Subversion Repositories neorv32

-- #################################################################################################

-- # << NEORV32 - CPU Co-Processor: MULDIV unit (RISC-V "M" Extension)>>                           #

-- # ********************************************************************************************* #

-- # Multiplier and Divider unit. Implements the RISC-V RV32-M CPU extension.                      #

-- # Multiplier core (signed/unsigned) uses serial algorithm. -> 32+4 cycles latency               #

-- # Divider core (unsigned) uses serial algorithm. -> 32+6 cycles latency                         #

-- # Multiplications can be mapped to DSP block when FAST_MUL_EN = true.                           #

-- # ********************************************************************************************* #

-- # BSD 3-Clause License                                                                          #

-- #                                                                                               #

-- # Copyright (c) 2021, Stephan Nolting. All rights reserved.                                     #

-- #                                                                                               #

-- # Redistribution and use in source and binary forms, with or without modification, are          #

-- # permitted provided that the following conditions are met:                                     #

-- #                                                                                               #

-- # 1. Redistributions of source code must retain the above copyright notice, this list of        #

-- #    conditions and the following disclaimer.                                                   #

-- #                                                                                               #

-- # 2. Redistributions in binary form must reproduce the above copyright notice, this list of     #

-- #    conditions and the following disclaimer in the documentation and/or other materials        #

-- #    provided with the distribution.                                                            #

-- #                                                                                               #

-- # 3. Neither the name of the copyright holder nor the names of its contributors may be used to  #

-- #    endorse or promote products derived from this software without specific prior written      #

-- #    permission.                                                                                #

-- #                                                                                               #

-- # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS   #

-- # OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF               #

-- # MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE    #

-- # COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,     #

-- # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE #

-- # GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED    #

-- # AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING     #

-- # NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED  #

-- # OF THE POSSIBILITY OF SUCH DAMAGE.                                                            #

-- # ********************************************************************************************* #

-- # The NEORV32 Processor - https://github.com/stnolting/neorv32              (c) Stephan Nolting #

-- #################################################################################################

library ieee;

use ieee.std_logic_1164.all;

use ieee.numeric_std.all;

library neorv32;

use neorv32.neorv32_package.all;

entity neorv32_cpu_cp_muldiv is

  generic (

    FAST_MUL_EN : boolean := false -- use DSPs for faster multiplication

  );

  port (

    -- global control --

    clk_i   : in  std_ulogic; -- global clock, rising edge

    rstn_i  : in  std_ulogic; -- global reset, low-active, async

    ctrl_i  : in  std_ulogic_vector(ctrl_width_c-1 downto 0); -- main control bus

    start_i : in  std_ulogic; -- trigger operation

    -- data input --

    rs1_i   : in  std_ulogic_vector(data_width_c-1 downto 0); -- rf source 1

    rs2_i   : in  std_ulogic_vector(data_width_c-1 downto 0); -- rf source 2

    -- result and status --

    res_o   : out std_ulogic_vector(data_width_c-1 downto 0); -- operation result

    valid_o : out std_ulogic -- data output valid

  );

end neorv32_cpu_cp_muldiv;

architecture neorv32_cpu_cp_muldiv_rtl of neorv32_cpu_cp_muldiv is

  -- advanced configuration --

  constant dsp_add_reg_stage_c : boolean := false; -- add another register stage to DSP-based multiplication for timing-closure

  -- operations --

  constant cp_op_mul_c    : std_ulogic_vector(2 downto 0) := "000"; -- mul

  constant cp_op_mulh_c   : std_ulogic_vector(2 downto 0) := "001"; -- mulh

  constant cp_op_mulhsu_c : std_ulogic_vector(2 downto 0) := "010"; -- mulhsu

  constant cp_op_mulhu_c  : std_ulogic_vector(2 downto 0) := "011"; -- mulhu

  constant cp_op_div_c    : std_ulogic_vector(2 downto 0) := "100"; -- div

  constant cp_op_divu_c   : std_ulogic_vector(2 downto 0) := "101"; -- divu

  constant cp_op_rem_c    : std_ulogic_vector(2 downto 0) := "110"; -- rem

  constant cp_op_remu_c   : std_ulogic_vector(2 downto 0) := "111"; -- remu

  -- controller --

  type state_t is (IDLE, DECODE, INIT_OPX, INIT_OPY, PROCESSING, FINALIZE, COMPLETED, FAST_MUL);

  signal state         : state_t;

  signal cnt           : std_ulogic_vector(4 downto 0);

  signal cp_op         : std_ulogic_vector(2 downto 0); -- operation to execute

  signal cp_op_ff      : std_ulogic_vector(2 downto 0); -- operation that was executed

  signal start         : std_ulogic;

  signal operation     : std_ulogic;

  signal opx, opy      : std_ulogic_vector(data_width_c-1 downto 0); -- input operands

  signal opx_is_signed : std_ulogic;

  signal opy_is_signed : std_ulogic;

  signal opy_is_zero   : std_ulogic;

  signal div_res_corr  : std_ulogic;

  signal valid         : std_ulogic;

