*misc

14 files changed, 648 insertions, 566 deletions

diff --git a/src/core/clip.c b/src/core/clip.c
index 382c766..1deb4ee 100644
--- a/src/core/clip.c
+++ b/src/core/clip.c
@@ -1,5 +1,6 @@
 #include "clip.h"
 #include "shader.h"
+#include "../util/type.h"
 
 typedef enum {
 	POSITIVE_W,
@@ -60,73 +61,75 @@ static float get_intersect_ratio(Vec4* prev, Vec4* curr, Plane plane) {
 		return 0;
 	}
 }
-
-static bool clip_against_plane(Plane plane, uint varying_flag, ClippedBuffer* src_buffer, ClippedBuffer* dst_buffer) {
-	int idx = 0;
-	for (int i = 0; i < src_buffer->count; ++i) {
-		ClippedVert* prev = &src_buffer->vertices[(i - 1 + src_buffer->count) % src_buffer->count];
-		ClippedVert* curr = &src_buffer->vertices[i];
-
-		bool prev_inside = is_inside_plane(&prev->clip_coord, plane);
-		bool curr_inside = is_inside_plane(&curr->clip_coord, plane);
-
-		if (prev_inside != curr_inside) {
-			ClippedVert* dst = &dst_buffer->vertices[idx];
-			float t = get_intersect_ratio(&prev->clip_coord, &curr->clip_coord, plane);
-			ssrS_lerpvec4(t, &prev->clip_coord, &curr->clip_coord, &dst->clip_coord);
-			/*set varying varibles*/
-			if (varying_flag & VARYING_BASIC) {
-				if (varying_flag & VARYING_POSITION) ssrS_lerpvec3(t, &prev->vertex.position, &curr->vertex.position, &dst->vertex.position);
-				if (varying_flag & VARYING_NORMAL) ssrS_lerpvec3(t, &prev->vertex.normal, &curr->vertex.normal, &dst->vertex.normal);
-				if (varying_flag & VARYING_TANGENT) ssrS_lerpvec3(t, &prev->vertex.tangent, &curr->vertex.tangent, &dst->vertex.tangent);
-				if (varying_flag & VARYING_TEXCOORD) ssrS_lerpvec2(t, &prev->vertex.texcoord, &curr->vertex.texcoord, &dst->vertex.texcoord);
-				if (varying_flag & VARYING_COLOR) ssrS_lerpcolor(t, prev->vertex.color, curr->vertex.color, &dst->vertex.color);
-				if (varying_flag & VARYING_JOINT) ssrS_lerpvec4(t, &prev->vertex.joint, &curr->vertex.joint, &dst->vertex.joint);
-				if (varying_flag & VARYING_WEIGHT) ssrS_lerpvec4(t, &prev->vertex.weight, &curr->vertex.weight, &dst->vertex.weight);
-			}
-			if (varying_flag & VARYING_EXTRA) {
-				int j = 0;
-				if (varying_flag & VARYING_NUM) {
-					for (j = 0; j < REG_NUM_COUNT; ++j) {
-						if (varying_flag & (VARYING_NUM_00 << j)) {
-							ssrS_lerpnum(t, prev->num[j], curr->num[j], &dst->num[j]);
-						}
-					}
-				}
-				if (varying_flag & VARYING_V2) {
-					for (j = 0; j < REG_V2_COUNT; ++j) {
-						if (varying_flag & (VARYING_V2_00 << j)) {
-							ssrS_lerpvec2(t, &prev->v2[j], &curr->v2[j], &dst->v2[j]);
-						}
-					}
-				}
-				if (varying_flag & VARYING_V3) {
-					for (j = 0; j < REG_V3_COUNT; ++j) {
-						if (varying_flag & (VARYING_V3_00 << j)) {
-							ssrS_lerpvec3(t, &prev->v3[j], &curr->v3[j], &dst->v3[j]);
-						}
-					}
-				}
-				if (varying_flag & VARYING_V4) {
-					for (j = 0; j < REG_V4_COUNT; ++j) {
-						if (varying_flag & (VARYING_V4_00 << j)) {
-							ssrS_lerpvec4(t, &prev->v4[j], &curr->v4[j], &dst->v4[j]);
-						}
-					}
-				}
-			}
-			++idx;
-		}
-
-		if (curr_inside) {
-			ClippedVert* dst = &dst_buffer->vertices[idx];
-			*dst = *curr;
-			++idx;
-		}
-	}
-	dst_buffer->count = idx;
-	return idx < 3;
-}
+//
+//static bool clip_against_plane(Plane plane, uint varying_flag, ClippedBuffer* src_buffer, ClippedBuffer* dst_buffer) {
+//	int idx = 0;
+//	for (int i = 0; i < src_buffer->count; ++i) {
+//		ClippedVert* prev = &src_buffer->vertices[(i - 1 + src_buffer->count) % src_buffer->count];
+//		ClippedVert* curr = &src_buffer->vertices[i];
+//
+//		bool prev_inside = is_inside_plane(&prev->clip_coord, plane);
+//		bool curr_inside = is_inside_plane(&curr->clip_coord, plane);
+//
+//		if (prev_inside != curr_inside) {
+//			ClippedVert* dst = &dst_buffer->vertices[idx];
+//			float t = get_intersect_ratio(&prev->clip_coord, &curr->clip_coord, plane);
+//			ssrS_lerpvec4(t, &prev->clip_coord, &curr->clip_coord, &dst->clip_coord);
+//			/*set varying varibles*/
+//			if (varying_flag & VARYING_BASIC) {
+//				if (varying_flag & VARYING_POSITION) ssrS_lerpvec3(t, &prev->vertex.position, &curr->vertex.position, &dst->vertex.position);
+//				if (varying_flag & VARYING_NORMAL) ssrS_lerpvec3(t, &prev->vertex.normal, &curr->vertex.normal, &dst->vertex.normal);
+//				if (varying_flag & VARYING_TANGENT) ssrS_lerpvec3(t, &prev->vertex.tangent, &curr->vertex.tangent, &dst->vertex.tangent);
+//				if (varying_flag & VARYING_TEXCOORD) ssrS_lerpvec2(t, &prev->vertex.texcoord, &curr->vertex.texcoord, &dst->vertex.texcoord);
+//				if (varying_flag & VARYING_COLOR) ssrS_lerpcolor(t, prev->vertex.color, curr->vertex.color, &dst->vertex.color);
+///*
+//				if (varying_flag & VARYING_JOINT) ssrS_lerpvec4(t, &prev->vertex.joint, &curr->vertex.joint, &dst->vertex.joint);
+//				if (varying_flag & VARYING_WEIGHT) ssrS_lerpvec4(t, &prev->vertex.weight, &curr->vertex.weight, &dst->vertex.weight);
+//*/
+//			}
+//			if (varying_flag & VARYING_EXTRA) {
+//				int j = 0;
+//				if (varying_flag & VARYING_NUM) {
+//					for (j = 0; j < REG_NUM_COUNT; ++j) {
+//						if (varying_flag & (VARYING_NUM_00 << j)) {
+//							ssrS_lerpnum(t, prev->num[j], curr->num[j], &dst->num[j]);
+//						}
+//					}
+//				}
+//				if (varying_flag & VARYING_V2) {
+//					for (j = 0; j < REG_V2_COUNT; ++j) {
+//						if (varying_flag & (VARYING_V2_00 << j)) {
+//							ssrS_lerpvec2(t, &prev->v2[j], &curr->v2[j], &dst->v2[j]);
+//						}
+//					}
+//				}
+//				if (varying_flag & VARYING_V3) {
+//					for (j = 0; j < REG_V3_COUNT; ++j) {
+//						if (varying_flag & (VARYING_V3_00 << j)) {
+//							ssrS_lerpvec3(t, &prev->v3[j], &curr->v3[j], &dst->v3[j]);
+//						}
+//					}
+//				}
+//				if (varying_flag & VARYING_V4) {
+//					for (j = 0; j < REG_V4_COUNT; ++j) {
+//						if (varying_flag & (VARYING_V4_00 << j)) {
+//							ssrS_lerpvec4(t, &prev->v4[j], &curr->v4[j], &dst->v4[j]);
+//						}
+//					}
+//				}
+//			}
+//			++idx;
+//		}
+//
+//		if (curr_inside) {
+//			ClippedVert* dst = &dst_buffer->vertices[idx];
+//			*dst = *curr;
+//			++idx;
+//		}
+//	}
+//	dst_buffer->count = idx;
+//	return idx < 3;
+//}
 
 #define CLIP(plane, from, to) \
 do {																					             \
@@ -136,6 +139,36 @@ do {																					             \
 	}																							           \
 }while(0)
 
