quantizer393.v

Go to the documentation of this file.

  // This file may be used to define same pre-processor macros to be included into each parsed file
`ifndef SYSTEM_DEFINES

  `define SYSTEM_DEFINES

  // TODO: Later compare instantiate/infer

  `define INSTANTIATE_DSP48E1

  `define DEBUG_DCT1D// undefine after debugging is over //  `define USE_OLD_DCT

  
// Parameters from x393_sata project

  `define USE_DRP

  `define ALIGN_CLOCKS

//  `define STRAIGHT_XCLK

  `define USE_DATASCOPE

//  `define DATASCOPE_INCOMING_RAW

  `define PRELOAD_BRAMS

//  `define AHCI_SATA 1

//  `define DEBUG_ELASTIC  

// End of parameters from x393_sata project

  
  `define PRELOAD_BRAMS

  `define DISPLAY_COMPRESSED_DATA    // if HISPI is not defined, parallel sensor interface is used for all channels

  `define HISPI /*************** CHANGE here and x393_hispi/x393_parallel in bitstream tool settings ****************///    `define USE_OLD_XDCT393  

//  `define USE_PCLK2X

//  `define USE_XCLK2X

  `define REVERSE_LANES 1  `define DEBUG_RING 1  `define USE_HARD_CURPARAMS// Adjustment of actual hardware may break simulation //  `define DEBUG_SENS_MEM_PAGES 1   

//  `define MCLK_VCO_MULT 16

// DDR3 memory speed grade and density

    `define sg25 1//  `define sg15E  1

//  `define sg187E  1

  `define den4096Mb 1    
  `define MCLK_VCO_MULT 16//  `define MCLK_VCO_MULT 18

//  `define MCLK_VCO_MULT 20

    
  `define MEMBRIDGE_DEBUG_WRITE 1// Enviroment-dependent options

  `ifdef IVERILOG

    `define SIMULATION

    `define OPEN_SOURCE_ONLY

  `endif

  
  `ifdef COCOTB

    `define SIMULATION

    `define OPEN_SOURCE_ONLY

  `endif

  
  `ifdef CVC

    `define SIMULATION

    `define OPEN_SOURCE_ONLY

  `endif // CVC

  
// will not use simultaneous reset in shift registers, just and input data with ~rst  

 `define SHREG_SEQUENTIAL_RESET 1// synthesis does to recognize global clock as G input of the primitive latch 

 `undef INFER_LATCHES

 // define when using CDC - it does not support them

 `undef IGNORE_ATTR

//`define MEMBRIDGE_DEBUG_READ 1

  `define use200Mhz 1  `define USE_CMD_ENCOD_TILED_32_RD 1    // chn 0 is read from memory and write to memory

 `define def_enable_mem_chn0

 `define def_read_mem_chn0

 `define def_write_mem_chn0

 `undef  def_scanline_chn0

 `undef  def_tiled_chn0

 
  // chn 1 is scanline r+w

 `define  def_enable_mem_chn1

 `define  def_read_mem_chn1

 `define  def_write_mem_chn1

 `define  def_scanline_chn1

 `undef   def_tiled_chn1


  // chn 2 is tiled r+w

 `define  def_enable_mem_chn2

 `define  def_read_mem_chn2

 `define  def_write_mem_chn2

 `undef   def_scanline_chn2

 `define  def_tiled_chn2


  // chn 3 is scanline r+w (reuse later)

 `define  def_enable_mem_chn3

 `define  def_read_mem_chn3

 `define  def_write_mem_chn3

 `define  def_scanline_chn3

 `undef   def_tiled_chn3


  // chn 4 is tiled r+w (reuse later)

 `define  def_enable_mem_chn4

 `define  def_read_mem_chn4

 `define  def_write_mem_chn4

 `undef   def_scanline_chn4

 `define  def_tiled_chn4


  // chn 5 is disabled

 `undef def_enable_mem_chn5


  // chn 6 is disabled

 `undef  def_enable_mem_chn6

 
  // chn 7 is disabled

 `undef  def_enable_mem_chn7

 
  // chn 8 is scanline w (sensor channel 0)

