comparison libmpcodecs/vf_spp.c @ 13720:821f464b4d90

adapting existing mmx/mmx2/sse/3dnow optimizations so they work on x86_64

13719:43ecd6a73ec0

			"paddw %%mm4, %%mm1	\n\t"
			"psraw %%mm2, %%mm0	\n\t"
			"psraw %%mm2, %%mm1	\n\t"
			"packuswb %%mm1, %%mm0	\n\t"
			"movq %%mm0, (%1) 	\n\t"
			"addl $16, %0		\n\t"
			"addl $8, %1		\n\t"
			"cmpl %2, %1		\n\t"
			" jb 1b			\n\t"
			: "+r" (src1), "+r"(dst1)
			: "r"(dst + width), "r"(dither[y]), "g"(log2_scale), "g"(6-log2_scale)
		);
		src += src_stride;

13720:821f464b4d90

			"paddw %%mm4, %%mm1	\n\t"
			"psraw %%mm2, %%mm0	\n\t"
			"psraw %%mm2, %%mm1	\n\t"
			"packuswb %%mm1, %%mm0	\n\t"
			"movq %%mm0, (%1) 	\n\t"
			"add $16, %0		\n\t"
			"add $8, %1		\n\t"
			"cmp %2, %1		\n\t"
			" jb 1b			\n\t"
			: "+r" (src1), "+r"(dst1)
			: "r"(dst + width), "r"(dither[y]), "g"(log2_scale), "g"(6-log2_scale)
		);
		src += src_stride;