  -- divider core --

  signal remainder        : std_ulogic_vector(data_width_c-1 downto 0);

  signal quotient         : std_ulogic_vector(data_width_c-1 downto 0);

  signal div_sub          : std_ulogic_vector(data_width_c   downto 0);

  signal div_sign_comp_in : std_ulogic_vector(data_width_c-1 downto 0);

  signal div_sign_comp    : std_ulogic_vector(data_width_c-1 downto 0);

  signal div_res          : std_ulogic_vector(data_width_c-1 downto 0);

  -- multiplier core --

  signal mul_product    : std_ulogic_vector(63 downto 0);

  signal mul_do_add     : std_ulogic_vector(data_width_c downto 0);

  signal mul_sign_cycle : std_ulogic;

  signal mul_p_sext     : std_ulogic;

  signal mul_op_x       : signed(32 downto 0); -- for using DSPs

  signal mul_op_y       : signed(32 downto 0); -- for using DSPs

  signal mul_buf_ff     : signed(65 downto 0); -- for using DSPs

  signal mul_buf2_ff    : signed(65 downto 0); -- for using DSPs

begin

  -- Co-Processor Controller ----------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  coprocessor_ctrl: process(rstn_i, clk_i)

  begin

    if (rstn_i = '0') then

      state        <= IDLE;

      opx          <= (others => '0');

      opy          <= (others => '0');

      cnt          <= (others => '0');

      start        <= '0';

      valid        <= '0';

      div_res_corr <= '0';

      opy_is_zero  <= '0';

      cp_op_ff     <= (others => '0');

    elsif rising_edge(clk_i) then

      -- defaults --

      start    <= '0';

      valid    <= '0';

      cp_op_ff <= cp_op;

      -- FSM --

      case state is

        when IDLE =>

          opx <= rs1_i;

          opy <= rs2_i;

          if (start_i = '1') then

            state <= DECODE;

          end if;

        when DECODE =>

          --

          if (cp_op = cp_op_div_c) then -- result sign compensation for div?

            div_res_corr <= opx(opx'left) xor opy(opy'left);

          elsif (cp_op = cp_op_rem_c) then -- result sign compensation for rem?

            div_res_corr <= opx(opx'left);

          else

            div_res_corr <= '0';

          end if;

          --

          if (or_all_f(opy) = '0') then -- *divide* by 0?

            opy_is_zero <= '1';

          else

            opy_is_zero <= '0';

          end if;

          --

          cnt   <= "11111";

          if (operation = '1') then -- division

            state <= INIT_OPX;

          else -- multiplication

            start <= '1';

            if (FAST_MUL_EN = true) then

              state <= FAST_MUL;

            else

              state <= PROCESSING;

            end if;

          end if;

        when INIT_OPX =>

          if ((opx(opx'left) and opx_is_signed) = '1') then -- signed division?

            opx <= div_sign_comp; -- make positive

          end if;

          state <= INIT_OPY;

        when INIT_OPY =>

          start <= '1';

          if ((opy(opy'left) and opy_is_signed) = '1') then -- signed division?

            opy <= div_sign_comp; -- make positive

          end if;

          state <= PROCESSING;

        when PROCESSING =>

          cnt <= std_ulogic_vector(unsigned(cnt) - 1);

          if (cnt = "00000") then

            state <= FINALIZE;

          end if;

        when FAST_MUL =>

          state <= FINALIZE;

        when FINALIZE =>

          state <= COMPLETED;

        when COMPLETED =>

          valid <= '1';

          state <= IDLE;

      end case;

    end if;

  end process coprocessor_ctrl;

  -- co-processor command --

  cp_op <= ctrl_i(ctrl_ir_funct3_2_c downto ctrl_ir_funct3_0_c);

  -- operation --

  operation <= '1' when (cp_op = cp_op_div_c) or (cp_op = cp_op_divu_c) or (cp_op = cp_op_rem_c) or (cp_op = cp_op_remu_c) else '0';

  -- opx (rs1) signed? --

  opx_is_signed <= '1' when (cp_op = cp_op_mulh_c) or (cp_op = cp_op_mulhsu_c) or (cp_op = cp_op_div_c) or (cp_op = cp_op_rem_c) else '0';

  -- opy (rs2) signed? --

  opy_is_signed <= '1' when (cp_op = cp_op_mulh_c) or (cp_op = cp_op_div_c) or (cp_op = cp_op_rem_c) else '0';

