add justfile and polish project structure

.justfile

@@ -0,0 +1,64 @@
+alias c := clean-build
+alias b := build
+alias r := run
+
+# default target
+default := "Visp_Pipeline"
+
+# clean all outpu files
+clean-all:
+  rm -rf build logs
+
+# clean build output files
+clean-build:
+  rm -rf ./build/*
+
+# build and run default target
+all: build run
+
+# clean and build
+rebuild: clean-build build
+
+# generate build files
+generate:
+  cmake . -B ./build/ -G Ninja
+
+# build default target
+build: generate
+  ninja -C ./build/ {{default}} -v
+
+# build all project
+build-all: generate
+  cmake --build ./build/ -v
+
+# build ISP target
+build-Visp: generate
+  cmake --bulid ./build/ -t Visp
+
+# build ISP Pipeline target
+build-Visp_Pipeline: generate
+  cmake --bulid ./build/ -t Visp_Pipeline
+
+# run default target
+run:
+  exec "./build/{{default}}"
+
+# run ISP
+run-Visp:
+  ./build/Visp
+
+# run ISP Pipeline
+run-Visp_Pipeline:
+  ./build/Visp_Pipeline
+
+# run default target with tracing wave
+trace:
+  exec "./build/{{default}}" +trace
+
+# run ISP with tracing wave
+trace-Visp:
+  ./build/Visp +trace
+
+# run ISP Pipeline with tracing wave
+trace-Visp_Pipeline:
+  ./build/Visp_Pipeline +trace

CMakeLists.txt

@@ -39,19 +39,25 @@ find_package(SystemCLanguage QUIET)
 # file(GLOB_RECURSE IMG_SRC ${PROJECT_SOURCE_DIR}/src/img_process/*.cpp)
 # add_library(img_process STATIC ${IMG_SRC})
 
+# Set input and output location
+set(INPUT_IMG ${PROJECT_SOURCE_DIR}/src/transform/test.bin)
+set(OUTPUT_DIR ${PROJECT_SOURCE_DIR}/logs/)
+add_compile_definitions(
+  INPUT_IMG="${INPUT_IMG}"
+  OUTPUT_DIR="${OUTPUT_DIR}"
+)
+
+# Get RTL source code dir
+SUBDIRLIST(RTL_SUBDIR ${PROJECT_SOURCE_DIR}/rtl)
+
+
 # ----------------------  EXE  ---------------------------
 #                         VISP
 # ----------------------  EXE  ---------------------------
 add_executable(Visp ${PROJECT_SOURCE_DIR}/src/sc_main.cpp)
-target_compile_definitions(Visp
-  PRIVATE INPUT_IMG="${PROJECT_SOURCE_DIR}/src/transform/test.bin"
-  PRIVATE OUTPUT_DIR="${PROJECT_SOURCE_DIR}/logs/"
-)
 target_include_directories(Visp PRIVATE ${PROJECT_SOURCE_DIR}/src/img_process)
 # target_link_libraries(Visp PRIVATE img_process)
 
-# Get RTL source code dir
-SUBDIRLIST(RTL_SUBDIR ${PROJECT_SOURCE_DIR}/rtl)
 # Add the Verilated circuit to the target
 verilate(Visp SYSTEMC COVERAGE TRACE
   INCLUDE_DIRS ${RTL_SUBDIR}
@@ -67,15 +73,13 @@ verilator_link_systemc(Visp)
 #                    VISP_Pipeline
 # ----------------------  EXE  ---------------------------
 add_executable(Visp_Pipeline ${PROJECT_SOURCE_DIR}/src/sc_main_pipeline.cpp)
-target_compile_definitions(Visp_Pipeline
+target_include_directories(
-  PRIVATE INPUT_IMG="${PROJECT_SOURCE_DIR}/src/transform/test.bin"
+  Visp_Pipeline
-  PRIVATE OUTPUT_DIR="${PROJECT_SOURCE_DIR}/logs/"
+  PRIVATE ${PROJECT_SOURCE_DIR}/src/img_process
+  PRIVATE ${PROJECT_SOURCE_DIR}/src
 )
-target_include_directories(Visp_Pipeline PRIVATE ${PROJECT_SOURCE_DIR}/src/img_process)
 # target_link_libraries(Visp_Pipeline PRIVATE img_process)
 
