From 185375ece1c5a3d93d32fec0c170b7e2ceed0b2b Mon Sep 17 00:00:00 2001 From: Alug Date: Thu, 5 Jun 2025 01:24:58 +0200 Subject: [PATCH] use ternaries for column clamp it is slightly faster for me lel also lets us get rid of an extra variable npow2min is also declared static as its always -1 --- src/r_draw_column.c | 100 ++++++++------------------------------------ 1 file changed, 17 insertions(+), 83 deletions(-) diff --git a/src/r_draw_column.c b/src/r_draw_column.c index 60c39db56..abfd515d4 100644 --- a/src/r_draw_column.c +++ b/src/r_draw_column.c @@ -28,8 +28,6 @@ void R_DrawColumn(void) UINT8 *restrict dest; intptr_t frac; intptr_t fracstep; - INT32 npow2min; - INT32 npow2max; count = dc_yh - dc_yl; @@ -59,8 +57,9 @@ void R_DrawColumn(void) const lighttable_t *restrict colormap = dc_colormap; intptr_t heightmask = dc_sourcelength-1; - npow2min = -1; - npow2max = dc_sourcelength; + + static const INT32 npow2min = -1; + const INT32 npow2max = dc_sourcelength; if (dc_sourcelength & heightmask) // not a power of 2 -- killough { @@ -89,17 +88,7 @@ void R_DrawColumn(void) // -1 is the lower clamp bound because column posts have a "safe" byte before the real data // and a few bytes after as well - - // jartha: faster on my AMD FX-6300 CPU. - // Faster than ternaries, faster than std::min/std::max. Don't ask me why. - // I tested by viewing a non-PO2 texture from a consistent distance so it covered the entire screen. - // The framerate difference was about 50 frames at 640x400. - INT32 n = frac >> FRACBITS; - if (n < npow2min) - n = npow2min; - if (n > npow2max) - n = npow2max; - *dest = colormap[source[n]]; + *dest = colormap[source[CLAMP(frac >> FRACBITS, npow2min, npow2max)]]; dest += vid.width; @@ -152,9 +141,6 @@ void R_Draw2sMultiPatchColumn(void) intptr_t frac; intptr_t fracstep; - INT32 npow2min; - INT32 npow2max; - count = dc_yh - dc_yl; if (count < 0) // Zero length, column does not exceed a pixel. @@ -183,8 +169,8 @@ void R_Draw2sMultiPatchColumn(void) intptr_t heightmask = dc_sourcelength-1; UINT8 val; - npow2min = -1; - npow2max = dc_sourcelength; + static const INT32 npow2min = -1; + const INT32 npow2max = dc_sourcelength; if (dc_sourcelength & heightmask) // not a power of 2 -- killough { @@ -213,18 +199,7 @@ void R_Draw2sMultiPatchColumn(void) // -1 is the lower clamp bound because column posts have a "safe" byte before the real data // and a few bytes after as well - - // jartha: faster on my AMD FX-6300 CPU. - // Faster than ternaries, faster than std::min/std::max. Don't ask me why. - // I tested by viewing a non-PO2 texture from a consistent distance so it covered the entire screen. - // The framerate difference was about 50 frames at 640x400. - INT32 n = frac >> FRACBITS; - if (n < npow2min) - n = npow2min; - if (n > npow2max) - n = npow2max; - - val = source[n]; + val = source[CLAMP(frac >> FRACBITS, npow2min, npow2max)]; if (val != TRANSPARENTPIXEL) { @@ -256,6 +231,7 @@ void R_Draw2sMultiPatchColumn(void) while ((count -= 2) >= 0) // texture height is a power of 2 { val = source[(frac>>FRACBITS) & heightmask]; + if (val != TRANSPARENTPIXEL) { *dest = colormap[val]; @@ -292,9 +268,6 @@ void R_Draw2sMultiPatchTranslucentColumn(void) intptr_t frac; intptr_t fracstep; - INT32 npow2min; - INT32 npow2max; - count = dc_yh - dc_yl; if (count < 0) // Zero length, column does not exceed a pixel. @@ -324,8 +297,8 @@ void R_Draw2sMultiPatchTranslucentColumn(void) intptr_t heightmask = dc_sourcelength-1; register UINT8 val; - npow2min = -1; - npow2max = dc_sourcelength; + static const INT32 npow2min = -1; + const INT32 npow2max = dc_sourcelength; if (dc_sourcelength & heightmask) // not a power of 2 -- killough { @@ -354,18 +327,7 @@ void R_Draw2sMultiPatchTranslucentColumn(void) // -1 is the lower clamp bound because column posts have a "safe" byte before the real data // and a few bytes after as well - - // jartha: faster on my AMD FX-6300 CPU. - // Faster than ternaries, faster than std::min/std::max. Don't ask me why. - // I tested by viewing a non-PO2 texture from a consistent distance so it covered the entire screen. - // The framerate difference was about 50 frames at 640x400. - INT32 n = frac >> FRACBITS; - if (n < npow2min) - n = npow2min; - if (n > npow2max) - n = npow2max; - - val = source[n]; + val = source[CLAMP(frac >> FRACBITS, npow2min, npow2max)]; if (val != TRANSPARENTPIXEL) { @@ -479,9 +441,6 @@ void R_DrawTranslucentColumn(void) intptr_t frac; intptr_t fracstep; - INT32 npow2min; - INT32 npow2max; - count = dc_yh - dc_yl + 1; if (count <= 0) // Zero length, column does not exceed a pixel. @@ -508,8 +467,8 @@ void R_DrawTranslucentColumn(void) const lighttable_t *restrict colormap = dc_colormap; intptr_t heightmask = dc_sourcelength-1; - npow2min = -1; - npow2max = dc_sourcelength; + static const INT32 npow2min = -1; + const INT32 npow2max = dc_sourcelength; if (dc_sourcelength & heightmask) { @@ -538,18 +497,7 @@ void R_DrawTranslucentColumn(void) // -1 is the lower clamp bound because column posts have a "safe" byte before the real data // and a few bytes after as well - - // jartha: faster on my AMD FX-6300 CPU. - // Faster than ternaries, faster than std::min/std::max. Don't ask me why. - // I tested by viewing a non-PO2 texture from a consistent distance so it covered the entire screen. - // The framerate difference was about 50 frames at 640x400. - INT32 n = frac >> FRACBITS; - if (n < npow2min) - n = npow2min; - if (n > npow2max) - n = npow2max; - - *dest = *(transmap + (colormap[source[n]]<<8) + (*dest)); + *dest = *(transmap + (colormap[source[CLAMP(frac >> FRACBITS, npow2min, npow2max)]]<<8) + (*dest)); dest += vid.width; @@ -591,9 +539,6 @@ void R_DrawTranslatedTranslucentColumn(void) intptr_t frac; intptr_t fracstep; - INT32 npow2min; - INT32 npow2max; - count = dc_yh - dc_yl + 1; if (count <= 0) // Zero length, column does not exceed a pixel. @@ -617,8 +562,8 @@ void R_DrawTranslatedTranslucentColumn(void) // This is as fast as it gets. intptr_t heightmask = dc_sourcelength-1; - npow2min = -1; - npow2max = dc_sourcelength; + static const INT32 npow2min = -1; + const INT32 npow2max = dc_sourcelength; if (dc_sourcelength & heightmask) { @@ -647,18 +592,7 @@ void R_DrawTranslatedTranslucentColumn(void) // -1 is the lower clamp bound because column posts have a "safe" byte before the real data // and a few bytes after as well - - // jartha: faster on my AMD FX-6300 CPU. - // Faster than ternaries, faster than std::min/std::max. Don't ask me why. - // I tested by viewing a non-PO2 texture from a consistent distance so it covered the entire screen. - // The framerate difference was about 50 frames at 640x400. - INT32 n = frac >> FRACBITS; - if (n < npow2min) - n = npow2min; - if (n > npow2max) - n = npow2max; - - *dest = *(dc_transmap + (dc_colormap[dc_translation[dc_source[n]]]<<8) + (*dest)); + *dest = *(dc_transmap + (dc_colormap[dc_translation[dc_source[CLAMP(frac >> FRACBITS, npow2min, npow2max)]]]<<8) + (*dest)); dest += vid.width;