur(x, y) = sum(box1(r.x, r.y) * input(x + r.x, y + r.y)) sum 40 test/correctness/convolution_multiple_kernels.cpp + sum(cast<uint16_t>(box2(r.x, r.y)) * input(x + r.x, y + r.y)); sum 11 test/correctness/explicit_inline_reductions.cpp f(x, y) = product(sum(r1, r1 + r3) + sum(r2, r2 * 2 + r3)); sum 12 test/correctness/explicit_inline_reductions.cpp f(r1, y) += product(r3, sum(r2, r1 + r2 + r3)); sum 109 test/correctness/extern_producer.cpp float error = evaluate_may_gpu<float>(sum(abs(output(r.x, r.y)))); sum 134 test/correctness/extern_producer.cpp float error_multi = evaluate<float>(sum(abs(output_multi(r.x, r.y)))); sum 46 test/correctness/extern_sort.cpp float error = evaluate_may_gpu<float>(sum(abs(reference(r) - output(r)))); sum 12 test/correctness/gpu_free_sync.cpp f(x, y) = sum(sqrt(sqrt(sqrt(sqrt(x+y+r))))); sum 16 test/correctness/gpu_vectorized_shared_memory.cpp g(x) += sum(r); sum 31 test/correctness/in_place.cpp float err = evaluate_may_gpu<float>(sum(abs(data(r) - reference_out(r)))); sum 18 test/correctness/inline_reduction.cpp Expr local_mean = sum(input_val) / 9.0f; sum 19 test/correctness/inline_reduction.cpp local_variance(x, y) = sum(input_val*input_val)/81.0f - local_mean*local_mean; sum 116 test/correctness/inline_reduction.cpp sum_implicit(_) = sum(input_3d(_, all_z)); sum 57 test/correctness/integer_powers.cpp rms_1() = sqrt(sum(pow(approx_result_1(r) - exact_result(r), 2), "rms_1_sum")); sum 58 test/correctness/integer_powers.cpp rms_2() = sqrt(sum(pow(approx_result_2(r) - exact_result(r), 2), "rms_2_sum")); sum 99 test/correctness/integer_powers.cpp rms_1() = sqrt(sum(pow(approx_result_1(r) - exact_result(r), 2), "rms_1_neg_sum")); sum 100 test/correctness/integer_powers.cpp rms_2() = sqrt(sum(pow(approx_result_2(r) - exact_result(r), 2), "rms_2_neg_sum")); sum 31 test/correctness/lazy_convolution.cpp blur(x, y) = select(f(x, y) > 0, sum(f(x+r.x, y+r.y)), 0); sum 205 test/correctness/likely.cpp g(c, x, y) = sum(select(nasty, likely(10), c + r)); sum 28 test/correctness/multi_way_select.cpp uint32_t err = evaluate_may_gpu<uint32_t>(sum(abs(g(r) - f(r)))); sum 19 test/correctness/parallel_reductions.cpp f1(i) = sum(i*16 + r1); sum 22 test/correctness/parallel_reductions.cpp f2() = sum(f1(r2)); sum 36 test/correctness/parallel_rvar.cpp int error = evaluate<int>(sum(f[0](r_check.x, r_check.y) - f[1](r_check.x, r_check.y))); sum 18 test/correctness/partition_loops_bug.cpp output(x,y) = sum(input_padded(x+rk.x,y+rk.y)); sum 28 test/correctness/random.cpp double mean = evaluate<double>(sum(val)) / (1024 * 1024); sum 29 test/correctness/random.cpp double variance = evaluate<double>(sum(pow(val - (float)mean, 2))) / (1024 * 1024 - 1); sum 36 test/correctness/random.cpp double mean_dx = evaluate<double>(sum(dx)) / (1024 * 1024); sum 37 test/correctness/random.cpp double variance_dx = evaluate<double>(sum(pow(dx - (float)mean_dx, 2))) / (1024 * 1024 - 1); sum 39 test/correctness/random.cpp double mean_dy = evaluate<double>(sum(dy)) / (1024 * 1024); sum 40 test/correctness/random.cpp double variance_dy = evaluate<double>(sum(pow(dy - (float)mean_dy, 2))) / (1024 * 1024 - 1); sum 99 test/correctness/random.cpp double e1 = evaluate<double>(sum(abs(v1 - v2))) / (1024 * 1024); sum 100 test/correctness/random.cpp double e2 = evaluate<double>(sum(abs(v1 - v3))) / (1024 * 1024); sum 125 test/correctness/random.cpp int set_bits = evaluate<int>(sum(popcount(val))); sum 136 test/correctness/random.cpp set_bits = evaluate<int>(sum(popcount(val2))); sum 163 test/correctness/random.cpp double f_var = evaluate<double>(sum(f_val * f_val)) / (S * S - 1); sum 164 test/correctness/random.cpp double g_var = evaluate<double>(sum(g_val * g_val)) / (S * S - 1); sum 22 test/correctness/realize_larger_than_two_gigs.cpp grand_total() = cast<uint8_t>(sum(big(r.x, r.y, r.z, r.w))); sum 33 test/correctness/reorder_rvars.cpp Expr check = sum(abs(f(r3.x, r3.y) - g(r3.x, r3.y))); sum 45 test/correctness/specialize_to_gpu.cpp uint32_t err = evaluate<uint32_t>(sum(abs(out(r) - reference(r)))); sum 8 test/correctness/vectorized_reduction_bug.cpp Func sum("sum"), foo("foo"); sum 14 test/correctness/vectorized_reduction_bug.cpp sum(x, y) += r.x; sum 16 test/correctness/vectorized_reduction_bug.cpp foo(x, y, c) = select(c == 3, 255, sum(x, y)); sum 43 test/correctness/vectorized_reduction_bug.cpp g(x, 0) = sum(h(x, RDom(0, 120))); sum 37 test/correctness/widening_reduction.cpp f(x, y) = u8_sat(sum(i16(input(x + r.x, y + r.y)) * kernel(r.x, r.y)) / 16); sum 11 test/error/ambiguous_inline_reductions.cpp f(x, y) = product(sum(r1, r1 + r3) + sum(r2, r2 * 2 + r3)); sum 17 test/error/ambiguous_inline_reductions.cpp f(r1, y) += product(sum(r2, r1 + r2 + r3)); sum 22 test/error/realize_constantly_larger_than_two_gigs.cpp grand_total() = cast<uint8_t>(sum(big(r.x, r.y, r.z))); sum 23 test/generator/embed_image_generator.cpp f(x, y, c) = sum(matrix(j, c) * input(x, y, j)); sum 17 test/generator/msan_generator.cpp msan_output(x, y, c) = sum(msan_extern_stage(r, y, c)); sum 27 test/opengl/conv_select.cpp f(x, y, c) = cast<float>(sum(input(coordx, y, c))); sum 15 test/opengl/inline_reduction.cpp f(x, y, c) = sum(cast<float>(r)); sum 24 test/opengl/sum_reduction.cpp g(x, y, c) = cast<float>(sum(input(coordx, y, c)) / sum(r) * 255.0f); sum 24 test/opengl/sumcolor_reduction.cpp g(x, y, c) = sum(input(x, y, r)); sum 49 test/performance/boundary_conditions.cpp g(x, y) = sum(f(x + r.x, y + r.y)); sum 29 test/performance/fast_pow.cpp f(x, y) = sum(pow_ref((x+1)/512.0f, (y+1+s)/512.0f)); sum 30 test/performance/fast_pow.cpp g(x, y) = sum(pow((x+1)/512.0f, (y+1+s)/512.0f)); sum 31 test/performance/fast_pow.cpp h(x, y) = sum(fast_pow((x+1)/512.0f, (y+1+s)/512.0f)); sum 259 tools/halide_image_info.h double sum = 0; sum 282 tools/halide_image_info.h sum += val; sum 300 tools/halide_image_info.h double avg = sum / cnt; sum 714 tutorial/lesson_09_update_definitions.cpp f1(x) = sum(r + x) * 7;