-# Get RTL source code dir
-SUBDIRLIST(RTL_SUBDIR ${PROJECT_SOURCE_DIR}/rtl)
 # Add the Verilated circuit to the target
 verilate(Visp_Pipeline SYSTEMC COVERAGE TRACE
   INCLUDE_DIRS ${RTL_SUBDIR}

FPGA.nix

@@ -19,6 +19,15 @@
   
   # Enable languages support
   # languages.cplusplus.enable = true;
+  languages.python = {
+    enable = true;
+    version = "3.11";
+    venv.enable = true;
+    venv.requirements = ''
+      teroshdl
+      cocotb
+    '';
+  };
   
   # When enter shell, exec ...
   enterShell = ''

flake.lock

@@ -191,6 +191,22 @@
         "type": "github"
       }
     },
+    "flake-compat_3": {
+      "flake": false,
+      "locked": {
+        "lastModified": 1696426674,
+        "narHash": "sha256-kvjfFW7WAETZlt09AgDn1MrtKzP7t90Vf7vypd3OL1U=",
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "rev": "0f9255e01c2351cc7d116c072cb317785dd33b33",
+        "type": "github"
+      },
+      "original": {
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "type": "github"
+      }
+    },
     "flake-parts": {
       "inputs": {
         "nixpkgs-lib": [
@@ -429,6 +445,27 @@
         "type": "github"
       }
     },
+    "nixpkgs-python": {
+      "inputs": {
+        "flake-compat": "flake-compat_3",
+        "nixpkgs": [
+          "nixpkgs"
+        ]
+      },
+      "locked": {
+        "lastModified": 1722978926,
+        "narHash": "sha256-sqOOEaKJJSUFBzag/cGeeXV491TrrVFY3DFBs1w20V8=",
+        "owner": "cachix",
+        "repo": "nixpkgs-python",
+        "rev": "7c550bca7e6cf95898e32eb2173efe7ebb447460",
+        "type": "github"
+      },
+      "original": {
+        "owner": "cachix",
+        "repo": "nixpkgs-python",
+        "type": "github"
+      }
+    },
     "nixpkgs-regression": {
       "locked": {
         "lastModified": 1643052045,
@@ -511,14 +548,18 @@
     },
     "nixpkgs_3": {
       "locked": {
-        "lastModified": 315532800,
+        "lastModified": 1728018373,
-        "narHash": "sha256-DXLcxy2Vmxj0vD/GACAsw/JCeyi3UJlVl/9iU+ZnPnU=",
+        "narHash": "sha256-NOiTvBbRLIOe5F6RbHaAh6++BNjsb149fGZd1T4+KBg=",
-        "type": "tarball",
+        "owner": "NixOS",
-        "url": "https://mirrors.ustc.edu.cn/nix-channels/nixpkgs-unstable/nixexprs.tar.xz"
+        "repo": "nixpkgs",
+        "rev": "bc947f541ae55e999ffdb4013441347d83b00feb",
+        "type": "github"
       },
       "original": {
-        "type": "tarball",
+        "owner": "NixOS",
-        "url": "https://mirrors.ustc.edu.cn/nix-channels/nixpkgs-unstable/nixexprs.tar.xz"
+        "ref": "nixos-unstable",
+        "repo": "nixpkgs",
+        "type": "github"
       }
     },
     "poetry2nix": {
@@ -614,7 +655,8 @@
     "root": {
       "inputs": {
         "devenv": "devenv",
-        "nixpkgs": "nixpkgs_3"
+        "nixpkgs": "nixpkgs_3",
+        "nixpkgs-python": "nixpkgs-python"
       }
     },
     "systems": {