 `define  def_enable_mem_chn8

 `undef   def_read_mem_chn8

 `define  def_write_mem_chn8

 `define  def_scanline_chn8

 `undef   def_tiled_chn8


  // chn 9 is scanline w (sensor channel 1)

 `define  def_enable_mem_chn9

 `undef   def_read_mem_chn9

 `define  def_write_mem_chn9

 `define  def_scanline_chn9

 `undef   def_tiled_chn9


  // chn 10 is scanline w (sensor channel 2)

 `define  def_enable_mem_chn10

 `undef   def_read_mem_chn10

 `define  def_write_mem_chn10

 `define  def_scanline_chn10

 `undef   def_tiled_chn10


  // chn 11 is scanline w (sensor channel 3)

 `define  def_enable_mem_chn11

 `undef   def_read_mem_chn11

 `define  def_write_mem_chn11

 `define  def_scanline_chn11

 `undef   def_tiled_chn11


  // chn 12 is tiled read (compressor channel 0)

 `define  def_enable_mem_chn12

 `define  def_read_mem_chn12

 `undef   def_write_mem_chn12

 `undef   def_scanline_chn12

 `define  def_tiled_chn12

 
  // chn 12 is tiled read (compressor channel 1)

 `define  def_enable_mem_chn13

 `define  def_read_mem_chn13

 `undef   def_write_mem_chn13

 `undef   def_scanline_chn13

 `define  def_tiled_chn13

 
  // chn 12 is tiled read (compressor channel 2)

 `define  def_enable_mem_chn14

 `define  def_read_mem_chn14

 `undef   def_write_mem_chn14

 `undef   def_scanline_chn14

 `define  def_tiled_chn14

 
  // chn 12 is tiled read (compressor channel 3)

 `define  def_enable_mem_chn15

 `define  def_read_mem_chn15

 `undef   def_write_mem_chn15

 `undef   def_scanline_chn15

 `define  def_tiled_chn15

`endif
  
`timescale 1ns/1ps

// will add extracted DC (8 bits) to data from DCT here that will make data 12 bits (signed) long.
// It will be possible to make a sequintial multiplier for DC - but I'll skip this now.
module quantizer393(
    input             clk,           // pixel clock, posedge
    input             en,   // enable (0 resets counter)
    input             mclk, // system clock to write tables
    input             tser_qe, // enable write to a quantization table
    input             tser_ce, // enable write to a coring table
    input             tser_a_not_d,  // address/not data distributed to submodules
    input      [ 7:0] tser_d,        // byte-wide serialized tables address/data to submodules
    input             ctypei,   // component type input (Y/C)
    input      [ 8:0] dci,      // [7:0]   - average value in a block - subtracted before DCT. now normal signed number
    input             first_stb, //this is first stb pulse in a frame
    input             stb,      // strobe that writes ctypei, dci
    input      [ 2:0] tsi,   // table (quality) select [2:0]
    input             pre_start,// marks first input pixel (one before)
    input             first_in, // first block in (valid @ start)
    output reg        first_out, // valid @ ds
    input      [12:0] di,    // [11:0] pixel data in (signed)
    output reg [12:0] do,    // [11:0] pixel data out (AC is only 9 bits long?) - changed to 10
    output            dv,    // data out valid
    output            ds,  // data out strobe (one ahead of the start of dv)
    output reg [15:0] dc_tdo, //[15:0], MSB aligned coefficient for the DC component (used in focus module)
    input             dcc_en,  // enable dcc (sync to beginning of a new frame)
    input      [ 2:0] hfc_sel, // hight frequency components select [2:0] (includes components with both numbers >=hfc_sel
                               // hfc_sel == 3'h7 - now high frequency output - just 3 words - brightness and 2 color diffs
    input             color_first, // first MCU in a frame
    input      [ 2:0] coring_num, // coring table pair number (0..7)
    output reg        dcc_vld, // single cycle when dcc_data is valid
    output     [15:0] dcc_data,  // [15:0] dc component data out (for reading by software) 
    input      [ 7:0] n000,      // input [7:0] number of zero pixels (255 if 256) - to be multiplexed with dcc
    input      [ 7:0] n255);     // input [7:0] number of 0xff pixels (255 if 256) - to be multiplexed with dcc