+ClippedBuffer clip_buffer;
+
+byte* clip_buffer_data[REG_TOTAL] =
+{
+	&clip_buffer.temp_reg_num[0],
+	&clip_buffer.temp_reg_num[1],
+	&clip_buffer.temp_reg_num[2],
+	&clip_buffer.temp_reg_num[3],
+	&clip_buffer.temp_reg_v2[0],
+	&clip_buffer.temp_reg_v2[1],
+	&clip_buffer.temp_reg_v2[2],
+	&clip_buffer.temp_reg_v2[3],
+	&clip_buffer.temp_reg_v2[4],
+	&clip_buffer.temp_reg_v2[5],
+	&clip_buffer.temp_reg_v3[0],
+	&clip_buffer.temp_reg_v3[1],
+	&clip_buffer.temp_reg_v3[2],
+	&clip_buffer.temp_reg_v3[3],
+	&clip_buffer.temp_reg_v3[4],
+	&clip_buffer.temp_reg_v3[5],
+	&clip_buffer.temp_reg_v3[6],
+	&clip_buffer.temp_reg_v3[7],
+	&clip_buffer.temp_reg_v4[0],
+	&clip_buffer.temp_reg_v4[1],
+	&clip_buffer.temp_reg_v4[2],
+	&clip_buffer.temp_reg_v4[3],
+	&clip_buffer.temp_reg_v4[4],
+	&clip_buffer.temp_reg_v4[5],
+};
+
 bool clip_triangle(Vec4* c0, Vec4* c1, Vec4* c2, Vert* v0, Vert* v1, Vert* v2, uint varying_flag, ClippedBuffer* buffer) {
 	bool is_visible = is_vertex_visible(c0)
 	&& is_vertex_visible(c1)
@@ -144,55 +177,53 @@ bool clip_triangle(Vec4* c0, Vec4* c1, Vec4* c2, Vert* v0, Vert* v1, Vert* v2, u
 	if (is_visible) {
 		return 0; /*no need to clip*/
 	}
-
-	/*clipping it*/
-
-	static ClippedBuffer temp;
-
-	/*copy vert data to temp*/
-	temp.count = 3;
-#define INIT_CLIP_VERT(idx) \
-	temp.vertices[idx].clip_coord = *c##idx;												 	 \
-	temp.vertices[idx].vertex = *v##idx;														 	 \
-	if (varying_flag & VARYING_EXTRA) {															 	 \
-		int i = 0;																										 	 \
-		if(varying_flag & VARYING_NUM) {															 	 \
-			for (i = 0; i < REG_NUM_COUNT; ++i) {													 \
-				if (varying_flag & (VARYING_NUM_00 << i))										 \
-					temp.vertices[idx].num[i] = reg_num[i].num[v##idx->index]; \
-			}																															 \
-		}																															 	 \
-		if(varying_flag & VARYING_V2) {															 	 \
-			for (i = 0; i < REG_V2_COUNT; ++i) {													 \
-				if (varying_flag & (VARYING_V2_00 << i))										 \
-					temp.vertices[idx].v2[i] = reg_v2[i].v2[v##idx->index];		 \
-			}																															 \
-		}																																 \
-		if(varying_flag & VARYING_V3) {															 	 \
-			for (i = 0; i < REG_V3_COUNT; ++i) {													 \
-				if (varying_flag & (VARYING_V3_00 << i))										 \
-					temp.vertices[idx].v3[i] = reg_v3[i].v3[v##idx->index];		 \
-			}																															 \
-		}																																 \
-		if(varying_flag & VARYING_V4) {															 	 \
-			for (i = 0; i < REG_V4_COUNT; ++i) {													 \
-				if (varying_flag & (VARYING_V4_00 << i))										 \
-					temp.vertices[idx].v4[i] = reg_v4[i].v4[v##idx->index];		 \
-			}																															 \
-		}																																 \
-	}
-	INIT_CLIP_VERT(0);
-	INIT_CLIP_VERT(1);
-	INIT_CLIP_VERT(2);
-#undef INIT_CLIP_VERT
-
-	CLIP(POSITIVE_W, &temp, buffer);
-	CLIP(POSITIVE_X, buffer, &temp);
-	CLIP(NEGATIVE_X, &temp, buffer);
-	CLIP(POSITIVE_Y, buffer, &temp);
-	CLIP(NEGATIVE_Y, &temp, buffer);
-	CLIP(POSITIVE_Z, buffer, &temp);
-	CLIP(NEGATIVE_Z, &temp, buffer);
-
-	return 1;
+//
+//	/*clipping it*/
+//
+//	/*copy vert data to clip_buffer*/
+//	clip_buffer.count = 3;
+//#define INIT_CLIP_VERT(idx) \
+//	clip_buffer.vertices[idx].clip_coord = *c##idx;												 	 \
+//	clip_buffer.vertices[idx].vertex = *v##idx;														 	 \
+//	if (varying_flag & VARYING_EXTRA) {															 	 \
+//		int i = 0;																										 	 \
+//		if(varying_flag & VARYING_NUM) {															 	 \
+//			for (i = 0; i < REG_NUM_COUNT; ++i) {													 \
+//				if (varying_flag & (VARYING_NUM_00 << i))										 \
+//					clip_buffer.vertices[idx].num[i] = reg_num[i].num[v##idx->index]; \
+//			}																															 \
+//		}																															 	 \
+//		if(varying_flag & VARYING_V2) {															 	 \
+//			for (i = 0; i < REG_V2_COUNT; ++i) {													 \
+//				if (varying_flag & (VARYING_V2_00 << i))										 \
+//					clip_buffer.vertices[idx].v2[i] = reg_v2[i].v2[v##idx->index];		 \
+//			}																															 \
+//		}																																 \
+//		if(varying_flag & VARYING_V3) {															 	 \
+//			for (i = 0; i < REG_V3_COUNT; ++i) {													 \
+//				if (varying_flag & (VARYING_V3_00 << i))										 \
+//					clip_buffer.vertices[idx].v3[i] = reg_v3[i].v3[v##idx->index];		 \
+//			}																															 \
+//		}																																 \
+//		if(varying_flag & VARYING_V4) {															 	 \
+//			for (i = 0; i < REG_V4_COUNT; ++i) {													 \
+//				if (varying_flag & (VARYING_V4_00 << i))										 \
+//					clip_buffer.vertices[idx].v4[i] = reg_v4[i].v4[v##idx->index];		 \
+//			}																															 \
+//		}																																 \
+//	}
+//	INIT_CLIP_VERT(0);
+//	INIT_CLIP_VERT(1);
+//	INIT_CLIP_VERT(2);
+//#undef INIT_CLIP_VERT
+//
+//	CLIP(POSITIVE_W, &clip_buffer, buffer);
+//	CLIP(POSITIVE_X, buffer, &clip_buffer);
+//	CLIP(NEGATIVE_X, &clip_buffer, buffer);
+//	CLIP(POSITIVE_Y, buffer, &clip_buffer);
+//	CLIP(NEGATIVE_Y, &clip_buffer, buffer);
+//	CLIP(POSITIVE_Z, buffer, &clip_buffer);
+//	CLIP(NEGATIVE_Z, &clip_buffer, buffer);
+//
+//	return 1;
 }
diff --git a/src/core/clip.h b/src/core/clip.h
index b6f2702..c22bb8d 100644
--- a/src/core/clip.h
+++ b/src/core/clip.h
@@ -6,25 +6,37 @@
 
 #define LERP(t,a,b) ((1-(t))*(a)+(t)*(b))
 
-typedef struct ClippedVert {
+typedef struct {
 	/*clipping coord*/
 	Vec4 clip_coord;
 	/*vertex data*/
 	Vert vertex; 
 	/*register values*/
+/*
 	float num[4];
 	Vec2 v2[6];
 	Vec3 v3[8];
 	Vec4 v4[6];
+*/
 } ClippedVert;
 
 #define CLIP_BUFFER_SIZE 6
 
-typedef struct ClippedBuffer {
+typedef struct {
 	ClippedVert vertices[CLIP_BUFFER_SIZE];
 	uint count;
+	/*temp register*/
+	float temp_reg_num[4][CLIP_BUFFER_SIZE];
+	Vec2 temp_reg_v2[6][CLIP_BUFFER_SIZE];
+	Vec3 temp_reg_v3[8][CLIP_BUFFER_SIZE];
+	Vec4 temp_reg_v4[6][CLIP_BUFFER_SIZE];
 } ClippedBuffer;
 
+ClippedBuffer clip_buffer;
+
+typedef void* (*BcpInterpolator)(Vec3* bc, void* a, void* b, void* c, void* out);
+typedef void* (*LinearInterpolator)(float t, void* a, void* b, void* c, void* out);
+
 bool clip_triangle(Vec4* c0, Vec4* c1, Vec4* c2, Vert* v0, Vert* v1, Vert* v2, uint varying_flag, ClippedBuffer* clipped);
 
 uint clip_line();
diff --git a/src/core/device.c b/src/core/device.c
index cf9fee2..3566448 100644
--- a/src/core/device.c
+++ b/src/core/device.c
@@ -21,7 +21,7 @@ typedef struct VertexAttr {
 /*
 ** 状态结构
 */
-struct {
+static struct {
 	/* MVP矩阵栈 */
 	Mat4 matrices[3][MATRIXDEPTH];
 	uint matrixtop[3];
@@ -33,7 +33,7 @@ struct {
 	uint buffersize; /*framebuffer size in bytess*/
 
 	/* zbuffer 深度会被映射到0~1的非线性区间*/
-	uint* zbuffer;
+	float* zbuffer;
 
 	/* 顶点数据 */
 	Vert* verts; uint nverts;
@@ -43,12 +43,19 @@ struct {
 	Program* program;
 	UniformCollection uniforms;
 
-	/*texture relavent*/
+	/* texture relavent */
 	FilterMode filtermode;
 	WrapMode wrapmode;
 
+	//FrameBuffer* framebuffer;
+
 	uint enable;
 
+	struct {
+		ssr_BlendFactor src;
+		ssr_BlendFactor dst;
+	} blendfactor;
+
 } state;
 
 /*
@@ -75,11 +82,14 @@ void ssr_init(ssr_Config* conf) {
 		state.framebuffer = conf->target;
 	}
 	state.buffersize = sizeof(Color) * config.width * config.height;
-	state.zbuffer = ssrM_newvector(uint, config.width * config.height);
-	memset(state.zbuffer, 0xff, sizeof(uint)*config.width*config.height);
+	state.zbuffer = ssrM_newvector(float, config.width * config.height);
+	//memset(state.zbuffer, 0xff, sizeof(uint)*config.width*config.height);
 
 	state.filtermode = FILTERMODE_POINT;
 	state.wrapmode = WRAPMODE_CLAMP;
+
+	state.blendfactor.src = BLENDFACTOR_SRC_ALPHA;
+	state.blendfactor.dst = BLENDFACTOR_ONE_MINUS_SRC_ALPHA;
 }
 
 float ssr_getaspect() {
@@ -110,6 +120,19 @@ WrapMode ssr_getwrapmode() {
 	return state.wrapmode;
 }
 
+void ssr_bindframebuffer(FrameBuffer* fbo) {
+
+}
+
+void ssr_unbindframebuffer() {
+
+}
+
+void ssr_setblendfunc(ssr_BlendFactor sfactor, ssr_BlendFactor dfactor) {
+	state.blendfactor.src = sfactor;
+	state.blendfactor.dst = dfactor;
+}
+
 void ssr_matrixmode(ssr_MatrixMode mode) {
 	state.matrixmode = mode;
 }
@@ -181,6 +204,11 @@ void ssr_getmvp(Mat4* out) {
 	mat4_multiply(&GETMATRIX(MATRIX_PROJECTION), out, out);
 }
 
+void ssr_getm(Mat4* out) {
+	ssr_assert(out);
+	*out = GETMATRIX(MATRIX_MODEL);
+}
+
 void ssr_getmv(Mat4* out) {
 	ssr_assert(out);
 	mat4_multiply(&GETMATRIX(MATRIX_VIEW), &GETMATRIX(MATRIX_MODEL), out);
@@ -240,7 +268,10 @@ void ssr_clearcolor(Color color) {
 }
 
 void ssr_cleardepth() {
-	memset(state.zbuffer, 0xff, sizeof(uint)*config.width*config.height);
+	//memset(state.zbuffer, 0xff, sizeof(uint)*config.width*config.height);
+	for (int i = 0; i < config.width * config.height; ++i) {
+		state.zbuffer[i] = 1;
+	}
 }
 
 void ssr_putpoint(uint screenx, uint screeny, Color color) {
@@ -249,24 +280,24 @@ void ssr_putpoint(uint screenx, uint screeny, Color color) {
 	BUFFER[screeny * config.width + screenx] = color;
 }
 
+void ssr_putpoint32(uint screenx, uint screeny, Color32* c32) {
+	Color c = color32_tocolor(c32);
+	ssr_putpoint(screenx, screeny, c);
+}
+
 Color ssr_color(unsigned char r, unsigned char g, unsigned char b, unsigned char a) {
 	unsigned int c = (a << 24) | (r << 16) | (g << 8) | b;
 	return c;
 }
 
-bool ssr_testdepth(uint x, uint y, uint depth) {
-	if (!contains(x, y, 0, config.width - 1, 0, config.height - 1))
-		return 0;
+bool ssr_testdepth(uint x, uint y, float depth){
 	uint off = x + y * config.width;
-	if (state.zbuffer[off] < depth)
-		return 0;
-	state.zbuffer[off] = depth;
-	return 1;
+	return state.zbuffer[off] >= depth;
 }
 
-bool ssr_testdepthf(uint x, uint y, float depth) {
-	uint d = UINT_MAX * (double)depth;
-	return ssr_testdepth(x, y, d);
+void ssr_writedepth(uint x, uint y, float depth) {
+	uint off = x + y * config.width;
+	state.zbuffer[off] = depth;
 }
 
 void ssrU_viewport(Vec2* p, Vec2* out) {
@@ -278,6 +309,26 @@ void ssrU_viewport(Vec2* p, Vec2* out) {
 	out->y = (int)round((1 - p->y) * (halfh - 0.5f));
 }
 