flake.nix

@@ -1,6 +1,8 @@
 {
   inputs = {
-    nixpkgs.url = "https://mirrors.ustc.edu.cn/nix-channels/nixpkgs-unstable/nixexprs.tar.xz";
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
+    nixpkgs-python.url = "github:cachix/nixpkgs-python";
+    nixpkgs-python.inputs.nixpkgs.follows = "nixpkgs";
     devenv.url = "github:cachix/devenv";
     devenv.inputs.nixpkgs.follows = "nixpkgs";
   };

rtl/Demosaic/Demosaic_Pipeline.sv

@@ -1,11 +1,11 @@
 `timescale 1ns / 1ps
 module Demosaic_Pipeline #(
-parameter WINDOW_LENGTH = 3,
+    parameter WINDOW_LENGTH = 3,
-parameter reg [15:0] TOTAL_WIDTH  = 512+3, // 总图像宽度
+    parameter reg [15:0] TOTAL_WIDTH  = 512+3, // 总图像宽度
-parameter reg [15:0] TOTAL_HEIGHT = 256+3, // 总图像高度
+    parameter reg [15:0] TOTAL_HEIGHT = 256+3, // 总图像高度
-parameter reg [ 1:0] RAW_TYPE  = 3,        // (0,0)位置算起RAW_TYPE的值
+    parameter reg [ 1:0] RAW_TYPE  = 3,        // (0,0)位置算起RAW_TYPE的值
-parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
+    parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
-)(
+  )(
     input wire clk,
     input wire reset,
 
@@ -20,7 +20,7 @@ parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
 
     output reg out_hsync,     // 行同步，一行第一个像素点输出的同时高电平
     output reg out_fsync      // 帧同步，一帧第一个像素点输出的同时高电平
-);
+  );
 
   localparam DATA_NUM = WINDOW_LENGTH*WINDOW_LENGTH;
   localparam PIPILINE = 3;
@@ -40,9 +40,12 @@ parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
   reg [1:0] raw_type;
 
   integer i;
-  always @(posedge clk) begin
+  always @(posedge clk)
-    if(reset) begin
+  begin
-      for(i=0;i<DATA_NUM;i=i+1) data_cache[i] <= 0;
+    if(reset)
+    begin
+      for(i=0;i<DATA_NUM;i=i+1)
+        data_cache[i] <= 0;
       pipeline_valid <= 0;
       {red, green, blue} <= 0;
       {out_data[2],out_data[1],out_data[0]}  <= 0;
@@ -51,36 +54,45 @@ parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
       pos_y <= ~0;
       temp_pos_x <= 0;
       temp_pos_y <= 0;
-    end else if(pipeline_running) begin
+    end
+    else if(pipeline_running)
+    begin
 
       pipeline_valid <= {pipeline_valid[PIPILINE-2:0], in_valid};
 
-      if(in_valid) begin
+      if(in_valid)
-        for(i=0;i<DATA_NUM;i=i+1) data_cache[i] <= in_data[i];
+      begin
+        for(i=0;i<DATA_NUM;i=i+1)
+          data_cache[i] <= in_data[i];
         pos_x <= (pos_x >= TOTAL_WIDTH - 1)?(0):(pos_x + 1);
         pos_y <= (pos_x >= TOTAL_WIDTH - 1)?((pos_y >= TOTAL_HEIGHT - 1)?(0):(pos_y + 1)):(pos_y);
       end
 