    
    wire       [3:0] tdco; // coring table output
    reg        [3:0] tbac; // coring memory table number (LSB - color)
    reg              coring_range; // input <16, use coring LUT
    wire      [15:0] tdo;
    reg       [ 9:0] tba;   // table output (use) address   
    wire      [15:0] zigzag_q;
    reg              wpage;
    reg              rpage;
    wire      [ 5:0] zwa;
    reg       [ 5:0] zra;
    reg       [12:0] qdo;
    reg       [12:0] qdo0;
    reg              zwe;
    reg       [12:0] d1;
    reg       [12:0] d2,d3; // registered data in, converted to sign+ absolute value
    wire      [27:0] qmul;
    wire             start_a;
    reg       [15:0] tdor;
    reg       [20:0] qmulr; // added 7 bits to total8 fractional for biasing/zero bin
    wire             start_out;
    wire             start_z;
    reg       [ 8:0] dc1;   // registered DC average - with restored sign   

// for fifo for ctype, dc
    wire            ctype;
    wire     [ 8:0] dc;
//    wire            next_dv;
//    reg      [ 2:0] last_dv; // last dv cycle (will turn of unless new ds)
    reg      [ 2:0] ds_r;
    reg      [ 3:0] ren;
    
    reg      [ 5:0] start;
    wire            dcc_stb;
    reg             dcc_run;
    reg             dcc_first;
    reg             dcc_Y;
    reg      [ 1:0] ctype_prev;
    reg      [12:0] dcc_acc;
    reg      [12:0] hfc_acc;
    wire            hfc_en;
    reg             hfc_copy; // copy hfc_acc to dcc_acc
    wire     [10:0] d2_dct;   // 11 bits enough, convetred to positive (before - 0 was in the middle - pixel value 128) - dcc only
    reg             sel_satnum; // select saturation numbers - dcc only
    reg      [15:0] pre_dc_tdo;
    wire            copy_dc_tdo;

    reg             first_interm; // valid @ ds

    wire     [ 2:0] ts;
    wire     [ 2:0] coring_sel;

    reg      [ 2:0] block_mem_ra;
    reg      [ 2:0] block_mem_wa;
    reg      [ 2:0] block_mem_wa_save;
    reg      [15:0] block_mem_ram[0:7];
    wire     [15:0] block_mem_o=block_mem_ram[block_mem_ra[2:0]];

    assign dc[8:0] =          block_mem_o[8:0];
    assign ctype =            block_mem_o[9];
    assign ts[2:0] =          block_mem_o[12:10];
    assign coring_sel[2:0] =  block_mem_o[15:13];

    assign start_a = start[5];
    assign start_z = start[4];
    assign dcc_stb = start[2];
    assign ds = ds_r[2];
    assign dv = ren[3];
    always @ (posedge clk) begin
        if (stb) block_mem_ram[block_mem_wa[2:0]] <= {coring_num[2:0],tsi[2:0], ctypei, dci[8:0]};

        if      (!en) block_mem_wa[2:0] <= 3'h0;
        else if (stb) block_mem_wa[2:0] <= block_mem_wa[2:0] +1;

        if (stb && first_stb)  block_mem_wa_save[2:0] <= block_mem_wa[2:0];

        if      (!en)       block_mem_ra[2:0] <= 3'h0;
        else if (pre_start) block_mem_ra[2:0] <= first_in?block_mem_wa_save[2:0]:(block_mem_ra[2:0] +1);
    end
 
    assign        d2_dct[10:0]={!d2[11] ^ ctype_prev[0], d2[9:0]}; 

    assign        dcc_data[15:0]=sel_satnum?
                    {n255[7:0],n000[7:0]}:
                    {dcc_first || (!dcc_Y && dcc_acc[12]) ,(!dcc_Y && dcc_acc[12]), (!dcc_Y && dcc_acc[12]), dcc_acc[12:0]};
//    assign         do[12:0]=zigzag_q[12:0];
    assign        qmul[27:0]=tdor[15:0]*d3[11:0];

    assign         start_out =   zwe && (zwa[5:0]== 6'h3f);   //adjust?
    assign         copy_dc_tdo = zwe && (zwa[5:0]== 6'h37);   // not critical