+void ssr_blend(Color32* src, Color32* dst, Color32* out) {
+	ssrU_blend(state.blendfactor.src, state.blendfactor.dst, src, dst, out);
+}
+
+void ssrU_blend(ssr_BlendFactor sfactor
+	, ssr_BlendFactor dfactor
+	, Color32* src
+	, Color32* dst
+	, Color32* out
+) {
+	ssr_assert(src && dst && out);
+	
+}
+
+Color32 ssr_getfbocolor(uint x, uint y) {
+	Color c = state.framebuffer[x + y * config.width];
+	Color32 c32; color_tocolor32(c, &c32);
+	return c32;
+}
+
 void ssr_bindvertices(Vert* verts, int nverts, uint* indices, int nprims) {
 	ssr_assert(verts && indices);
 	state.verts = verts;
@@ -301,7 +352,7 @@ FragmentShaderIn ssr_frag_in;
 static struct {
 	Vec4* coords;
 	uint length;
-} clipping_buffer;
+} clip_coords;
 
 static ClippedBuffer clipped_buffer; /*clipping result*/
 
@@ -310,6 +361,7 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 
 	static Mat4 mvp, mv;
 
+	/*set built-in uniforms*/
 	ssr_getmvp(&mvp);
 	ssr_getmv(&mv);
 	state.uniforms.model      = &GETMATRIX(MATRIX_MODEL);
@@ -323,15 +375,17 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 
 	/*prepare registers if necessary*/
 	if (use_extra_varyings) {
-		ssrS_setregisters(varying_flag, state.nverts);
+		ssrS_openregs(varying_flag);
+		ssrS_setregisters(state.nverts);
+		ssrS_setactiveregr();
 	}
 
 	/*resize clipping coords buffer*/
-	if (clipping_buffer.length < state.nverts) {
+	if (clip_coords.length < state.nverts) {
 		ssrM_resizevector(
 			Vec4, 
-			clipping_buffer.coords, 
-			clipping_buffer.length, 
+			clip_coords.coords, 
+			clip_coords.length, 
 			state.nverts, 
 			FALSE
 		);
@@ -343,9 +397,9 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 		ssr_vert_in.vertex = vert;
 		/*set register pointers*/
 		if (use_extra_varyings) {
-			ssrS_setupregisterpoints(varying_flag, vert->index);
+			ssrS_setupregisterpoints(vert->index);
 		}
-		state.program->vertexshader(&state.uniforms, &ssr_vert_in, &clipping_buffer.coords[i]);
+		state.program->vertexshader(&state.uniforms, &ssr_vert_in, &clip_coords.coords[i]);
 	}
 
 	/*set register pointer to frag-in*/
@@ -363,9 +417,9 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 			i1 = state.indices[i * 3 + 1];
 			i2 = state.indices[i * 3 + 2];
 
-			h0 = &clipping_buffer.coords[i0],
-			h1 = &clipping_buffer.coords[i1],
-			h2 = &clipping_buffer.coords[i2];
+			h0 = &clip_coords.coords[i0],
+			h1 = &clip_coords.coords[i1],
+			h2 = &clip_coords.coords[i2];
 
 			v0 = &state.verts[i0];
 			v1 = &state.verts[i1];
@@ -379,8 +433,9 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 				ab.y = h1->y * w1 - h0->y * w0;
 				ac.x = h2->x * w2 - h0->x * w0;
 				ac.y = h2->y * w2 - h0->y * w0;
-				if (ab.x * ac.y - ab.y * ac.x <= 0) 
+				if (ab.x * ac.y - ab.y * ac.x <= 0) {
 					continue; /*cull*/
+				}
 			}
 
 			/*clipping*/
@@ -388,20 +443,21 @@ void ssr_draw(ssr_PrimitiveType primitive) {
 
 			/*rasterization*/
 			if (!clipped) {
-				ssrR_triangle(h0, h1, h2, v0, v1, v2, TRUE, NULL, NULL, NULL, state.program, &state.uniforms);
+				/*ssrS_setactiveregr();*/
+				ssrR_triangle(h0, h1, h2, v0, v1, v2, state.program, &state.uniforms);
 			} else {
+				ssrS_setactiveregc();
 				if (clipped_buffer.count >= 3) {
 					ClippedVert* vt0 = &clipped_buffer.vertices[0], *vt1, *vt2;
 					for (int i = 1; i <= clipped_buffer.count - 2; ++i) {
 						vt1 = &clipped_buffer.vertices[i];
 						vt2 = &clipped_buffer.vertices[i + 1];
 						h0 = &vt0->clip_coord; h1 = &vt1->clip_coord; h2 = &vt2->clip_coord;
-						v0 = &vt0->vertex; v1 = &vt1->vertex; v2 = &vt2->vertex;
-						ssrR_triangle(h0, h1, h2, v0, v1, v2, FALSE, vt0, vt1, vt2, state.program, &state.uniforms);
+						v0 = &vt0->vertex;     v1 = &vt1->vertex;     v2 = &vt2->vertex;
+						ssrR_triangle(h0, h1, h2, v0, v1, v2, state.program, &state.uniforms);
 					}
 				}
 			}
-
 		}
 	} else if (primitive == PRIMITIVE_POINT) {
 
diff --git a/src/core/device.h b/src/core/device.h
index b265471..d8daf76 100644
--- a/src/core/device.h
+++ b/src/core/device.h
@@ -8,30 +8,47 @@
 #include "shader.h"
 #include "rasterizer.h"
 #include "vert.h"
+#include "framebuffer.h"
 
-typedef enum ssr_MatrixMode {
+typedef enum {
 	MATRIX_MODEL = 0,
 	MATRIX_VIEW = 1,
 	MATRIX_PROJECTION = 2,
 } ssr_MatrixMode;
 
-typedef struct ssr_Config {
+typedef struct {
 	int width, height;
 	bool dbuffer;/* double buffer? */
 	Color* target; /* screen target buffer */
 } ssr_Config;
 
-typedef enum ssr_PrimitiveType {
+typedef enum {
 	PRIMITIVE_POINT,
 	PRIMITIVE_LINE,
 	PRIMITIVE_TRIANGLE,
 } ssr_PrimitiveType;
 
-typedef enum ssr_EnableMask {
-	ENABLEMASK_BACKFACECULL = 1 , 
-	ENABLEMASK_DEPTHTEST = 1 << 1,
+typedef enum {
+	ENABLEMASK_BACKFACECULL = 1,
+	ENABLEMASK_DEPTHTEST    = 1 << 1,
+	ENABLEMASK_MULTISAMPLE  = 1 << 2,
+	ENABLEMASK_BLEND        = 1 << 3,
+	ENABLEMASK_WRITEDEPTH   = 1 << 4,
 } ssr_EnableMask;
 
+typedef enum {
+	BLENDFACTOR_ONE,
+	BLENDFACTOR_ZERO,
+	BLENDFACTOR_SRC_COLOR, 
+	BLENDFACTOR_ONE_MINUS_SRC_COLOR, 
+	BLENDFACTOR_DST_COLOR, 
+	BLENDFACTOR_ONE_MINUS_DST_COLOR,
+	BLENDFACTOR_SRC_ALPHA,
+	BLENDFACTOR_ONE_MINUS_SRC_ALPHA,
+	BLENDFACTOR_DST_ALPHA,
+	BLENDFACTOR_ONE_MINUS_DST_ALPHA,
+} ssr_BlendFactor;
+
 void ssr_init(ssr_Config* config);
 
 float ssr_getaspect();
@@ -61,6 +78,7 @@ void ssr_ortho(float l, float r, float b, float t, float n, float f);
 
 void ssr_viewport(float l, float r, float b, float t);
 
+void ssr_getm(Mat4* out);
 void ssr_getmv(Mat4* out);
 void ssr_getmvp(Mat4* out);
 
@@ -86,19 +104,28 @@ void ssr_clearcolor(Color color);
 void ssr_cleardepth();
 
 void ssr_putpoint(uint screenx, uint screeny, Color color); /*直接修改某个位置的颜色*/
+void ssr_putpoint32(uint screenx, uint screeny, Color32* c32);
 
 void ssr_present();
 
 Color ssr_color(unsigned char r, unsigned char g, unsigned char b, unsigned char a);
 
-//bool ssr_cutline(int* x0, int* y0, int* x1, int* y1);
+bool ssr_testdepth(uint x, uint y, float depth);
+void ssr_writedepth(uint x, uint y, float depth);
 
-bool ssr_testdepth(uint x, uint y, uint depth); /*尝试写入深度，如果可绘制返回true，否则discard*/
-bool ssr_testdepthf(uint x, uint y, float depth01);
+void ssr_blend(Color32* src, Color32* dst, Color32* out);
 
 void ssrU_viewport(Vec2* p, Vec2* out);
+void ssrU_blend(ssr_BlendFactor sfactor, ssr_BlendFactor dfactor, Color32* src, Color32* dst, Color* out);
+
+Color32 ssr_getfbocolor(uint x, uint y);
 
 FilterMode ssr_getfiltermode();
 WrapMode ssr_getwrapmode();
 
+void ssr_bindframebuffer(FrameBuffer* fbo);
+void ssr_unbindframebuffer();
+
+void ssr_setblendfunc(ssr_BlendFactor sfactor, ssr_BlendFactor dfactor);
+
 #endif
\ No newline at end of file
diff --git a/src/core/framebuffer.c b/src/core/framebuffer.c
new file mode 100644
index 0000000..e69de29
--- /dev/null
+++ b/src/core/framebuffer.c
diff --git a/src/core/framebuffer.h b/src/core/framebuffer.h
new file mode 100644
index 0000000..3f03200
--- /dev/null
+++ b/src/core/framebuffer.h
@@ -0,0 +1,16 @@
+#ifndef _SOFTSHADEROOM_FRAMEBUFFER_H_
+#define _SOFTSHADEROOM_FRAMEBUFFER_H_
+
+#include "texture.h"
+
+/*frame buffer object*/
+
+typedef Texture RenderTexture;
+
+typedef struct FrameBuffer{
+	RenderTexture rt;
+} FrameBuffer;
+
+FrameBuffer* fbo_create(uint height, uint width);
+
+#endif
\ No newline at end of file
diff --git a/src/core/mem.h b/src/core/mem.h
index 70e8011..b51661b 100644
--- a/src/core/mem.h
+++ b/src/core/mem.h
@@ -3,9 +3,6 @@
 
 #include <memory.h>
 
-void ssrM_freemem();
-void ssrM_realloc();
-
 #define ssrM_copy memcpy
 
 #define ssrM_zero(obj, size) memset(obj, 0, size)
@@ -13,11 +10,10 @@ void ssrM_realloc();
 #define ssrM_newvector(T, c) ((T*)calloc(c, sizeof(T)))
 #define ssrM_free(obj) free(obj)
 
-/*扩容一倍*/
-#define ssrM_scalevector(T, v, curlen) \
+#define ssrM_scalevector(T, v, curlen)  \
 do{																		  \
 	ssr_assert(v);												\
-	curlen = curlen << 1;									\
+	curlen <<= 1;									        \
 	v = (T*)realloc(v, sizeof(T)*curlen);	\
 }while(0)
 
diff --git a/src/core/rasterizer.c b/src/core/rasterizer.c
index bf210b7..772eb29 100644
--- a/src/core/rasterizer.c
+++ b/src/core/rasterizer.c
@@ -84,18 +84,11 @@ static void puttriangle(Vec2* A, Vec2* B, Vec2* C, Color c) {
 
 extern FragmentShaderIn ssr_frag_in;
 