-      if(pipeline_valid[0]) begin
+      if(pipeline_valid[0])
+      begin
         temp_pos_x <= pos_x;
         temp_pos_y <= pos_y;
         case (raw_type)
-          0: begin  // Missing B, R on G
+          0:
+          begin  // Missing B, R on G
             blue  <= (data_cache[1] + data_cache[7]) >> 1;
             red   <= (data_cache[3] + data_cache[5]) >> 1;
             green <= data_cache[4];
           end
-          1: begin  // Missing G, R on B
+          1:
+          begin  // Missing G, R on B
             green <= (data_cache[1] + data_cache[3] + data_cache[5] + data_cache[7]) >> 2;
             red <= (data_cache[0] + data_cache[2] + data_cache[6] + data_cache[8]) >> 2;
             blue <= data_cache[4];
           end
-          2: begin  // Missing G, B on R
+          2:
+          begin  // Missing G, B on R
             green <= (data_cache[1] + data_cache[3] + data_cache[5] + data_cache[7]) >> 2;
             blue <= (data_cache[0] + data_cache[2] + data_cache[6] + data_cache[8]) >> 2;
             red <= data_cache[4];
           end
-          3: begin  // Missing B, R on G
+          3:
+          begin  // Missing B, R on G
             red   <= (data_cache[1] + data_cache[7]) >> 1;
             blue  <= (data_cache[3] + data_cache[5]) >> 1;
             green <= data_cache[4];
@@ -88,7 +100,8 @@ parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
         endcase
       end
 
-      if(pipeline_valid[1]) begin
+      if(pipeline_valid[1])
+      begin
         {out_data[2],out_data[1],out_data[0]}  <= {red,blue,green};
         out_hsync <= (temp_pos_x == 0);
         out_fsync <= ((temp_pos_x == 0) && (temp_pos_y == 0));
@@ -98,13 +111,20 @@ parameter reg [ 4:0] DATA_WIDTH = 16       // 输入/输出数据位宽
 
   // 0:gr 1:rg 2:bg 3:gb  窗口右移，0<->1 2<->3; 窗口下移，0<->2，1<->3。
   //   bg   gb   gr   rg
-  always @(*) begin
+  always @(*)
-    if(reset) raw_type = RAW_TYPE;
+  begin
-    else case (RAW_TYPE)
+    if(reset)
-          2'b00: raw_type = {pos_y[0], pos_x[0]};
+      raw_type = RAW_TYPE;
-          2'b01: raw_type = {pos_y[0], ~pos_x[0]};
+    else
-          2'b10: raw_type = {~pos_y[0], pos_x[0]};
+    case (RAW_TYPE)
-          2'b11: raw_type = {~pos_y[0], ~pos_x[0]};
+      2'b00:
+        raw_type = {pos_y[0], pos_x[0]};
+      2'b01:
+        raw_type = {pos_y[0], ~pos_x[0]};
+      2'b10:
+        raw_type = {~pos_y[0], pos_x[0]};
+      2'b11:
+        raw_type = {~pos_y[0], ~pos_x[0]};
     endcase
   end
 

src/sc_main_pipeline.cpp

@@ -12,7 +12,6 @@
 #include <systemc>
 
 // Include common routines
-#include <string>
 #include <sys/stat.h> // mkdir
 #include <utility>
 #include <vector>
@@ -22,8 +21,8 @@
 // Include model header, generated from Verilating "isp.v"
 #include "Visp_Pipeline.h"
 
-// Write Pictures
+// Include testbench written by systemc
-#include "bitmap_image.hpp"
+#include "tb_isp.hpp"
 
 // Image Parameters
 static const uint16_t IN_WIDTH = 1936;
@@ -56,161 +55,6 @@ static const double contrast = 1.2;
 using namespace sc_core;
 using namespace sc_dt;
 
