Prototype for SDL_Loading a RIX3

rix.c

#ifndef _GNU_SOURCE
#	define _GNU_SOURCE
#endif
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdint.h>
/* x86
#if HAVE_MMX || if HAVE_SSE || if HAVE_AVX2
#	include <xmmintrin.h>
#endif*/
/* PowerPC
#if HAVE_ALTIVEC
#	include <altivec.h>
#endif*/
/* ARM
#if HAVE_NEON
#	include <neon.h>
#endif*/

/*#include <SDL2/SDL.h>*/
#include <rix/rix.h>
#include <dataio/BigEndian.h>

/* Returns 0b00011111 subtract failed criterions
 * Signature is 0th bit
 * File size is 1st bit
 * Width is 2nd bit
 * Height is 3rd bit
 * Control is 4th bit
 * We lack use for 5th, 6th and 7th bit
 * Consider 31 a 100% match */

char IsFalloutRIX(FILE * Splash) /* File must be set on beginning */
{
	char result = 0;
	size_t ActualSize = 0;
	struct RIXheader RIX = {{0,0,0,0},0,0,0};
	if(!ftell(Splash))
	{
		fseek(Splash,0,SEEK_SET);
	}
	fread(&RIX.Signature,4,1,Splash);
	if(!BigEndian())
	{
		fread(&RIX.Width,2,1,Splash);
		fread(&RIX.Height,2,1,Splash);
		fread(&RIX.Unknown,2,1,Splash);
	}
	else
	{
		RIX.Width = ReadAlienEndian16(Splash);
		RIX.Height = ReadAlienEndian16(Splash);
		RIX.Unknown = ReadAlienEndian16(Splash);
	}
	fseek(Splash,0,SEEK_END);
	ActualSize = ftell(Splash);
	fseek(Splash,0,SEEK_SET);
	if(!strncmp(RIX.Signature,"RIX3",4))
	{
		result += 1;
	}
	if(ActualSize == 307978)
	{
		result += 2;
	}
	if(RIX.Width == 640)
	{
		result += 4;
	}
	if(RIX.Height == 480)
	{
		result += 8;
	}
	if(RIX.Unknown == 0x00AF)
	{
		result += 16;
	}
	return result;
}
/* MMX */
char * FalloutRIX2BMPMem(FILE * RIX)
{
	char RIXPalette[768];
	char * bitmap = malloc(921654);
	int32_t Position;
	uint16_t Counter;
	union
	{
		uint16_t w[8];
		uint64_t q[2];
	} Index;
	union
	{
		uint8_t b[24];
		uint64_t q[3];
	} BGR;
/* Simple BMP header for a 640x480x24 pic */
	char bmpheader[54] = {	66,77,54,16,14,0,0,0,0,0,54,0,0,0,40,0,0,0,128,2,0,0,224,1,0,0,
				1,0,24,0,0,0,0,0,0,16,14,0,19,11,0,0,19,11,0,0,0,0,0,0,0,0,0,0};
	memcpy(bitmap,bmpheader,54);
	Index.q[0] = 0x0003000300030003;
	Index.q[1] = 0x0404040404040404;
	__asm__("movq mm7, %0\n" : "=m"(Index.q[0]));
	__asm__("movq mm6, %0\n" : "=m"(Index.q[1])); /* We're initializing our multipliers */
	fseek(RIX,10,SEEK_SET);
	fread(RIXPalette,768,1,RIX);
	for(Position = 919734; Position >= 54; Position -= 1920)
	{
		for(Counter = 0; Counter < 1920; Counter += 24)
		{
			Index.w[0] = fgetc(RIX);
			Index.w[1] = fgetc(RIX);
			Index.w[2] = fgetc(RIX);
			Index.w[3] = fgetc(RIX);
			Index.w[4] = fgetc(RIX);
			Index.w[5] = fgetc(RIX);
			Index.w[6] = fgetc(RIX);
			Index.w[7] = fgetc(RIX);
			__asm__("movq mm0, %0\n"
					"pmullw mm0, mm7\n"
					"movq %0, mm0\n" : "=m"(Index.q[0]));
			__asm__("movq mm1, %0\n"
					"pmullw mm1, mm7\n"
					"movq %0, mm1\n" : "=m"(Index.q[1]));
			BGR.b[0] = RIXPalette[Index.w[0]+2];
			BGR.b[1] = RIXPalette[Index.w[0]+1];
			BGR.b[2] = RIXPalette[Index.w[0]];
			BGR.b[3] = RIXPalette[Index.w[1]+2];
			BGR.b[4] = RIXPalette[Index.w[1]+1];
			BGR.b[5] = RIXPalette[Index.w[1]];
			BGR.b[6] = RIXPalette[Index.w[2]+2];
			BGR.b[7] = RIXPalette[Index.w[2]+1];
			BGR.b[8] = RIXPalette[Index.w[2]];
			BGR.b[9] = RIXPalette[Index.w[3]+2];
			BGR.b[10] = RIXPalette[Index.w[3]+1];
			BGR.b[11] = RIXPalette[Index.w[3]];
			BGR.b[12] = RIXPalette[Index.w[4]+2];
			BGR.b[13] = RIXPalette[Index.w[4]+1];
			BGR.b[14] = RIXPalette[Index.w[4]];
			BGR.b[15] = RIXPalette[Index.w[5]+2];
			BGR.b[16] = RIXPalette[Index.w[5]+1];
			BGR.b[17] = RIXPalette[Index.w[5]];
			BGR.b[18] = RIXPalette[Index.w[6]+2];
			BGR.b[19] = RIXPalette[Index.w[6]+1];
			BGR.b[20] = RIXPalette[Index.w[6]];
			BGR.b[21] = RIXPalette[Index.w[7]+2];
			BGR.b[22] = RIXPalette[Index.w[7]+1];
			BGR.b[23] = RIXPalette[Index.w[7]];
			__asm__("movq mm2, %0\n"
					"pmulhw mm2, mm6\n"
					"movq %0, mm2\n" : "=m"(BGR.q[0]));
			__asm__("movq mm3, %0\n"
					"pmulhw mm3, mm6\n"
					"movq %0, mm3\n" : "=m"(BGR.q[1]));
			__asm__("movq mm4, %0\n"
					"pmulhw mm4, mm6\n"
					"movq %0, mm4\n" : "=m"(BGR.q[2]));
			memcpy(bitmap+Position+Counter,BGR.b,24);
		}
	}
	return bitmap;
}

int main(int argc, char * argv[])
{
	int result = 0;
	FILE * input = NULL;
	FILE * output = NULL;
	char * memory = NULL;
	if(argc != 3)
	{
		exit(1);
	}
	input = fopen(argv[1],"rb");
	if(IsFalloutRIX(input) == 31)
	{
		memory = FalloutRIX2BMPMem(input);
		output = fopen(argv[2],"wb");
		fwrite(memory,921654,1,output);
		free(memory);
		fclose(output);
	}
	fclose(input);
	return result;
}

Revision 6 produces a file whose difference is presented here http://imgur.com/a/Xngv4