-void ssrR_triangle(
-	Vec4* CA, Vec4* CB, Vec4* CC, 
-	Vert* A, Vert* B, Vert* C, 
-	bool from_reg, 
-	ClippedVert* CVA, ClippedVert* CVB, ClippedVert* CVC, 
-	Program* program, 
-	UniformCollection* uniforms
-) {
+void ssrR_triangle( Vec4* CA, Vec4* CB, Vec4* CC, Vert* A, Vert* B, Vert* C, Program* program, UniformCollection* uniforms ) {
 	ssr_assert(CA && CB && CC && program);
 
 	static Vec3 SA, SB, SC;
-	static Vec3 bcp;
+	static Vec3 bc;
 
 	vec4_dividew(CA, &SA); ssrU_viewport(&SA, &SA);
 	vec4_dividew(CB, &SB); ssrU_viewport(&SB, &SB);
@@ -104,11 +97,11 @@ void ssrR_triangle(
 	puttriangle(&SA, &SB, &SC, 0xffff0000);
 	return;
 */
-	Vec3 *sa = &SA, *sb = &SB, *sc = &SC, *tmp; Vec4* v4tmp; Vert* vtemp; ClippedVert* cvtmp;
+	Vec3 *sa = &SA, *sb = &SB, *sc = &SC, *tmp; Vec4* v4tmp; Vert* vtemp;
 #define swap(t, a, b) {t = a; a = b; b = t;} /*sort in y axis*/
-	if (sb->y < sa->y) { swap(tmp, sa, sb); swap(v4tmp, CA, CB); swap(vtemp, A, B); if (!from_reg) swap(cvtmp, CVA, CVB); }
-	if (sc->y < sb->y) { swap(tmp, sb, sc); swap(v4tmp, CB, CC); swap(vtemp, B, C); if (!from_reg) swap(cvtmp, CVC, CVB); }
-	if (sb->y < sa->y) { swap(tmp, sa, sb); swap(v4tmp, CA, CB); swap(vtemp, A, B); if (!from_reg) swap(cvtmp, CVA, CVB); }
+	if (sb->y < sa->y) { swap(tmp, sa, sb); swap(v4tmp, CA, CB); swap(vtemp, A, B); }
+	if (sc->y < sb->y) { swap(tmp, sb, sc); swap(v4tmp, CB, CC); swap(vtemp, B, C); }
+	if (sb->y < sa->y) { swap(tmp, sa, sb); swap(v4tmp, CA, CB); swap(vtemp, A, B); }
 #undef swap
 
 	Vec2 AB = {sb->x - sa->x, sb->y - sa->y}, AC = { sc->x - sa->x, sc->y - sa->y };
@@ -118,7 +111,7 @@ void ssrR_triangle(
 	float invkAB = (sb->x - sa->x) / (sb->y - sa->y + EPSILON);
 	float invkBC = (sc->x - sb->x) / (sc->y - sb->y + EPSILON);
 	
-	Color color;
+	Color32 color;
 	float from, to;
 	Vec2 p;
 	float depth;
@@ -126,115 +119,67 @@ void ssrR_triangle(
 	
 	uint varying_flag = program->varying_flag;
 
-#define DO_INTERPOLATION  \
-	if (varying_flag & VARYING_BASIC) {																																										 																		 \
-		if (varying_flag & VARYING_POSITION) ssrS_bcpvec3(&bcp, &A->position, &B->position, &C->position, &ssr_frag_in.position); 															 \
-		if (varying_flag & VARYING_NORMAL) ssrS_bcpvec3(&bcp, &A->normal, &B->normal, &C->normal, &ssr_frag_in.normal);					 																 \
-		if (varying_flag & VARYING_TANGENT) ssrS_bcpvec3(&bcp, &A->tangent, &B->tangent, &C->tangent, &ssr_frag_in.tangent);			 															 \
-		if (varying_flag & VARYING_TEXCOORD) ssrS_bcpvec2(&bcp, &A->texcoord, &B->texcoord, &C->texcoord, &ssr_frag_in.texcoord); 															 \
-		if (varying_flag & VARYING_JOINT) ssrS_bcpvec4(&bcp, &A->joint, &B->joint, &C->joint, &ssr_frag_in.joint);								 															 \
-		if (varying_flag & VARYING_WEIGHT) ssrS_bcpvec4(&bcp, &A->weight, &B->weight, &C->weight, &ssr_frag_in.weight);					 																 \
-		if (varying_flag & VARYING_COLOR) ssrS_bcpcolor(&bcp, A->color, B->color, C->color, &ssr_frag_in.color);									 															 \
-	}																																																																													 \
-	if (from_reg) { /*extra varying*/																																																													 \
-		uint i_a = A->index, i_b = B->index, i_c = C->index;																																																		 \
-		if (varying_flag & VARYING_NUM) {																																																											 \
-			if (varying_flag & VARYING_NUM_00) reg_num_00 = ssrS_bcpnum(&bcp, reg_num[0].num[i_a], reg_num[0].num[i_b], reg_num[0].num[i_c], &ssr_frag_in.num[0]); \
-			if (varying_flag & VARYING_NUM_01) reg_num_01 = ssrS_bcpnum(&bcp, reg_num[1].num[i_a], reg_num[1].num[i_b], reg_num[1].num[i_c], &ssr_frag_in.num[1]); \
-			if (varying_flag & VARYING_NUM_02) reg_num_02 = ssrS_bcpnum(&bcp, reg_num[2].num[i_a], reg_num[2].num[i_b], reg_num[2].num[i_c], &ssr_frag_in.num[2]); \
-			if (varying_flag & VARYING_NUM_03) reg_num_03 = ssrS_bcpnum(&bcp, reg_num[3].num[i_a], reg_num[3].num[i_b], reg_num[3].num[i_c], &ssr_frag_in.num[3]); \
-		}																																																																												 \
-		if (varying_flag & VARYING_V2) {																																																												 \
-			if (varying_flag & VARYING_V2_00) reg_v2_00 = ssrS_bcpvec2(&bcp, &reg_v2[0].v2[i_a], &reg_v2[0].v2[i_b], &reg_v2[0].v2[i_c], &ssr_frag_in.v2[0]);			 \
-			if (varying_flag & VARYING_V2_01) reg_v2_01 = ssrS_bcpvec2(&bcp, &reg_v2[1].v2[i_a], &reg_v2[1].v2[i_b], &reg_v2[1].v2[i_c], &ssr_frag_in.v2[1]);			 \
-			if (varying_flag & VARYING_V2_02) reg_v2_02 = ssrS_bcpvec2(&bcp, &reg_v2[2].v2[i_a], &reg_v2[2].v2[i_b], &reg_v2[2].v2[i_c], &ssr_frag_in.v2[2]);			 \
-			if (varying_flag & VARYING_V2_03) reg_v2_03 = ssrS_bcpvec2(&bcp, &reg_v2[3].v2[i_a], &reg_v2[3].v2[i_b], &reg_v2[3].v2[i_c], &ssr_frag_in.v2[3]);			 \
-			if (varying_flag & VARYING_V2_04) reg_v2_04 = ssrS_bcpvec2(&bcp, &reg_v2[4].v2[i_a], &reg_v2[4].v2[i_b], &reg_v2[4].v2[i_c], &ssr_frag_in.v2[4]);			 \
-			if (varying_flag & VARYING_V2_05) reg_v2_05 = ssrS_bcpvec2(&bcp, &reg_v2[5].v2[i_a], &reg_v2[5].v2[i_b], &reg_v2[5].v2[i_c], &ssr_frag_in.v2[5]);			 \
-		}																																																																												 \
-		if (varying_flag & VARYING_V3) {																																																												 \
-			if (varying_flag & VARYING_V3_00) reg_v3_00 = ssrS_bcpvec3(&bcp, &reg_v3[0].v3[i_a], &reg_v3[0].v3[i_b], &reg_v3[0].v3[i_c], &ssr_frag_in.v3[0]);			 \
-			if (varying_flag & VARYING_V3_01) reg_v3_01 = ssrS_bcpvec3(&bcp, &reg_v3[1].v3[i_a], &reg_v3[1].v3[i_b], &reg_v3[1].v3[i_c], &ssr_frag_in.v3[1]);			 \
-			if (varying_flag & VARYING_V3_02) reg_v3_02 = ssrS_bcpvec3(&bcp, &reg_v3[2].v3[i_a], &reg_v3[2].v3[i_b], &reg_v3[2].v3[i_c], &ssr_frag_in.v3[2]);			 \
-			if (varying_flag & VARYING_V3_03) reg_v3_03 = ssrS_bcpvec3(&bcp, &reg_v3[3].v3[i_a], &reg_v3[3].v3[i_b], &reg_v3[3].v3[i_c], &ssr_frag_in.v3[3]);			 \
-			if (varying_flag & VARYING_V3_04) reg_v3_04 = ssrS_bcpvec3(&bcp, &reg_v3[4].v3[i_a], &reg_v3[4].v3[i_b], &reg_v3[4].v3[i_c], &ssr_frag_in.v3[4]);			 \
-			if (varying_flag & VARYING_V3_05) reg_v3_05 = ssrS_bcpvec3(&bcp, &reg_v3[5].v3[i_a], &reg_v3[5].v3[i_b], &reg_v3[5].v3[i_c], &ssr_frag_in.v3[5]);			 \
-			if (varying_flag & VARYING_V3_06) reg_v3_06 = ssrS_bcpvec3(&bcp, &reg_v3[6].v3[i_a], &reg_v3[6].v3[i_b], &reg_v3[6].v3[i_c], &ssr_frag_in.v3[6]);			 \
-			if (varying_flag & VARYING_V3_07) reg_v3_07 = ssrS_bcpvec3(&bcp, &reg_v3[7].v3[i_a], &reg_v3[7].v3[i_b], &reg_v3[7].v3[i_c], &ssr_frag_in.v3[7]);			 \
-		}																																																																												 \
-		if (varying_flag & VARYING_V4) {																																																												 \
-			if (varying_flag & VARYING_V4_00) reg_v4_00 = ssrS_bcpvec4(&bcp, &reg_v4[0].v4[i_a], &reg_v4[0].v4[i_b], &reg_v4[0].v4[i_c], &ssr_frag_in.v4[0]);			 \
-			if (varying_flag & VARYING_V4_01) reg_v4_01 = ssrS_bcpvec4(&bcp, &reg_v4[1].v4[i_a], &reg_v4[1].v4[i_b], &reg_v4[1].v4[i_c], &ssr_frag_in.v4[1]);			 \
-			if (varying_flag & VARYING_V4_02) reg_v4_02 = ssrS_bcpvec4(&bcp, &reg_v4[2].v4[i_a], &reg_v4[2].v4[i_b], &reg_v4[2].v4[i_c], &ssr_frag_in.v4[2]);			 \
-			if (varying_flag & VARYING_V4_03) reg_v4_03 = ssrS_bcpvec4(&bcp, &reg_v4[3].v4[i_a], &reg_v4[3].v4[i_b], &reg_v4[3].v4[i_c], &ssr_frag_in.v4[3]);			 \
-			if (varying_flag & VARYING_V4_04) reg_v4_04 = ssrS_bcpvec4(&bcp, &reg_v4[4].v4[i_a], &reg_v4[4].v4[i_b], &reg_v4[4].v4[i_c], &ssr_frag_in.v4[4]);			 \
-			if (varying_flag & VARYING_V4_05) reg_v4_05 = ssrS_bcpvec4(&bcp, &reg_v4[5].v4[i_a], &reg_v4[5].v4[i_b], &reg_v4[5].v4[i_c], &ssr_frag_in.v4[5]);			 \
-		}																																																																												 \
-	} else {																																																																									 \
-		if (varying_flag & VARYING_NUM) {																																																											 \
-			if (varying_flag & VARYING_NUM_00) reg_num_00 = ssrS_bcpnum(&bcp, CVA->num[0], CVB->num[0], CVC->num[0], &ssr_frag_in.num[0]);												 \
-			if (varying_flag & VARYING_NUM_01) reg_num_01 = ssrS_bcpnum(&bcp, CVA->num[1], CVB->num[1], CVC->num[1], &ssr_frag_in.num[1]);												 \
-			if (varying_flag & VARYING_NUM_02) reg_num_02 = ssrS_bcpnum(&bcp, CVA->num[2], CVB->num[2], CVC->num[2], &ssr_frag_in.num[2]);												 \
-			if (varying_flag & VARYING_NUM_03) reg_num_03 = ssrS_bcpnum(&bcp, CVA->num[3], CVB->num[3], CVC->num[3], &ssr_frag_in.num[3]);												 \
-		}																																																																												 \
-		if (varying_flag & VARYING_V2) {																																																												 \
-			if (varying_flag & VARYING_V2_00) reg_v2_00 = ssrS_bcpvec2(&bcp, &CVA->v2[0], &CVB->v2[0], &CVC->v2[0], &ssr_frag_in.v2[0]);													 \
-			if (varying_flag & VARYING_V2_01) reg_v2_01 = ssrS_bcpvec2(&bcp, &CVA->v2[1], &CVB->v2[1], &CVC->v2[1], &ssr_frag_in.v2[1]);													 \
-			if (varying_flag & VARYING_V2_02) reg_v2_02 = ssrS_bcpvec2(&bcp, &CVA->v2[2], &CVB->v2[2], &CVC->v2[2], &ssr_frag_in.v2[2]);													 \
-			if (varying_flag & VARYING_V2_03) reg_v2_03 = ssrS_bcpvec2(&bcp, &CVA->v2[3], &CVB->v2[3], &CVC->v2[3], &ssr_frag_in.v2[3]);													 \
-			if (varying_flag & VARYING_V2_04) reg_v2_04 = ssrS_bcpvec2(&bcp, &CVA->v2[4], &CVB->v2[4], &CVC->v2[4], &ssr_frag_in.v2[4]);													 \
-			if (varying_flag & VARYING_V2_05) reg_v2_05 = ssrS_bcpvec2(&bcp, &CVA->v2[5], &CVB->v2[5], &CVC->v2[5], &ssr_frag_in.v2[5]);													 \
-		}																																																																												 \
-		if (varying_flag & VARYING_V3) {																																																												 \
-			if (varying_flag & VARYING_V3_00) reg_v3_00 = ssrS_bcpvec3(&bcp, &CVA->v3[0], &CVB->v3[0], &CVC->v3[0], &ssr_frag_in.v3[0]);													 \
-			if (varying_flag & VARYING_V3_01) reg_v3_01 = ssrS_bcpvec3(&bcp, &CVA->v3[1], &CVB->v3[1], &CVC->v3[1], &ssr_frag_in.v3[1]);													 \
-			if (varying_flag & VARYING_V3_02) reg_v3_02 = ssrS_bcpvec3(&bcp, &CVA->v3[2], &CVB->v3[2], &CVC->v3[2], &ssr_frag_in.v3[2]);													 \
-			if (varying_flag & VARYING_V3_03) reg_v3_03 = ssrS_bcpvec3(&bcp, &CVA->v3[3], &CVB->v3[3], &CVC->v3[3], &ssr_frag_in.v3[3]);													 \
-			if (varying_flag & VARYING_V3_04) reg_v3_04 = ssrS_bcpvec3(&bcp, &CVA->v3[4], &CVB->v3[4], &CVC->v3[4], &ssr_frag_in.v3[4]);													 \
-			if (varying_flag & VARYING_V3_05) reg_v3_05 = ssrS_bcpvec3(&bcp, &CVA->v3[5], &CVB->v3[5], &CVC->v3[5], &ssr_frag_in.v3[5]);													 \
-			if (varying_flag & VARYING_V3_06) reg_v3_06 = ssrS_bcpvec3(&bcp, &CVA->v3[6], &CVB->v3[6], &CVC->v3[6], &ssr_frag_in.v3[6]);													 \
-			if (varying_flag & VARYING_V3_07) reg_v3_07 = ssrS_bcpvec3(&bcp, &CVA->v3[7], &CVB->v3[7], &CVC->v3[7], &ssr_frag_in.v3[7]);													 \
-		}																																																																												 \
-		if (varying_flag & VARYING_V4) {																																																												 \
-			if (varying_flag & VARYING_V4_00) reg_v4_00 = ssrS_bcpvec4(&bcp, &CVA->v4[0], &CVB->v4[0], &CVC->v4[0], &ssr_frag_in.v4[0]);													 \
-			if (varying_flag & VARYING_V4_01) reg_v4_01 = ssrS_bcpvec4(&bcp, &CVA->v4[1], &CVB->v4[1], &CVC->v4[1], &ssr_frag_in.v4[1]);													 \
-			if (varying_flag & VARYING_V4_02) reg_v4_02 = ssrS_bcpvec4(&bcp, &CVA->v4[2], &CVB->v4[2], &CVC->v4[2], &ssr_frag_in.v4[2]);													 \
-			if (varying_flag & VARYING_V4_03) reg_v4_03 = ssrS_bcpvec4(&bcp, &CVA->v4[3], &CVB->v4[3], &CVC->v4[3], &ssr_frag_in.v4[3]);													 \
-			if (varying_flag & VARYING_V4_04) reg_v4_04 = ssrS_bcpvec4(&bcp, &CVA->v4[4], &CVB->v4[4], &CVC->v4[4], &ssr_frag_in.v4[4]);													 \
-			if (varying_flag & VARYING_V4_05) reg_v4_05 = ssrS_bcpvec4(&bcp, &CVA->v4[5], &CVB->v4[5], &CVC->v4[5], &ssr_frag_in.v4[5]);													 \
-		}																																																																												 \
-	}
-
 	Vec3 s[2], u;
 	s[0].x = sc->x - sa->x; s[0].y = sb->x - sa->x;
 	s[1].x = sc->y - sa->y; s[1].y = sb->y - sa->y;
 