//    assign next_dv=en && (ds || (dv && (zra[5:0]!=6'h00)));    
    always @ (posedge clk) begin
        d1[12:0]      <= di[12:0];
//inv_sign
        dc1[8:0] <= start[0]?dc[8:0]:9'b0;   // sync to d1[8:0]ctype valid at start, not later
        d2[12:0] <= {dc1[8],dc1[8:0],3'b0} + d1[12:0];
        d3[12]   <= d2[12];
        d3[11:0] <= d2[12]? -d2[11:0]:d2[11:0];

        if (start[0] || !en) tba[9:6] <= {ts[2:0],ctype};
      
/// TODO - make sure ctype switches at needed time (compensate if needed) *****************************************
        if (start[3] || !en) tbac[3:0] <= {coring_sel[2:0],ctype}; // table number to use

        if      (start[0])        tba[5:0] <= 6'b0;
        else if (tba[5:0]!=6'h3f) tba[5:0] <= tba[5:0]+1;
        
        tdor[15:0]  <= tdo[15:0]; // registered table data out
        
        if (start[3])  pre_dc_tdo[15:0] <= tdor[15:0]; //16-bit q. tables)
        
        if (copy_dc_tdo) dc_tdo[15:0]     <= pre_dc_tdo[15:0];
        
        qmulr[19:0] <= qmul[27:8]; // absolute value
        qmulr[20]   <= d3[12];     // sign
        qdo0[12]    <= qmulr[20];  // sign
      
// tdco[3:0] - same timing as qdo0;      
// use lookup table from 8 bits of absolute value (4.4 - 4 fractional) to calculate 4 bit coring output that would replace output
// if input is less thahn 16. For larger values the true rounding will be used.

// Absolute values here have quantization coefficients already applied, so we can use the same coring table for all DCT coefficients.
// there are be 16 tables - 8 Y/C pairs to switch
        qdo0[11:0]  <= qmulr[19:8] + qmulr[7]; // true rounding of the absolute value 
        coring_range<= !(|qmulr[19:12]) && !(&qmulr[11:7]) ; // valid with qdo0
        qdo[11:0]   <= coring_range? (qdo0[12]?-{8'h0,tdco[3:0]}:{8'h0,tdco[3:0]}):(qdo0[12]?-qdo0[11:0]:qdo0[11:0]);
        qdo[12]     <= qdo0[12] && (!coring_range || (tdco[3:0]!=4'h0)); 

        if (start_out)    rpage <= wpage;

//        last_dv <= {last_dv[1:0], en && (zra[5:0] == 6'h3f)};
        
        if (start_out)    zra[5:0] <= 6'b0;
//        else if (zra[5:0]!=6'h3f)   zra[5:0] <= zra[5:0]+1; // conserving energy
        else if (ren[0])  zra[5:0] <= zra[5:0]+1; // conserving energy
        
        ds_r    <= {ds_r[1:0], en && start_out};
        
        if      (!en)                 ren[0] <= 0;
        else if (start_out)           ren[0] <= 1;
        else if ((zra[5:0] == 6'h3f)) ren[0] <= 0;
        
        if      (!en)                 ren[3:1] <= 0;
        else                          ren[3:1] <= ren [2:0]; 
        
        if (ren[2])                   do[12:0] <= zigzag_q[12:0];
        
        if (start_a)   first_interm <= first_in;
        if (start_out) first_out    <=first_interm;
// zwe???
        zwe <= en && (start_a || (zwe && (zwa[5:0]!=6'h3f)));
        if          (!en) wpage <= 1'b0;
        else if (start_a) wpage <= ~wpage;
    end


    always @ (posedge clk) begin
        sel_satnum <= dcc_run && (start[0]? (ctype_prev[1:0]==2'b10): sel_satnum);
        
        hfc_copy <= dcc_run && (hfc_sel[2:0]!=3'h7) && (tba[5:0]==6'h1f) && ctype_prev[0] && ctype_prev[1];
        
        start[5:0] <= {start[4:0], pre_start}; // needed?
        