-SC_MODULE(TB_ISP) {
-  sc_in_clk clk;
-  sc_in<bool> rst;
-
-  sc_in<bool> in_ready;         // next module ready to receive data
-  sc_out<bool> out_valid;       // next module data valid signal
-  sc_out<uint32_t> out_data[3]; // next module receive data
-
-  sc_in<bool> in_valid;    // this module receive data valid signal
-  sc_out<bool> out_ready;  // this module ready to receive data
-  sc_in<uint32_t> in_data; // this module receive data
-
-  bool is_done;                // when receive all data
-  std::vector<uint16_t> image; // the data of image
-  std::vector<uint32_t> process_image = std::vector<uint32_t>(
-      OUT_SIZE, 0); // after isp process, the data of image
-
-  SC_CTOR(TB_ISP) {
-    SC_CTHREAD(sendData, clk.pos()); // when clk posedge, exec sendData
-    reset_signal_is(rst, true);      // set rst signal
-
-    SC_CTHREAD(readData, clk.pos());
-    reset_signal_is(rst, true); // set rst signal
-  }
-
-  void sendData(void) {
-    // init var
-    uint16_t pos_x = 0, pos_y = 0, cnt_flame = 0;
-    bool is_finish = false; // when send all data
-    // reset
-    out_valid = false;
-    for (auto &data : out_data)
-      data = 0;
-
-    while (true) {
-      if (in_ready && !is_finish) {
-        // valid and send data
-        out_valid = true;
-        out_data[0] = image[(pos_y + 0) * IN_WIDTH + pos_x];
-        out_data[1] = image[(pos_y + 1) * IN_WIDTH + pos_x];
-        out_data[2] = image[(pos_y + 2) * IN_WIDTH + pos_x];
-
-        // print data
-        std::printf("x=%4d, y=%4d, data=0x%04x\t", pos_x, pos_y,
-                    image[(pos_y + 0) * IN_WIDTH + pos_x]);
-        std::printf("x=%4d, y=%4d, data=0x%04x\t", pos_x, pos_y,
-                    image[(pos_y + 1) * IN_WIDTH + pos_x]);
-        std::printf("x=%4d, y=%4d, data=0x%04x\n", pos_x, pos_y,
-                    image[(pos_y + 2) * IN_WIDTH + pos_x]);
-        pos_x++;
-
-        // calculate position and recognize when to finish
-        if (pos_x >= IN_WIDTH) {
-          pos_x = 0;
-          pos_y++;
-        }
-        if (pos_y >= IN_HEIGHT - 1) {
-          pos_y = 0;
-          cnt_flame++;
-        }
-        if (cnt_flame >= FLAMES) {
-          is_finish = true;
-        }
-      } else {
-        out_valid = false;
-      }
-
-      // wait for next clk
-      wait();
-    }
-  }
-
-  void readData(void) {
-    // init local var
-    uint16_t pos_x = 0, pos_y = 0, cnt_flame = 0;
-    uint32_t last_data = 0, cnt = 0;
-    bool is_finish = false;
-    // reset
-    out_ready = false;
-    is_done = false;
-
-    while (true) {
-      if (!is_finish) {
-        out_ready = true;
-
-        // when data valid, write it down
-        if (in_valid) {
-          process_image[pos_y * OUT_WIDTH + pos_x] = in_data;
-
-          // calculate position
-          pos_x++;
-
-          if (pos_x >= OUT_WIDTH) {
-            pos_x = 0;
-            pos_y++;
-          }
-          if (pos_y >= OUT_HEIGHT) {
-            pos_y = 0;
-            saveData(
-                ("output_img_" + std::to_string(cnt_flame) + ".bmp").c_str());
-            cnt_flame++;
-          }
-          if (cnt_flame >= FLAMES) {
-            is_finish = true;
-          }
-        }
-      } else {
-        out_ready = false;
-      }
-
-      // when no data send, give finish signal
-      if (is_finish && (last_data == in_data)) {
-        cnt++;
-        if (cnt >= 100000L) { // when receive many times the same data
-          is_done = true;
-          std::printf("Finish Reading data; pos_x = %d, pos_y = %d\n", pos_x,
-                      pos_y);
-        }
-      } else {
-        cnt = 0;
-      }
-      last_data = in_data;
-
-      // wait for next clk
-      wait();
-    }
-  }
-
-  bool saveData(const char *name) {
-    bool ret = true;
-
-    // Check Image Size
-    if (process_image.size() > OUT_SIZE) {
-      std::cout << "Process Image Over Size!!!\n"
-                << "Image Size:" << process_image.size() << "\n";
-      return false;
-    }
-
-    // Write BMP image
-    bitmap_image bmp(OUT_WIDTH, OUT_HEIGHT);
-    if (!bmp) {
-      std::cout << "Output Image Open Failed!!!\n";
-      return false;
-    }
-    for (int y = 0; y < OUT_HEIGHT; y++)
-      for (int x = 0; x < OUT_WIDTH; x++)
-        bmp.set_pixel(x, y,
-                      (process_image[y * OUT_WIDTH + x] & 0x00ff0000) >> 16,
-                      (process_image[y * OUT_WIDTH + x] & 0x0000ff00) >> 8,
-                      (process_image[y * OUT_WIDTH + x] & 0x000000ff) >> 0);
-    bmp.save_image(std::string(OUTPUT_DIR) + name);
-    return ret;
-  }
-};
-
 int sc_main(int argc, char *argv[]) {
   std::printf("Enter into sc_main\n");
 