+	bool depth_test   = ssr_isenable(ENABLEMASK_DEPTHTEST);
+	bool multi_sample = ssr_isenable(ENABLEMASK_MULTISAMPLE);
+	bool blend        = ssr_isenable(ENABLEMASK_BLEND);
+	bool write_depth  = ssr_isenable(ENABLEMASK_WRITEDEPTH);
+
+#define DO_INTERPOLATION  \
+	if (varying_flag & VARYING_BASIC) {																																										 		 \
+		if (varying_flag & VARYING_POSITION) ssrS_bcpvec3(&bc, &A->position, &B->position, &C->position, &ssr_frag_in.position); \
+		if (varying_flag & VARYING_NORMAL) ssrS_bcpvec3(&bc, &A->normal, &B->normal, &C->normal, &ssr_frag_in.normal);					 \
+		if (varying_flag & VARYING_TANGENT) ssrS_bcpvec3(&bc, &A->tangent, &B->tangent, &C->tangent, &ssr_frag_in.tangent);			 \
+		if (varying_flag & VARYING_TEXCOORD) ssrS_bcpvec2(&bc, &A->texcoord, &B->texcoord, &C->texcoord, &ssr_frag_in.texcoord); \
+/*		if (varying_flag & VARYING_JOINT) ssrS_bcpvec4(&bc, &A->joint, &B->joint, &C->joint, &ssr_frag_in.joint);								 \
+		if (varying_flag & VARYING_WEIGHT) ssrS_bcpvec4(&bc, &A->weight, &B->weight, &C->weight, &ssr_frag_in.weight);					 \
+	*/	if (varying_flag & VARYING_COLOR) ssrS_bcpcolor(&bc, A->color, B->color, C->color, &ssr_frag_in.color);									 \
+	}	
+
 #define RENDER_TRIANGLE \
-	for (p.y = FROM->y; p.y < TO->y + OFFSET; ++p.y) {							 \
-		SET_FROM_AND_TO	                                               \
-		s[1].z = sa->y - p.y;																			     \
-		for (p.x = from; order * (p.x - to) <= 0; p.x += order) {	     \
-			s[0].z = sa->x - p.x;																		     \
-			vec3_cross(&s[0], &s[1], &u);													       \
-			discardif(compare(u.z, 0));														       \
-			u.z = 1.f / u.z;																			       \
-			bcp.x = 1.f - (u.x + u.y) * u.z;											       \
-			bcp.y = u.y * u.z;																		       \
-			bcp.z = u.x * u.z;																		       \
-			discardif(bcp.x < 0 || bcp.y < 0 || bcp.z < 0);				       \
-			/*perspective correction*/																	 \
-			bcp.x *= CAw; bcp.y *= CBw; bcp.z *= CCw;							       \
-			vec3_scale(&bcp, 1.f / (bcp.x + bcp.y + bcp.z), &bcp);       \
-			/*depth test*/																					     \
-			if(ssr_isenable(ENABLEMASK_DEPTHTEST)){											 \
-				depth = (bcp.x*sa->z+bcp.y*sb->z+bcp.z*sc->z) * 0.5f + 0.5f;\
-				discardif(!ssr_testdepthf(p.x, p.y, depth));						    \
-			}																														 \
-			/*set varying variables*/																		 \
-			DO_INTERPOLATION																						 \
-			if(program->fragmentshader(uniforms, &ssr_frag_in, &color))	 \
-				ssr_putpoint(p.x, p.y, color);													   \
-		}																													     \
+	for (p.y = FROM->y; p.y < TO->y + OFFSET; ++p.y) {							   \
+		SET_FROM_AND_TO	                                                 \
+		s[1].z = sa->y - p.y;																			       \
+		for (p.x = from; order * (p.x - to) <= 0; p.x += order) {	       \
+			s[0].z = sa->x - p.x;																		       \
+			vec3_cross(&s[0], &s[1], &u);													         \
+			discardif(compare(u.z, 0));														         \
+			u.z = 1.f / u.z;																			         \
+			bc.x = 1.f - (u.x + u.y) * u.z;											           \
+			bc.y = u.y * u.z;																		           \
+			bc.z = u.x * u.z;																		           \
+			discardif(bc.x < 0 || bc.y < 0 || bc.z < 0);				           \
+			/*perspective correction*/																	   \
+			bc.x *= CAw; bc.y *= CBw; bc.z *= CCw;							           \
+			vec3_scale(&bc, 1.f / (bc.x + bc.y + bc.z), &bc);              \
+			/*early depth testing*/																			   \
+			if(depth_test ){											                         \
+				depth = bc.x*sa->z+bc.y*sb->z+bc.z*sc->z;                    \
+				discardif(!ssr_testdepth(p.x, p.y, depth));						       \
+			}																														   \
+			/*set varying variables*/																		   \
+			DO_INTERPOLATION																							 \
+			ssrS_solveregs(&bc, A->index, B->index, C->index);             \
+			/*enter fragment shader*/																			 \
+			if(program->fragmentshader(uniforms, &ssr_frag_in, &color))	{	 \
+				/*MSAA*/																										 \
+				if(multi_sample) {																					 \
+																																		 \
+				}																														 \
+				/*blend*/																										 \
+				if(blend) {																									 \
+					Color32 dst; dst = ssr_getfbocolor(p.x, p.y);							 \
+					ssr_blend(&color, &dst, &color);													 \
+				}																														 \
+				if(write_depth) {																						 \
+					ssr_writedepth(p.x, p.y, depth);													 \
+				}																														 \
+				ssr_putpoint32(p.x, p.y, &color);														 \
+			}																															 \
+		}																													       \
 	}
 