        if    (!dcc_en) dcc_run <= 1'b0;
        else if (start[0]) dcc_run <= 1'b1;
        
        if (!dcc_en)    ctype_prev[1:0] <= 2'b11;
        else if (start[0]) ctype_prev[1:0] <= {ctype_prev[0],ctype && dcc_run}; 
        
        if (dcc_stb || hfc_copy) dcc_acc[12:0] <= hfc_copy?
                                                hfc_acc[12:0]:
                                               {(d2_dct[10]&&ctype_prev[0]),(d2_dct[10]&&ctype_prev[0]),d2_dct[10:0]}+((ctype_prev[0] || ctype_prev[1])?13'h0:dcc_acc[12:0]);
                                               
        if (!dcc_run || hfc_copy) hfc_acc <=13'b0;
        else if (hfc_en) hfc_acc <= hfc_acc + {2'b0, d3[10:0]};
        
        if (dcc_stb) dcc_first <= color_first && dcc_run && dcc_stb && ctype && !ctype_prev[0];
        
        if (dcc_stb) dcc_Y <= dcc_run && dcc_stb && ctype && !ctype_prev[0];
        
        dcc_vld <= (dcc_run && dcc_stb && (ctype || ctype_prev[0] || sel_satnum)) || hfc_copy;
    end
    
    wire          twqe;
    wire          twce;
    wire  [15:0]  tdi;
    wire  [22:0]  ta;
    

    table_ad_receive #( // here may be changed to 8-bit from 16-bit
        .MODE_16_BITS (1),
        .NUM_CHN      (2)
    ) table_ad_receive_i (
        .clk       (mclk),              // input
        .a_not_d   (tser_a_not_d),      // input
        .ser_d     (tser_d),            // input[7:0] 
        .dv        ({tser_ce,tser_qe}), // input[1:0] 
        .ta        (ta), // output[22:0] 
        .td        (tdi), // output[15:0] 
        .twe       ({twce,twqe}) // output[1:0] 
    );


//    SRL16 i_hfc_en (.Q(hfc_en), .A0(1'b1), .A1(1'b0), .A2(1'b0), .A3(1'b0), .CLK(clk),
//                    .D(((tba[2:0]>hfc_sel[2:0]) || (tba[5:3]>hfc_sel[2:0])) && dcc_run && !ctype_prev[0])); // dly=1+1
    dly_16 #(.WIDTH(1)) i_hfc_en (
        .clk(clk),
        .rst(1'b0),
        .dly(4'd1),
        .din(((tba[2:0]>hfc_sel[2:0]) || (tba[5:3]>hfc_sel[2:0])) && dcc_run && !ctype_prev[0]),
        .dout(hfc_en));   // dly=1+1

    zigzag393 i_zigzag(   .clk(clk),
                     .start(start_z),
                      .q(zwa[5:0]));

    // All memories below are non-registered, see if they can be made registered
    ram18_var_w_var_r #(
        .REGISTERS    (0),
        .LOG2WIDTH_WR (4),
        .LOG2WIDTH_RD (4),
        .DUMMY        (0)
`ifdef PRELOAD_BRAMS, .INIT_00 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_01 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_02 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_03 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_04 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_05 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_06 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_07 (256'hFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF)
, .INIT_08 (256'h174615551555333340005555800080001555199A200033335555800080005555)
, .INIT_09 (256'h155510000F0F199A2AAB40005555555517461249174620003333555555555555)
, .INIT_0A (256'h0E390B210C31100013B117462492333311110C310BA312491746249240004000)
, .INIT_0B (256'h0CCD0C310CCD0BA30CCD0D790E3912490CCD0AAB0AAB0C310F0F100013B1199A)
, .INIT_0C (256'h0CCD0CCD0CCD0CCD13B13333400040000CCD0CCD0CCD0CCD1C72333340005555)
, .INIT_0D (256'h0CCD0CCD0CCD0CCD0CCD0CCD13B11C720CCD0CCD0CCD0CCD0CCD174633333333)
, .INIT_0E (256'h0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD)
, .INIT_0F (256'h0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD0CCD)