  -- Multiplier Core (signed/unsigned) ------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  multiplier_core: process(clk_i)

  begin

    if rising_edge(clk_i) then

      -- ---------------------------------------------------------

      if (FAST_MUL_EN = false) then -- use small iterative computation

        if (start = '1') then -- start new multiplication

          mul_product(63 downto 32) <= (others => '0');

          mul_product(31 downto 00) <= opy;

        elsif ((state = PROCESSING) or (state = FINALIZE)) and (operation = '0') then

          mul_product(63 downto 31) <= mul_do_add(32 downto 0);

          mul_product(30 downto 00) <= mul_product(31 downto 1);

        end if;

      -- ---------------------------------------------------------

      else -- use direct approach using (several!) DSP blocks

        if (start = '1') then

          mul_op_x <= signed((opx(opx'left) and opx_is_signed) & opx);

          mul_op_y <= signed((opy(opy'left) and opy_is_signed) & opy);

        end if;

        mul_buf_ff <= mul_op_x * mul_op_y;

        if (dsp_add_reg_stage_c = true) then -- add another reg stage?

          mul_buf2_ff <= mul_buf_ff;

          mul_product <= std_ulogic_vector(mul_buf2_ff(63 downto 0)); -- let the register balancing do the magic here

        else

          mul_product <= std_ulogic_vector(mul_buf_ff(63 downto 0)); -- let the register balancing do the magic here

        end if;

      end if;

    end if;

  end process multiplier_core;

  -- MUL: do another addition --

  mul_update: process(mul_product, mul_sign_cycle, mul_p_sext, opx_is_signed, opx)

  begin

    -- current bit of opy to take care of --

    if (mul_product(0) = '1') then -- multiply with 1

      if (mul_sign_cycle = '1') then -- for signed operations only: take care of negative weighted MSB -> multiply with -1

        mul_do_add <= std_ulogic_vector(unsigned(mul_p_sext & mul_product(63 downto 32)) - unsigned((opx(opx'left) and opx_is_signed) & opx));

      else -- multiply with +1

        mul_do_add <= std_ulogic_vector(unsigned(mul_p_sext & mul_product(63 downto 32)) + unsigned((opx(opx'left) and opx_is_signed) & opx));

      end if;

    else -- multiply with 0

      mul_do_add <= mul_p_sext & mul_product(63 downto 32);

    end if;

  end process mul_update;

  -- sign control --

  mul_sign_cycle <= opy_is_signed when (state = FINALIZE) else '0';

  mul_p_sext     <= mul_product(mul_product'left) and opx_is_signed;

  -- Divider Core (unsigned) ----------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  divider_core: process(clk_i)

  begin

    if rising_edge(clk_i) then

      if (start = '1') then -- start new division

        quotient  <= opx;

        remainder <= (others => '0');

      elsif ((state = PROCESSING) or (state = FINALIZE)) and (operation = '1') then -- running?

        quotient <= quotient(30 downto 0) & (not div_sub(32));

        if (div_sub(32) = '0') then -- still overflowing

          remainder <= div_sub(31 downto 0);

        else -- underflow

          remainder <= remainder(30 downto 0) & quotient(31);

        end if;

      end if;

    end if;

  end process divider_core;

  -- DIV: try another subtraction --

  div_sub <= std_ulogic_vector(unsigned('0' & remainder(30 downto 0) & quotient(31)) - unsigned('0' & opy));

  -- Div sign compensation --

  div_sign_comp_in <= opx when (state = INIT_OPX) else

                      opy when (state = INIT_OPY) else

                      quotient when ((cp_op = cp_op_div_c) or (cp_op = cp_op_divu_c)) else remainder;

  div_sign_comp <= std_ulogic_vector(0 - unsigned(div_sign_comp_in));

  -- result sign correction --

  div_res <= div_sign_comp when (div_res_corr = '1') and (opy_is_zero = '0') else div_sign_comp_in;

  -- Data Output ----------------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  operation_result: process(valid, cp_op_ff, mul_product, div_res, quotient, opy_is_zero, rs1_i, remainder)

  begin

    if (valid = '1') then

      case cp_op_ff is

        when cp_op_mul_c =>

          res_o <= mul_product(31 downto 00);

        when cp_op_mulh_c | cp_op_mulhsu_c | cp_op_mulhu_c =>

          res_o <= mul_product(63 downto 32);

        when cp_op_div_c =>

          res_o <= div_res;

        when cp_op_divu_c =>

          res_o <= quotient;

        when cp_op_rem_c =>

          if (opy_is_zero = '0') then

            res_o <= div_res;

          else

            res_o <= rs1_i;

          end if;

        when others => -- cp_op_remu_c

          res_o <= remainder;

      end case;

    else

      res_o <= (others => '0');

    end if;

  end process operation_result;

  -- status output --

  valid_o <= valid;

end neorv32_cpu_cp_muldiv_rtl;