 #define FROM    sa 
diff --git a/src/core/rasterizer.h b/src/core/rasterizer.h
index eceb8c7..8e35511 100644
--- a/src/core/rasterizer.h
+++ b/src/core/rasterizer.h
@@ -13,11 +13,7 @@ bool ssrR_barycentric(Vec2* A, Vec2* B, Vec2* C, Vec2* p, Vec3* out); /*计算重心
 void ssrR_center(Vec2* A, Vec2* B, Vec2* C, Vec2* out); /*获得重心*/
 bool ssrR_ispointintriangle(Vec2* A, Vec2* B, Vec2* C, Vec2* p);
 
-void ssrR_triangle(Vec4* CA, Vec4* CB, Vec4* CC, 
-	Vert* A, Vert* B, Vert* C, 
-	bool from_reg, 
-	ClippedVert* CVA, ClippedVert* CVB, ClippedVert* CVC, 
-	Program* program, UniformCollection* uniforms
+void ssrR_triangle(Vec4* CA, Vec4* CB, Vec4* CC, Vert* A, Vert* B, Vert* C, Program* program, UniformCollection* uniforms
 );
 
 #endif
\ No newline at end of file
diff --git a/src/core/shader.c b/src/core/shader.c
index a508820..0490362 100644
--- a/src/core/shader.c
+++ b/src/core/shader.c
@@ -31,199 +31,182 @@ Vec4* reg_v4_03;
 Vec4* reg_v4_04;
 Vec4* reg_v4_05;
 
-Register reg_num[4] = {
-	{ 0, REGTYPE_NUM, NULL },
-	{ 0, REGTYPE_NUM, NULL },
-	{ 0, REGTYPE_NUM, NULL },
-	{ 0, REGTYPE_NUM, NULL },
+extern FragmentShaderIn ssr_frag_in;
+
+Register registers[REG_TOTAL] = {
+	/*4 float registers*/
+	{ 0, REGTYPE_NUM, sizeof(float), NULL },
+	{ 0, REGTYPE_NUM, sizeof(float), NULL },
+	{ 0, REGTYPE_NUM, sizeof(float), NULL },
+	{ 0, REGTYPE_NUM, sizeof(float), NULL },
+	/*6 vec2 registers*/								  
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	{ 0, REGTYPE_VEC2, sizeof(Vec2), NULL },
+	/*8 vec3 registers*/								  
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	{ 0, REGTYPE_VEC3, sizeof(Vec3), NULL },
+	/*6 vec4 registers*/								  
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL },
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL },
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL },
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL },
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL },
+	{ 0, REGTYPE_VEC4, sizeof(Vec4), NULL }
 };
 
-Register reg_v2[6] = {
-	{ 0, REGTYPE_VEC2, NULL },
-	{ 0, REGTYPE_VEC2, NULL },
-	{ 0, REGTYPE_VEC2, NULL },
-	{ 0, REGTYPE_VEC2, NULL },
-	{ 0, REGTYPE_VEC2, NULL },
-	{ 0, REGTYPE_VEC2, NULL },
+ActiveReg active_regs[REG_TOTAL] = {
+	/*4 float registers*/
+	{ sizeof(float), NULL, &reg_num_00, ssrS_bcpnum, &ssr_frag_in.num[0] },
+	{ sizeof(float), NULL, &reg_num_01, ssrS_bcpnum, &ssr_frag_in.num[1] },
+	{ sizeof(float), NULL, &reg_num_02, ssrS_bcpnum, &ssr_frag_in.num[2] },
+	{ sizeof(float), NULL, &reg_num_03, ssrS_bcpnum, &ssr_frag_in.num[3] },
+	/*6 vec2 registers*/
+	{ sizeof(Vec2), NULL, &reg_v2_00, ssrS_bcpvec2, &ssr_frag_in.v2[0] },
+	{ sizeof(Vec2), NULL, &reg_v2_01, ssrS_bcpvec2, &ssr_frag_in.v2[1] },
+	{ sizeof(Vec2), NULL, &reg_v2_02, ssrS_bcpvec2, &ssr_frag_in.v2[2] },
+	{ sizeof(Vec2), NULL, &reg_v2_03, ssrS_bcpvec2, &ssr_frag_in.v2[3] },
+	{ sizeof(Vec2), NULL, &reg_v2_04, ssrS_bcpvec2, &ssr_frag_in.v2[4] },
+	{ sizeof(Vec2), NULL, &reg_v2_05, ssrS_bcpvec2, &ssr_frag_in.v2[5] },
+	/*8 vec3 registers*/																					  
+	{ sizeof(Vec3), NULL, &reg_v3_00, ssrS_bcpvec3, &ssr_frag_in.v3[0] },
+	{ sizeof(Vec3), NULL, &reg_v3_01, ssrS_bcpvec3, &ssr_frag_in.v3[1] },
+	{ sizeof(Vec3), NULL, &reg_v3_02, ssrS_bcpvec3, &ssr_frag_in.v3[2] },
+	{ sizeof(Vec3), NULL, &reg_v3_03, ssrS_bcpvec3, &ssr_frag_in.v3[3] },
+	{ sizeof(Vec3), NULL, &reg_v3_04, ssrS_bcpvec3, &ssr_frag_in.v3[4] },
+	{ sizeof(Vec3), NULL, &reg_v3_05, ssrS_bcpvec3, &ssr_frag_in.v3[5] },
+	{ sizeof(Vec3), NULL, &reg_v3_06, ssrS_bcpvec3, &ssr_frag_in.v3[6] },
+	{ sizeof(Vec3), NULL, &reg_v3_07, ssrS_bcpvec3, &ssr_frag_in.v3[7] },
+	/*6 vec4 registers*/																					  
+	{ sizeof(Vec4), NULL, &reg_v4_00, ssrS_bcpvec4, &ssr_frag_in.v4[0] },
+	{ sizeof(Vec4), NULL, &reg_v4_01, ssrS_bcpvec4, &ssr_frag_in.v4[1] },
+	{ sizeof(Vec4), NULL, &reg_v4_02, ssrS_bcpvec4, &ssr_frag_in.v4[2] },
+	{ sizeof(Vec4), NULL, &reg_v4_03, ssrS_bcpvec4, &ssr_frag_in.v4[3] },
+	{ sizeof(Vec4), NULL, &reg_v4_04, ssrS_bcpvec4, &ssr_frag_in.v4[4] },
+	{ sizeof(Vec4), NULL, &reg_v4_05, ssrS_bcpvec4, &ssr_frag_in.v4[5] }
 };
 
-Register reg_v3[8] = {
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-	{ 0, REGTYPE_VEC3, NULL },
-};
-
-Register reg_v4[6] = {
-	{ 0, REGTYPE_VEC4, NULL },
-	{ 0, REGTYPE_VEC4, NULL },
-	{ 0, REGTYPE_VEC4, NULL },
-	{ 0, REGTYPE_VEC4, NULL },
-	{ 0, REGTYPE_VEC4, NULL },
-	{ 0, REGTYPE_VEC4, NULL },
-};
-
-void ssrS_setregisters(uint flag, int capacity) {
-#define reg_scale_num(i) \
-if (flag & VARYING_NUM_##i) ssrM_rescalevector(float, reg_num[##i].num, reg_num[##i].length, capacity, FALSE)
-
-#define reg_scale_v2(i) \
-if (flag & VARYING_V2_##i) ssrM_rescalevector(Vec2, reg_v2[##i].v2, reg_v2[##i].length, capacity, FALSE)
-
-#define reg_scale_v3(i) \
-if (flag & VARYING_V3_##i) ssrM_rescalevector(Vec3, reg_v3[##i].v3, reg_v3[##i].length, capacity, FALSE)
-
-#define reg_scale_v4(i) \
-if (flag & VARYING_V4_##i) ssrM_rescalevector(Vec4, reg_v4[##i].v4, reg_v4[##i].length, capacity, FALSE)
-
-	if (!(flag & VARYING_EXTRA))
+/*设置这个draw call开启的寄存器*/
+void ssrS_openregs(uint varying_flag) {
+	if (varying_flag & VARYING_EXTRA == 0)
+	{
+		open_regsi[0] = -1;
 		return;
-
-	if (flag & VARYING_NUM) {
-		reg_scale_num(00);
-		reg_scale_num(01);
-		reg_scale_num(02);
-		reg_scale_num(03);
-	}
-	if (flag & VARYING_V2) {
-		reg_scale_v2(00);
-		reg_scale_v2(01);
-		reg_scale_v2(02);
-		reg_scale_v2(03);
-		reg_scale_v2(04);
-		reg_scale_v2(05);
 	}
-	if (flag & VARYING_V3) {
-		reg_scale_v3(00);
-		reg_scale_v3(01);
-		reg_scale_v3(02);
-		reg_scale_v3(03);
-		reg_scale_v3(04);
-		reg_scale_v3(05);
-		reg_scale_v3(06);
-		reg_scale_v3(07);
-	}
-	if (flag & VARYING_V4) {
-		reg_scale_v4(00);
-		reg_scale_v4(01);
-		reg_scale_v4(02);
-		reg_scale_v4(03);
-		reg_scale_v4(04);
-		reg_scale_v4(05);
+	int j = 0;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		if (varying_flag & VARYING_EXTRA == 0)
+			break;
+		if (varying_flag & (1 << i)) {
+			open_regsi[j++] = i;
+			varying_flag &= ~(1 << i);
+		}
 	}
+	if(j < REG_TOTAL) 
+		open_regsi[j] = -1;
 }
 
-void ssrS_setupregisterpoints(uint extra_varying_flag, int idx) {
-#define set_reg_num_pointer(i)\
-		if (extra_varying_flag & VARYING_NUM_##i) reg_num_##i = &reg_num[##i].num[idx];
-
-#define set_reg_v2_pointer(i)\
-		if (extra_varying_flag & VARYING_V2_##i) reg_v2_##i = &reg_v2[##i].v2[idx];
-
-#define set_reg_v3_pointer(i)\
-		if (extra_varying_flag & VARYING_V3_##i) reg_v3_##i = &reg_v3[##i].v3[idx];
-
-#define set_reg_v4_pointer(i)\
-		if (extra_varying_flag & VARYING_V4_##i) reg_v4_##i = &reg_v4[##i].v4[idx];
-
-	if (!(extra_varying_flag & VARYING_EXTRA))
-		return;
-	if (extra_varying_flag & VARYING_NUM) {
-		set_reg_num_pointer(00);
-		set_reg_num_pointer(01);
-		set_reg_num_pointer(02);
-		set_reg_num_pointer(03);
-	}
-	if (extra_varying_flag & VARYING_V2) {
-		set_reg_v2_pointer(00);
-		set_reg_v2_pointer(01);
-		set_reg_v2_pointer(02);
-		set_reg_v2_pointer(03);
-		set_reg_v2_pointer(04);
-		set_reg_v2_pointer(05);
-	}
-	if (extra_varying_flag & VARYING_V3) {
-		set_reg_v3_pointer(00);
-		set_reg_v3_pointer(01);
-		set_reg_v3_pointer(02);
-		set_reg_v3_pointer(03);
-		set_reg_v3_pointer(04);
-		set_reg_v3_pointer(05);
-		set_reg_v3_pointer(06);
-		set_reg_v3_pointer(07);
-	}
-	if (extra_varying_flag & VARYING_V4) {
-		set_reg_v4_pointer(00);
-		set_reg_v4_pointer(01);
-		set_reg_v4_pointer(02);
-		set_reg_v4_pointer(03);
-		set_reg_v4_pointer(04);
-		set_reg_v4_pointer(05);
+void ssrS_setactiveregr() { /*set active reg data from registers*/
+	int index = 0;
+	ActiveReg* reg; 
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break; 
+		reg = &active_regs[index]; 
+		reg->data = registers[index].data;
 	}
 }
 