@@ -314,7 +158,7 @@ int sc_main(int argc, char *argv[]) {
   isp.gain_blue(gain_blue);
 
   // Construct testbench module
-  TB_ISP tb_isp("tb_isp");
+  TB_ISP tb_isp("tb_isp", IN_WIDTH, IN_HEIGHT, OUT_WIDTH, OUT_HEIGHT, FLAMES, OUTPUT_DIR);
   tb_isp.image = std::move(in_image);
   tb_isp.clk(clk);
   tb_isp.rst(rst);

src/tb_isp.hpp

@@ -0,0 +1,180 @@
+#ifndef __TB_ISP_H__
+#define __TB_ISP_H__
+
+#include <cstdint>
+#include <systemc>
+
+// Write Pictures
+#include "bitmap_image.hpp"
+
+SC_MODULE(TB_ISP) {
+  sc_core::sc_in_clk clk;
+  sc_core::sc_in<bool> rst;
+
+  sc_core::sc_in<bool> in_ready;         // next module ready to receive data
+  sc_core::sc_out<bool> out_valid;       // next module data valid signal
+  sc_core::sc_out<uint32_t> out_data[3]; // next module receive data
+
+  sc_core::sc_in<bool> in_valid;    // this module receive data valid signal
+  sc_core::sc_out<bool> out_ready;  // this module ready to receive data
+  sc_core::sc_in<uint32_t> in_data; // this module receive data
+
+  const uint16_t IN_WIDTH;
+  const uint16_t IN_HEIGHT;
+  const uint32_t IN_SIZE;
+  const uint16_t OUT_WIDTH;
+  const uint16_t OUT_HEIGHT;
+  const uint32_t OUT_SIZE;
+  const uint32_t FLAMES;
+  const std::string OUT_DIR;
+
+  bool is_done;                        // when receive all data
+  std::vector<uint16_t> image;         // the data of image
+  std::vector<uint32_t> process_image; // after isp process, the data of image
+
+  SC_CTOR(TB_ISP, const uint16_t in_width, const uint16_t in_height,
+          const uint16_t out_width, const uint16_t out_height,
+          const uint32_t cnt_flame, const std::string& out_dir)
+      : IN_WIDTH(in_width), IN_HEIGHT(in_height), IN_SIZE(in_width * in_height),
+        OUT_WIDTH(out_width), OUT_HEIGHT(out_height),
+        OUT_SIZE(out_width * out_height), FLAMES(cnt_flame),
+        OUT_DIR(out_dir),
+        process_image(std::vector<uint32_t>(out_width * out_height, 0)) {
+    SC_CTHREAD(sendData, clk.pos()); // when clk posedge, exec sendData
+    reset_signal_is(rst, true);      // set rst signal
+
+    SC_CTHREAD(readData, clk.pos());
+    reset_signal_is(rst, true); // set rst signal
+  }
+
+  void sendData(void) {
+    // init var
+    uint16_t pos_x = 0, pos_y = 0, cnt_flame = 0;
+    bool is_finish = false; // when send all data
+    // reset
+    out_valid = false;
+    for (auto &data : out_data)
+      data = 0;
+
+    while (true) {
+      if (in_ready && !is_finish) {
+        // valid and send data
+        out_valid = true;
+        out_data[0] = image[(pos_y + 0) * IN_WIDTH + pos_x];
+        out_data[1] = image[(pos_y + 1) * IN_WIDTH + pos_x];
+        out_data[2] = image[(pos_y + 2) * IN_WIDTH + pos_x];
+
+        // print data
+        std::printf("x=%4d, y=%4d, data=0x%04x\t", pos_x, pos_y,
+                    image[(pos_y + 0) * IN_WIDTH + pos_x]);
+        std::printf("x=%4d, y=%4d, data=0x%04x\t", pos_x, pos_y,