`endif
    ) i_quant_table (
        .rclk         (clk),                          // input
        .raddr        ({tba[9:6],tba[2:0],tba[5:3]}), // input[8:0] 
        .ren          (1'b1),                         // input
        .regen        (1'b1),                         // input
        .data_out     (tdo[15:0]),                    // output[15:0] 
        .wclk         (mclk),                         // input
        .waddr        (ta[9:0]),                      // input[8:0] 
        .we           (twqe),                         // input
        .web          (4'hf),                         // input[3:0] 
        .data_in      (tdi[15:0])                     // input[15:0] 
    );

    ram18_var_w_var_r #(
        .REGISTERS    (0),
        .LOG2WIDTH_WR (4),
        .LOG2WIDTH_RD (2),
        .DUMMY        (0)
`ifdef PRELOAD_BRAMS, .INIT_00 (256'h4444444433333333333333332222222222222222111111111111111100000000)
, .INIT_01 (256'h8888888877777777777777776666666666666666555555555555555544444444)
, .INIT_02 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_03 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_04 (256'h4444444433333333333333332222222222222222111111111111111100000000)
, .INIT_05 (256'h8888888877777777777777776666666666666666555555555555555544444444)
, .INIT_06 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_07 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_08 (256'h4444444433333333333333322222222222222221111111111111100000000000)
, .INIT_09 (256'h8888888877777777777777776666666666666666555555555555555544444444)
, .INIT_0A (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_0B (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_0C (256'h4444444433333333333333322222222222222221111111111111100000000000)
, .INIT_0D (256'h8888888877777777777777776666666666666666555555555555555544444444)
, .INIT_0E (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_0F (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_10 (256'h4444444333333333333333322222222222222211111111111000000000000000)
, .INIT_11 (256'h8888888877777777777777776666666666666666555555555555555444444444)
, .INIT_12 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_13 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_14 (256'h4444444333333333333333322222222222222211111111111000000000000000)
, .INIT_15 (256'h8888888877777777777777776666666666666666555555555555555444444444)
, .INIT_16 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_17 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_18 (256'h4444443333333333333222222222211111111111100000000000000000000000)
, .INIT_19 (256'h8888888877777777777777766666666666666665555555555555555444444444)
, .INIT_1A (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_1B (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_1C (256'h4444443333333333333222222222211111111111100000000000000000000000)
, .INIT_1D (256'h8888888877777777777777766666666666666665555555555555555444444444)
, .INIT_1E (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999988888888)
, .INIT_1F (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_20 (256'h3333333332222222222211111111111111000000000000000000000000000000)
, .INIT_21 (256'h8888888777777777777777766666666666666665555555555555544444444444)
, .INIT_22 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999888888888)
, .INIT_23 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_24 (256'h3333333332222222222211111111111111000000000000000000000000000000)
, .INIT_25 (256'h8888888777777777777777766666666666666665555555555555544444444444)
, .INIT_26 (256'hCCCCCCCCBBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA999999999999999888888888)
, .INIT_27 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_28 (256'h2222222222211111111111111110000000000000000000000000000000000000)
, .INIT_29 (256'h8888888777777777777777666666666666655555555554444444443333333333)
, .INIT_2A (256'hCCCCCCCCBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA9999999999999999888888888)
, .INIT_2B (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_2C (256'h2222222222211111111111111110000000000000000000000000000000000000)
, .INIT_2D (256'h8888888777777777777777666666666666655555555554444444443333333333)
, .INIT_2E (256'hCCCCCCCCBBBBBBBBBBBBBBBAAAAAAAAAAAAAAAA9999999999999999888888888)
, .INIT_2F (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCC)
, .INIT_31 (256'h5555444444443333333333322222222222221111111111111111111111110000)
, .INIT_32 (256'hCCCCCCCBBBBBBBBBBBBBBAAAAAAAAA9999999998888888777777766666665555)
, .INIT_33 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCCC)
, .INIT_35 (256'h5555444444443333333333322222222222221111111111111111111111110000)
, .INIT_36 (256'hCCCCCCCBBBBBBBBBBBBBBAAAAAAAAA9999999998888888777777766666665555)
, .INIT_37 (256'hFFFFFFFFFFFFFFFFFFFFFFFFEEEEEEEEEEEEEEEDDDDDDDDDDDDDDDDCCCCCCCCC)
, .INIT_3B (256'hFFFFFFFFFEEDCBA9987766555444333332222222211111111111111111111100)
, .INIT_3F (256'hFFFFFFFFFEEDCBA9987766555444333332222222211111111111111111111100)