-void ssrS_setregtofragin(uint extra_varying_flag, FragmentShaderIn* frag_in){
-#define set_reg_num_pointer_to_fragin(i)\
-		if (extra_varying_flag & VARYING_NUM_##i) reg_num_##i = &frag_in->num[##i];
+extern byte* clip_buffer_data[REG_TOTAL];
 
-#define set_reg_v2_pointer_to_fragin(i)\
-		if (extra_varying_flag & VARYING_V2_##i) reg_v2_##i = &frag_in->v2[##i];
-
-#define set_reg_v3_pointer_to_fragin(i)\
-		if (extra_varying_flag & VARYING_V3_##i) reg_v3_##i = &frag_in->v3[##i];
-
-#define set_reg_v4_pointer_to_fragin(i)\
-		if (extra_varying_flag & VARYING_V4_##i) reg_v4_##i = &frag_in->v4[##i];
-
-	if (!(extra_varying_flag & VARYING_EXTRA))
-		return;
-	if (extra_varying_flag & VARYING_NUM) {
-		set_reg_num_pointer_to_fragin(00);
-		set_reg_num_pointer_to_fragin(01);
-		set_reg_num_pointer_to_fragin(02);
-		set_reg_num_pointer_to_fragin(03);
+void ssrS_setactiveregc() { /*set active reg data from clipping buffer*/
+	int index = 0;
+	ActiveReg* reg;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break;
+		reg = &active_regs[index];
+		reg->data = clip_buffer_data[index];
 	}
-	if (extra_varying_flag & VARYING_V2) {
-		set_reg_v2_pointer_to_fragin(00);
-		set_reg_v2_pointer_to_fragin(01);
-		set_reg_v2_pointer_to_fragin(02);
-		set_reg_v2_pointer_to_fragin(03);
-		set_reg_v2_pointer_to_fragin(04);
-		set_reg_v2_pointer_to_fragin(05);
+}
+
+/*计算寄存器中的值并输出*/
+void ssrS_solveregs(Vec3* bc, uint a, uint b, uint c) {
+	int index = 0;
+	uint stride = 0;
+	ActiveReg* reg;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break;
+		reg = &active_regs[index];
+		stride = reg->element_size;
+		reg->accessor = reg->bcp_interpolator(
+			bc
+			, &reg->data[a*stride]
+			, &reg->data[b*stride]
+			, &reg->data[c*stride]
+			, reg->output
+		);
 	}
-	if (extra_varying_flag & VARYING_V3) {
-		set_reg_v3_pointer_to_fragin(00);
-		set_reg_v3_pointer_to_fragin(01);
-		set_reg_v3_pointer_to_fragin(02);
-		set_reg_v3_pointer_to_fragin(03);
-		set_reg_v3_pointer_to_fragin(04);
-		set_reg_v3_pointer_to_fragin(05);
-		set_reg_v3_pointer_to_fragin(06);
-		set_reg_v3_pointer_to_fragin(07);
+}
+
+/*给寄存器扩容（如果需要的话）*/
+void ssrS_setregisters(int capacity) {
+	Register* reg;
+	byte* data;
+	uint index;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break;
+		reg = &registers[index];
+		data = reg->data;
+		if (reg->length >= capacity)
+			continue; 
+		if (!data)
+			data = (byte*)calloc(capacity * reg->element_size, sizeof(byte));
+		else
+			data = (byte*)realloc(data, capacity * reg->element_size);
+		reg->data = data;
+		reg->length = capacity;
 	}
-	if (extra_varying_flag & VARYING_V4) {
-		set_reg_v4_pointer_to_fragin(00);
-		set_reg_v4_pointer_to_fragin(01);
-		set_reg_v4_pointer_to_fragin(02);
-		set_reg_v4_pointer_to_fragin(03);
-		set_reg_v4_pointer_to_fragin(04);
-		set_reg_v4_pointer_to_fragin(05);
+}
+
+/*进入vert shader前设置寄存器指针到对应顶点的数据在寄存器中的位置*/
+void ssrS_setupregisterpoints(int idx) {
+	ActiveReg* reg; 
+	uint index;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break;
+		reg = &active_regs[index];
+		*reg->accessor = &reg->data[idx * reg->element_size];
 	}
 }
 
-Color32 texture2d(Texture* tex, Vec2* uv) {
-	FilterMode filter_mode = ssr_getfiltermode();
-	WrapMode wrap_mode = ssr_getwrapmode();
-	return texture_sampling(tex, filter_mode, wrap_mode, uv->x, uv->y);
+/*进入frag shader前，将寄存器指针设置到fragment-in对应的位置*/
+void ssrS_setregtofragin() {
+	ActiveReg* reg;
+	uint index;
+	for (int i = 0; i < REG_TOTAL; ++i) {
+		index = open_regsi[i];
+		if (index == -1) break;
+		reg = &active_regs[index];
+		*reg->accessor = reg->output;
+	}
 }
 
-float* ssrS_bcpnum(Vec3* bc, float A, float B, float C, float* out) {
+float* ssrS_bcpnum(Vec3* bc, float* A, float* B, float* C, float* out) {
 	ssr_assert(bc && out);
-	*out = bc->A * A + bc->B * B + bc->C * C;
+	*out = bc->A * *A + bc->B * *B + bc->C * *C;
 	return out;
 }
 
diff --git a/src/core/shader.h b/src/core/shader.h
index 2e87ac5..d74581d 100644
--- a/src/core/shader.h
+++ b/src/core/shader.h
@@ -4,8 +4,9 @@
 #include "../math/math.h"
 #include "vert.h"
 #include "texture.h"
+#include "clip.h"
 
-typedef struct UniformCollection {
+typedef struct {
 	/*built in varaibles*/
 	Mat4* model;
 	Mat4* view;
@@ -23,12 +24,13 @@ typedef struct UniformCollection {
 	void* userdata;
 } UniformCollection;
 
-#define TEX(i) (uniforms->var_tex[i])
-#define UM4(i) (&uniforms->var_mat4[i])
-#define UV2(i) (&uniforms->var_vec2[i])
-#define UV3(i) (&uniforms->var_vec3[i])
-#define UV4(i) (&uniforms->var_vec4[i])
-#define UN(i) (&uniforms->var_num[i])
+#define UTEX(i) (uniforms->var_tex[i])
+#define UM4(i)  (&uniforms->var_mat4[i])
+#define UV2(i)  (&uniforms->var_vec2[i])
+#define UV3(i)  (&uniforms->var_vec3[i])
+#define UV4(i)  (&uniforms->var_vec4[i])
+#define UN(i)   (&uniforms->var_num[i])
+#define UU      (uniforms->userdata)
 
 typedef struct VertexShaderIn {
 	Vert* vertex;
@@ -39,10 +41,10 @@ typedef void(*VertexShader)(UniformCollection* uniforms, VertexShaderIn* in, Vec
 typedef struct FragmentShaderIn {
 	/*value from vertices interpolation*/
 	Vec3 position;
-	Vec3 normal; 
-	Vec3 tangent; 
-	Vec2 texcoord; 
-	Color color; 
+	Vec3 normal;
+	Vec3 tangent;
+	Vec2 texcoord;
+	Color color;
 	Vec4 joint;
 	Vec4 weight;
 	/*value from registers interpolation*/
@@ -52,68 +54,68 @@ typedef struct FragmentShaderIn {
 	Vec4 v4[6];
 } FragmentShaderIn;
 
-typedef bool(*FragmentShader)(UniformCollection* uniforms, FragmentShaderIn* in, Color* color);
+typedef bool(*FragmentShader)(UniformCollection* uniforms, FragmentShaderIn* in, Color32* color);
 
-typedef struct Program {
+typedef struct {
 	VertexShader vertexshader;
-	FragmentShader fragmentshader; 
+	FragmentShader fragmentshader;
 	uint varying_flag;
 } Program;
 
 #define VARYING_NONE     (0)
 
-#define VARYING_POSITION (1U << 31)
-#define VARYING_NORMAL   (1U << 30)
-#define VARYING_TANGENT  (1U << 29)
-#define VARYING_TEXCOORD (1U << 28)
-#define VARYING_COLOR		 (1U << 27)
-#define VARYING_JOINT		 (1U << 26)
-#define VARYING_WEIGHT   (1U << 25)
+#define VARYING_POSITION (1u << 31)
+#define VARYING_NORMAL   (1u << 30)
+#define VARYING_TANGENT  (1u << 29)
+#define VARYING_TEXCOORD (1u << 28)
+#define VARYING_COLOR		 (1u << 27)
+#define VARYING_JOINT		 (1u << 26)
+#define VARYING_WEIGHT   (1u << 25)
 
-#define VARYING_UNUSED   (1U << 24) 
+#define VARYING_UNUSED   (1u << 24) 
 
 #define VARYING_BASIC  ( VARYING_UNUSED | VARYING_POSITION | VARYING_NORMAL | VARYING_TANGENT | VARYING_TEXCOORD | VARYING_COLOR | VARYING_JOINT | VARYING_WEIGHT )
 
-#define VARYING_NUM_00		 (1U << 20)
-#define VARYING_NUM_01		 (1U << 21)
-#define VARYING_NUM_02		 (1U << 22)
-#define VARYING_NUM_03		 (1U << 23)
+#define VARYING_NUM_00		 (1u)
+#define VARYING_NUM_01		 (1u << 1)
+#define VARYING_NUM_02		 (1u << 2)
+#define VARYING_NUM_03		 (1u << 3)
 
 #define VARYING_NUM      (VARYING_NUM_00 | VARYING_NUM_01 | VARYING_NUM_02 | VARYING_NUM_03 )
 
-#define VARYING_V2_00		 (1U)
-#define VARYING_V2_01		 (1U << 1)
-#define VARYING_V2_02		 (1U << 2)
-#define VARYING_V2_03		 (1U << 3)
-#define VARYING_V2_04		 (1U << 4)
-#define VARYING_V2_05		 (1U << 5)
+#define VARYING_V2_00		 (1u << 4)
+#define VARYING_V2_01		 (1u << 5)
+#define VARYING_V2_02		 (1u << 6)
+#define VARYING_V2_03		 (1u << 7)
+#define VARYING_V2_04		 (1u << 8)
+#define VARYING_V2_05		 (1u << 9)
 
 #define VARYING_V2       (VARYING_V2_00 | VARYING_V2_01 | VARYING_V2_02 | VARYING_V2_03 | VARYING_V2_04 | VARYING_V2_05)
 
-#define VARYING_V3_00		 (1U << 6)
-#define VARYING_V3_01		 (1U << 7)
-#define VARYING_V3_02		 (1U << 8)
-#define VARYING_V3_03		 (1U << 9)
-#define VARYING_V3_04		 (1U << 10)
-#define VARYING_V3_05		 (1U << 11)
-#define VARYING_V3_06		 (1U << 12)
-#define VARYING_V3_07		 (1U << 13)
+#define VARYING_V3_00		 (1u << 10)
+#define VARYING_V3_01		 (1u << 11)
+#define VARYING_V3_02		 (1u << 12)
+#define VARYING_V3_03		 (1u << 13)
+#define VARYING_V3_04		 (1u << 14)
+#define VARYING_V3_05		 (1u << 15)
+#define VARYING_V3_06		 (1u << 16)
+#define VARYING_V3_07		 (1u << 17)
 
 #define VARYING_V3       (VARYING_V3_00 | VARYING_V3_01 | VARYING_V3_02 | VARYING_V3_03 | VARYING_V3_04 | VARYING_V3_05| VARYING_V3_06| VARYING_V3_07)
 