+                    image[(pos_y + 1) * IN_WIDTH + pos_x]);
+        std::printf("x=%4d, y=%4d, data=0x%04x\n", pos_x, pos_y,
+                    image[(pos_y + 2) * IN_WIDTH + pos_x]);
+        pos_x++;
+
+        // calculate position and recognize when to finish
+        if (pos_x >= IN_WIDTH) {
+          pos_x = 0;
+          pos_y++;
+        }
+        if (pos_y >= IN_HEIGHT - 1) {
+          pos_y = 0;
+          cnt_flame++;
+        }
+        if (cnt_flame >= FLAMES) {
+          is_finish = true;
+        }
+      } else {
+        out_valid = false;
+      }
+
+      // wait for next clk
+      wait();
+    }
+  }
+
+  void readData(void) {
+    // init local var
+    uint16_t pos_x = 0, pos_y = 0, cnt_flame = 0;
+    uint32_t last_data = 0, cnt = 0;
+    bool is_finish = false;
+    // reset
+    out_ready = false;
+    is_done = false;
+
+    while (true) {
+      if (!is_finish) {
+        out_ready = true;
+
+        // when data valid, write it down
+        if (in_valid) {
+          process_image[pos_y * OUT_WIDTH + pos_x] = in_data;
+
+          // calculate position
+          pos_x++;
+
+          if (pos_x >= OUT_WIDTH) {
+            pos_x = 0;
+            pos_y++;
+          }
+          if (pos_y >= OUT_HEIGHT) {
+            pos_y = 0;
+            saveData(
+                ("output_img_" + std::to_string(cnt_flame) + ".bmp").c_str());
+            cnt_flame++;
+          }
+          if (cnt_flame >= FLAMES) {
+            is_finish = true;
+          }
+        }
+      } else {
+        out_ready = false;
+      }
+
+      // when no data send, give finish signal
+      if (is_finish && (last_data == in_data)) {
+        cnt++;
+        if (cnt >= 100000L) { // when receive many times the same data
+          is_done = true;
+          std::printf("Finish Reading data; pos_x = %d, pos_y = %d\n", pos_x,
+                      pos_y);
+        }
+      } else {
+        cnt = 0;
+      }
+      last_data = in_data;
+
+      // wait for next clk
+      wait();
+    }
+  }
+
+  bool saveData(const char *name) {
+    bool ret = true;
+
+    // Check Image Size
+    if (process_image.size() > OUT_SIZE) {
+      std::cout << "Process Image Over Size!!!\n"
+                << "Image Size:" << process_image.size() << "\n";
+      return false;
+    }
+
+    // Write BMP image
+    bitmap_image bmp(OUT_WIDTH, OUT_HEIGHT);
+    if (!bmp) {
+      std::cout << "Output Image Open Failed!!!\n";
+      return false;
+    }
+    for (int y = 0; y < OUT_HEIGHT; y++)
+      for (int x = 0; x < OUT_WIDTH; x++)
+        bmp.set_pixel(x, y,
+                      (process_image[y * OUT_WIDTH + x] & 0x00ff0000) >> 16,
+                      (process_image[y * OUT_WIDTH + x] & 0x0000ff00) >> 8,
+                      (process_image[y * OUT_WIDTH + x] & 0x000000ff) >> 0);
+    bmp.save_image(std::string(OUT_DIR) + name);
+    return ret;
+  }
+};
+
+#endif