2690fps

1 files changed, 29 insertions, 29 deletions

diff --git a/render.c b/render.c
index 26b3c19..6e399ef 100644
--- a/render.c
+++ b/render.c
@@ -16,32 +16,32 @@ static void set_decay(uint16_t old_weight) {
 }
 
 /* [=]===^=[ apply_phosphor_decay ]=================================================================^===[=] */
-__attribute__((always_inline, hot))
-static inline void apply_phosphor_decay(void) {
-	// PROFILE_FUNCTION();
-	__m256i old_weight = _mm256_set1_epi16(_old_weight);
-	__m256i new_weight = _mm256_set1_epi16(_new_weight);
-	__m128i alpha_mask = _mm_set1_epi32(0x000000ff);
-	uint32_t * restrict src = buffer;
-	uint32_t * restrict dst = display_buffer;
-
-	for(uint32_t y = 0; y < BUFFER_HEIGHT; ++y, src += BUFFER_WIDTH, dst += BUFFER_WIDTH) {
-		for(uint32_t x = 0; x < BUFFER_WIDTH; x += 4) {
-			_mm_prefetch((char*)&src[x + 2 * BUFFER_WIDTH], _MM_HINT_T0);
-			_mm_prefetch((char*)&dst[x + 2 * BUFFER_WIDTH], _MM_HINT_T0);
-
-			__m128i new_pixels = _mm_load_si128((__m128i*)&src[x]);
-			__m128i old_pixels = _mm_load_si128((__m128i*)&dst[x]);
-
-			__m256i old_lo = _mm256_cvtepu8_epi16(old_pixels);
-			__m256i new_lo = _mm256_cvtepu8_epi16(new_pixels);
-
-			__m256i blended = _mm256_adds_epu16(_mm256_mullo_epi16(old_lo, old_weight), _mm256_mullo_epi16(new_lo, new_weight));
-			blended = _mm256_srli_epi16(blended, 8);
-
-			__m128i final_pixels = _mm_packus_epi16(_mm256_castsi256_si128(blended), _mm256_extracti128_si256(blended, 1));
-			final_pixels = _mm_or_si128(final_pixels, _mm_and_si128(old_pixels, alpha_mask));
-			_mm_store_si128((__m128i*)&dst[x], final_pixels);
-		}
-	}
-}
+// __attribute__((always_inline, hot))
+// static inline void apply_phosphor_decay(void) {
+// 	// PROFILE_FUNCTION();
+// 	__m256i old_weight = _mm256_set1_epi16(_old_weight);
+// 	__m256i new_weight = _mm256_set1_epi16(_new_weight);
+// 	__m128i alpha_mask = _mm_set1_epi32(0x000000ff);
+// 	uint32_t * restrict src = buffer;
+// 	uint32_t * restrict dst = display_buffer;
+
+// 	for(uint32_t y = 0; y < BUFFER_HEIGHT; ++y, src += BUFFER_WIDTH, dst += BUFFER_WIDTH) {
+// 		for(uint32_t x = 0; x < BUFFER_WIDTH; x += 4) {
+// 			_mm_prefetch((char*)&src[x + 2 * BUFFER_WIDTH], _MM_HINT_T0);
+// 			_mm_prefetch((char*)&dst[x + 2 * BUFFER_WIDTH], _MM_HINT_T0);
+
+// 			__m128i new_pixels = _mm_load_si128((__m128i*)&src[x]);
+// 			__m128i old_pixels = _mm_load_si128((__m128i*)&dst[x]);
+
+// 			__m256i old_lo = _mm256_cvtepu8_epi16(old_pixels);
+// 			__m256i new_lo = _mm256_cvtepu8_epi16(new_pixels);
+
+// 			__m256i blended = _mm256_adds_epu16(_mm256_mullo_epi16(old_lo, old_weight), _mm256_mullo_epi16(new_lo, new_weight));
+// 			blended = _mm256_srli_epi16(blended, 8);
+
+// 			__m128i final_pixels = _mm_packus_epi16(_mm256_castsi256_si128(blended), _mm256_extracti128_si256(blended, 1));
+// 			final_pixels = _mm_or_si128(final_pixels, _mm_and_si128(old_pixels, alpha_mask));
+// 			_mm_store_si128((__m128i*)&dst[x], final_pixels);
+// 		}
+// 	}
+// }