-#define VARYING_V4_00		 (1U << 14)
-#define VARYING_V4_01		 (1U << 15)
-#define VARYING_V4_02		 (1U << 16)
-#define VARYING_V4_03		 (1U << 17)
-#define VARYING_V4_04		 (1U << 18)
-#define VARYING_V4_05		 (1U << 19)
+#define VARYING_V4_00		 (1u << 18)
+#define VARYING_V4_01		 (1u << 19)
+#define VARYING_V4_02		 (1u << 20)
+#define VARYING_V4_03		 (1u << 21)
+#define VARYING_V4_04		 (1u << 22)
+#define VARYING_V4_05		 (1u << 23)
 
 #define VARYING_V4       (VARYING_V4_00 | VARYING_V4_01 | VARYING_V4_02 | VARYING_V4_03 | VARYING_V4_04 | VARYING_V4_05)
 
 #define VARYING_EXTRA    (VARYING_NUM | VARYING_V2 | VARYING_V3 | VARYING_V4)
 
-float* ssrS_bcpnum(Vec3* bc, float A, float B, float C, float* out);
 Color* ssrS_bcpcolor(Vec3* bc, Color A, Color B, Color C, Color* out);
+float* ssrS_bcpnum(Vec3* bc, float* A, float* B, float* C, float* out);
 Vec2*  ssrS_bcpvec2(Vec3* bc, Vec2* A, Vec2* B, Vec2* C, Vec2* out);
 Vec3* ssrS_bcpvec3(Vec3* bc, Vec3* A, Vec3* B, Vec3* C, Vec3* out);
 Vec4* ssrS_bcpvec4(Vec3* bc, Vec4* A, Vec4* B, Vec4* C, Vec4* out);
@@ -124,23 +126,24 @@ void ssrS_lerpvec2(float t, Vec2* A, Vec2* B, Vec2* out);
 void ssrS_lerpvec3(float t, Vec3* A, Vec3* B, Vec3* out);
 void ssrS_lerpvec4(float t, Vec4* A, Vec4* B, Vec4* out);
 
-void ssrS_setregisters(uint flag, int capacity);
-
 /*
 ** 顶点数据外，还提供额外的寄存器用来存储vertex shader输出的顶点额外数据
 ** 共24个
 */
-enum RegisterType {
+typedef enum {
 	REGTYPE_NUM = 1,
 	REGTYPE_VEC2,
 	REGTYPE_VEC4,
 	REGTYPE_VEC3,
-};
+} RegisterType;
 
-typedef struct Register {
+typedef struct {
 	uint length;
 	uint type;
+	uint element_size; /*element in bytes, 4\8\12\16*/
 	union {
+		byte* data;
+		/*pecific*/
 		float* num;
 		Vec2* v2;
 		Vec3* v3;
@@ -148,52 +151,52 @@ typedef struct Register {
 	};
 } Register;
 
+typedef struct {
+	uint element_size;
+	byte* data; /*mutable, either of registers or temp-registers*/
+	byte** accessor;
+	BcpInterpolator bcp_interpolator;
+	byte* output; /*fragment-in*/
+} ActiveReg;
+
+#define REG_TOTAL 24
 #define REG_NUM_COUNT 4
 #define REG_V2_COUNT 6 
 #define REG_V3_COUNT 8
 #define REG_V4_COUNT 6
 
-Register reg_num[4];
-Register reg_v2[6];
-Register reg_v3[8];
-Register reg_v4[6];
-
-/*寄存器指针，使用错误可能会出现野指针*/
-
-float* reg_num_00;
-float* reg_num_01;
-float* reg_num_02;
-float* reg_num_03;
-
-Vec2* reg_v2_00;
-Vec2* reg_v2_01;
-Vec2* reg_v2_02;
-Vec2* reg_v2_03;
-Vec2* reg_v2_04;
-Vec2* reg_v2_05;
-
-Vec3* reg_v3_00;
-Vec3* reg_v3_01;
-Vec3* reg_v3_02;
-Vec3* reg_v3_03;
-Vec3* reg_v3_04;
-Vec3* reg_v3_05;
-Vec3* reg_v3_06;
-Vec3* reg_v3_07;
-
-Vec4* reg_v4_00;
-Vec4* reg_v4_01;
-Vec4* reg_v4_02;
-Vec4* reg_v4_03;
-Vec4* reg_v4_04;
-Vec4* reg_v4_05;
+Register registers[REG_TOTAL];
+ActiveReg active_regs[REG_TOTAL];
+uint open_regsi[REG_TOTAL]; /*draw call用到的寄存器，可以索引到registers和active_regs*/
+
+#define REG_NUM_OFFSET 0
+#define REG_V2_OFFSET 4
+#define REG_V3_OFFSET 10
+#define REG_V4_OFFSET 18
+#define REG_NUM(i) (registers[i])
+#define REG_V2(i)  (registers[REG_V2_OFFSET + i])
+#define REG_V3(i)  (registers[REG_V3_OFFSET + i])
+#define REG_V4(i)  (registers[REG_V4_OFFSET + i])
+
+/*寄存器指针accessor，指向寄存器中的某个值，用于在shader里快速访问，使用错误可能会出现野指针*/
+float *reg_num_00, *reg_num_01, *reg_num_02, *reg_num_03;
+Vec2 *reg_v2_00, *reg_v2_01, *reg_v2_02, *reg_v2_03, *reg_v2_04, *reg_v2_05;
+Vec3 *reg_v3_00, *reg_v3_01, *reg_v3_02, *reg_v3_03, *reg_v3_04, *reg_v3_05, *reg_v3_06, *reg_v3_07;
+Vec4 *reg_v4_00, *reg_v4_01, *reg_v4_02, *reg_v4_03, *reg_v4_04, *reg_v4_05;
+
+void ssrS_openregs(uint varying_flag);
+void ssrS_setactiveregr();
+void ssrS_setactiveregc();
+void ssrS_solveregs(Vec3* bc, uint a, uint b, uint c);
+void ssrS_setregisters(int capacity);
 
 /*设置寄存器指针，指向寄存器（在fragment阶段，指针会指向fragmentIn结构）*/
-void ssrS_setupregisterpoints(uint extra_varying_flag, int idx);
+void ssrS_setupregisterpoints(int idx);
 
 /*设置寄存器指针，指向fragIn结构*/
-void ssrS_setregtofragin(uint extra_varying_flag, FragmentShaderIn* frag_in);
+void ssrS_setregtofragin();
 
-Color32 texture2d(Texture* tex, Vec2* uv);
+#define tex2d(tex, uv) \
+texture_sampling(tex, ssr_getfiltermode(), ssr_getwrapmode(), (uv).x, (uv).y)
 
 #endif
\ No newline at end of file
diff --git a/src/core/texture.c b/src/core/texture.c
index 5b23220..c7dc884 100644
--- a/src/core/texture.c
+++ b/src/core/texture.c
@@ -16,6 +16,15 @@ static void texture_abgr2argb(Texture* tex) {
 	}
 }
 
+static void texture_abgr2c32(Color* abgr, Color32* color, int count) {
+	for (int i = 0; i < count; ++i) {
+		color[i].r = (abgr[i] & 0xff) / 255.f;
+		color[i].g = ((abgr[i] >> 8) & 0xff) / 255.f;
+		color[i].b = ((abgr[i] >> 16) & 0xff) / 255.f;
+		color[i].a = ((abgr[i] >> 24) & 0xff) / 255.f;
+	}
+}
+
 Texture* texture_loadfromfile(const char* path) {
 	ssr_assert(path);
 	FILE* file = fopen(path, "rb");
@@ -38,13 +47,13 @@ Texture* texture_loadfromfile(const char* path) {
 	if (!pixels) return NULL;
 	Texture* texture = ssrM_new(Texture);
 	texture->width = width; texture->height = height;
-	texture->pixels = pixels;
-	texture_abgr2argb(texture);
+	texture->pixels = ssrM_newvector(Color32, width*height);
+	texture_abgr2c32(pixels, texture->pixels, width*height);
+	ssrM_free(pixels);
 	return texture;
 }
 
 static Color32 sampling(Texture* tex, WrapMode wrap_mode, int x, int y) {
-	Color32 color; 
 	if (wrap_mode == WRAPMODE_CLAMP) { /*clamp*/
 		x = clamp(x, 0, tex->width - 1);
 		y = clamp(y, 0, tex->height - 1);
@@ -52,30 +61,36 @@ static Color32 sampling(Texture* tex, WrapMode wrap_mode, int x, int y) {
 		x = x % tex->width; 
 		y = y % tex->height;
 	}
-	color_tocolor32(tex->pixels[x + y * tex->width], &color);
-	return color;
+	return tex->pixels[x + y * tex->width];
 }
 
 Color32 texture_sampling(Texture* tex, FilterMode filter_mode, WrapMode wrap_mode, float x01, float y01) {
 	ssr_assert(tex);
-	int x = x01 * tex->width, y = (1 - y01) * tex->height; /*map to texture coordinate*/
+	float x = x01 * tex->width, y = (1 - y01) * tex->height; /*map to texture coordinate*/
+	int x_min = floor(x), y_min = floor(y);
+	int x_max = ceil(x), y_max = ceil(y);
 	if (filter_mode == FILTERMODE_POINT) {
-		return sampling(tex, wrap_mode, x, y);
+		int px = (x_max - x > x - x_min) ? x_min : x_max;
+		int py = (y_max - y > y - y_min) ? y_min : y_max;
+
+		return sampling(tex, wrap_mode, px, py);
 	}
 	else if (filter_mode == FILTERMODE_BILINEAR) {
-		int x_min = floor(x), y_min = floor(y);
-		int x_max = ceil(x), y_max = ceil(y);
 		static Color32 tl, tr, bl, br, t, b, out;
+
 		tl = sampling(tex, wrap_mode, x_min, y_min);
 		tr = sampling(tex, wrap_mode, x_max, y_min);
 		bl = sampling(tex, wrap_mode, x_min, y_max);
 		br = sampling(tex, wrap_mode, x_max, y_max);
+
 		vec4_lerp(&tl, &tr, x - x_min, &t);
 		vec4_lerp(&bl, &br, x - x_min, &b);
 		vec4_lerp(&t, &b, y - y_min, &out);
+
 		return out;
 	}
+	/*
 	else if (filter_mode == FILTERMODE_TRILINEAR) {
-		/*TODO*/
 	}
+	*/
 }
diff --git a/src/core/texture.h b/src/core/texture.h
index 49fed4b..9e0d884 100644
--- a/src/core/texture.h
+++ b/src/core/texture.h
@@ -3,21 +3,21 @@
 
 #include "vert.h"
 
-typedef enum FilterMode {
+typedef enum {
 	FILTERMODE_POINT, 
 	FILTERMODE_BILINEAR,
-	FILTERMODE_TRILINEAR,
+	/*FILTERMODE_TRILINEAR,*/
 } FilterMode;
 
-typedef enum WrapMode{
+typedef enum {
 	WRAPMODE_REPEAT,
 	WRAPMODE_CLAMP,
 } WrapMode;
 
-typedef struct Texture{
+typedef struct {
 	uint width, height;
-	Color* pixels;
-}Texture;
+	Color32* pixels;
+} Texture;
 
 Texture* texture_loadfromfile(const char* path);
 
diff --git a/src/core/vert.h b/src/core/vert.h
index a85e38c..d594442 100644
--- a/src/core/vert.h
+++ b/src/core/vert.h
@@ -25,11 +25,13 @@ typedef struct Vert {
 	uint index;
 	Vec3 position;
 	Vec3 normal;
-	Vec3 tangent;
+	Vec4 tangent;
 	Vec2 texcoord;
 	Color color;
+/*
 	Vec4 joint;
 	Vec4 weight;
+*/
 } Vert;
 
 #endif
\ No newline at end of file