`endif
    ) i_coring_table (
        .rclk         (clk), // input
        .raddr        ({tbac[3:0],qmulr[11:4]}), // input[10:0] 
        .ren          (1'b1),                    // input
        .regen        (1'b1),                    // input
        .data_out     (tdco[3:0]),               // output[3:0] 
        .wclk         (mclk),                    // input
        .waddr        (ta[9:0]),                 // input[9:0] 
        .we           (twce),                    // input
        .web          (4'hf),                    // input[3:0] 
        .data_in      (tdi[15:0])                // input[15:0] 
    );

    ram18_var_w_var_r #(
        .REGISTERS    (1),
        .LOG2WIDTH_WR (4),
        .LOG2WIDTH_RD (4),
        .DUMMY        (0)
    ) i_zigzagbuf (
        .rclk         (clk),                     // input
        .raddr        ({3'b0,rpage,zra[5:0]}),   // input[9:0] 
        .ren          (ren[0]),                  // input
        .regen        (ren[1]),                  // input
        .data_out     (zigzag_q[15:0]),          // output[15:0] 
        .wclk         (clk),                     // input
        .waddr        ({3'b0,wpage,zwa[5:0]}),   // input[9:0] 
        .we           (zwe),                     // input
        .web          (4'hf),                    // input[3:0] 
        .data_in      ({3'b0,qdo[12:0]})         // input[15:0] 
    );


endmodule

module zigzag393 (
    input            clk,           // system clock, posedge
    input            start,
    output reg [5:0] q);

    reg        [5:0] a;
    wire      [ 4:0] rom_a;
    reg       [ 5:0] rom_q;

    assign   rom_a[4:0]=a[5]?(~a[4:0]):a[4:0];

    always @ (posedge clk) begin
        if (start)   a[5:0] <= 6'b0;
        else   if (a[5:0]!=6'h3f) a[5:0] <= a[5:0]+1;
    end

    // ROM (combinatorial)
    always @(rom_a) case (rom_a)
        5'h00: rom_q <= 6'h00;
        5'h01: rom_q <= 6'h02;
        5'h02: rom_q <= 6'h03;
        5'h03: rom_q <= 6'h09;
        5'h04: rom_q <= 6'h0a;
        5'h05: rom_q <= 6'h14;
        5'h06: rom_q <= 6'h15;
        5'h07: rom_q <= 6'h23;
        5'h08: rom_q <= 6'h01;
        5'h09: rom_q <= 6'h04;
        5'h0a: rom_q <= 6'h08;
        5'h0b: rom_q <= 6'h0b;
        5'h0c: rom_q <= 6'h13;
        5'h0d: rom_q <= 6'h16;
        5'h0e: rom_q <= 6'h22;
        5'h0f: rom_q <= 6'h24;
        5'h10: rom_q <= 6'h05;
        5'h11: rom_q <= 6'h07;
        5'h12: rom_q <= 6'h0c;
        5'h13: rom_q <= 6'h12;
        5'h14: rom_q <= 6'h17;
        5'h15: rom_q <= 6'h21;
        5'h16: rom_q <= 6'h25;
        5'h17: rom_q <= 6'h30;
        5'h18: rom_q <= 6'h06;
        5'h19: rom_q <= 6'h0d;
        5'h1a: rom_q <= 6'h11;
        5'h1b: rom_q <= 6'h18;
        5'h1c: rom_q <= 6'h20;
        5'h1d: rom_q <= 6'h26;
        5'h1e: rom_q <= 6'h2f;
        5'h1f: rom_q <= 6'h31;
    endcase
    // add symmetrical part
    always @ (posedge clk) q[5:0]   <= a[5]? (~rom_q[5:0]):rom_q[5:0];
endmodule