mirror of https://github.com/YosysHQ/yosys.git
564 lines
15 KiB
Verilog
564 lines
15 KiB
Verilog
// The core logic primitive of the Cyclone V/10GX is the Adaptive Logic Module
|
|
// (ALM). Each ALM is made up of an 8-input, 2-output look-up table, covered
|
|
// in this file, connected to combinational outputs, a carry chain, and four
|
|
// D flip-flops (which are covered as MISTRAL_FF in dff_sim.v).
|
|
//
|
|
// The ALM is vertically symmetric, so I find it helps to think in terms of
|
|
// half-ALMs, as that's predominantly the unit that synth_intel_alm uses.
|
|
//
|
|
// ALMs are quite flexible, having multiple modes.
|
|
//
|
|
// Normal (combinational) mode
|
|
// ---------------------------
|
|
// The ALM can implement:
|
|
// - a single 6-input function (with the other inputs usable for flip-flop access)
|
|
// - two 5-input functions that share two inputs
|
|
// - a 5-input and a 4-input function that share one input
|
|
// - a 5-input and a 3-or-less-input function that share no inputs
|
|
// - two 4-or-less-input functions that share no inputs
|
|
//
|
|
// Normal-mode functions are represented as MISTRAL_ALUTN cells with N inputs.
|
|
// It would be possible to represent a normal mode function as a single cell -
|
|
// the vendor cyclone{v,10gx}_lcell_comb cell does exactly that - but I felt
|
|
// it was more user-friendly to print out the specific function sizes
|
|
// separately.
|
|
//
|
|
// With the exception of MISTRAL_ALUT6, you can think of two normal-mode cells
|
|
// fitting inside a single ALM.
|
|
//
|
|
// Extended (7-input) mode
|
|
// -----------------------
|
|
// The ALM can also fit a 7-input function made of two 5-input functions that
|
|
// share four inputs, multiplexed by another input.
|
|
//
|
|
// Because this can't accept arbitrary 7-input functions, Yosys can't handle
|
|
// it, so it doesn't have a cell, but I would likely call it MISTRAL_ALUT7(E?)
|
|
// if it did, and it would take up a full ALM.
|
|
//
|
|
// It might be possible to add an extraction pass to examine all ALUT5 cells
|
|
// that feed into ALUT3 cells to see if they can be combined into an extended
|
|
// ALM, but I don't think it will be worth it.
|
|
//
|
|
// Arithmetic mode
|
|
// ---------------
|
|
// In arithmetic mode, each half-ALM uses its carry chain to perform fast addition
|
|
// of two four-input functions that share three inputs. Oddly, the result of
|
|
// one of the functions is inverted before being added (you can see this as
|
|
// the dot on a full-adder input of Figure 1-8 in the Handbook).
|
|
//
|
|
// The cell for an arithmetic-mode half-ALM is MISTRAL_ALM_ARITH. One idea
|
|
// I've had (or rather was suggested by mwk) is that functions that feed into
|
|
// arithmetic-mode cells could be packed directly into the arithmetic-mode
|
|
// cell as a function, which reduces the number of ALMs needed.
|
|
//
|
|
// Shared arithmetic mode
|
|
// ----------------------
|
|
// Shared arithmetic mode looks a lot like arithmetic mode, but here the
|
|
// output of every other four-input function goes to the input of the adder
|
|
// the next bit along. What this means is that adding three bits together can
|
|
// be done in an ALM, because functions can be used to implement addition that
|
|
// then feeds into the carry chain. This means that three bits can be added per
|
|
// ALM, as opposed to two in the arithmetic mode.
|
|
//
|
|
// Shared arithmetic mode doesn't currently have a cell, but I intend to add
|
|
// it as MISTRAL_ALM_SHARED, and have it occupy a full ALM. Because it adds
|
|
// three bits per cell, it makes addition shorter and use less ALMs, but
|
|
// I don't know enough to tell whether it's more efficient to use shared
|
|
// arithmetic mode to shorten the carry chain, or plain arithmetic mode with
|
|
// the functions packed in.
|
|
|
|
`default_nettype none
|
|
|
|
// Cyclone V LUT output timings (picoseconds):
|
|
//
|
|
// CARRY A B C D E F G
|
|
// COMBOUT - 605 583 510 512 - 97 400 (LUT6)
|
|
// COMBOUT - 602 583 457 510 302 93 483 (LUT7)
|
|
// SUMOUT 368 1342 1323 887 927 - 785 -
|
|
// CARRYOUT 71 1082 1062 866 813 - 1198 -
|
|
|
|
(* abc9_lut=2, lib_whitebox *)
|
|
module MISTRAL_ALUT6(input A, B, C, D, E, F, output Q);
|
|
|
|
parameter [63:0] LUT = 64'h0000_0000_0000_0000;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 605;
|
|
(B => Q) = 583;
|
|
(C => Q) = 510;
|
|
(D => Q) = 512;
|
|
(E => Q) = 400;
|
|
(F => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 275;
|
|
(B => Q) = 272;
|
|
(C => Q) = 175;
|
|
(D => Q) = 165;
|
|
(E => Q) = 162;
|
|
(F => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = LUT >> {F, E, D, C, B, A};
|
|
|
|
endmodule
|
|
|
|
|
|
(* abc9_lut=1, lib_whitebox *)
|
|
module MISTRAL_ALUT5(input A, B, C, D, E, output Q);
|
|
|
|
parameter [31:0] LUT = 32'h0000_0000;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 583;
|
|
(B => Q) = 510;
|
|
(C => Q) = 512;
|
|
(D => Q) = 400;
|
|
(E => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 272;
|
|
(B => Q) = 175;
|
|
(C => Q) = 165;
|
|
(D => Q) = 162;
|
|
(E => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = LUT >> {E, D, C, B, A};
|
|
|
|
endmodule
|
|
|
|
|
|
(* abc9_lut=1, lib_whitebox *)
|
|
module MISTRAL_ALUT4(input A, B, C, D, output Q);
|
|
|
|
parameter [15:0] LUT = 16'h0000;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 510;
|
|
(B => Q) = 512;
|
|
(C => Q) = 400;
|
|
(D => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 175;
|
|
(B => Q) = 165;
|
|
(C => Q) = 162;
|
|
(D => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = LUT >> {D, C, B, A};
|
|
|
|
endmodule
|
|
|
|
|
|
(* abc9_lut=1, lib_whitebox *)
|
|
module MISTRAL_ALUT3(input A, B, C, output Q);
|
|
|
|
parameter [7:0] LUT = 8'h00;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 510;
|
|
(B => Q) = 400;
|
|
(C => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 165;
|
|
(B => Q) = 162;
|
|
(C => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = LUT >> {C, B, A};
|
|
|
|
endmodule
|
|
|
|
|
|
(* abc9_lut=1, lib_whitebox *)
|
|
module MISTRAL_ALUT2(input A, B, output Q);
|
|
|
|
parameter [3:0] LUT = 4'h0;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 400;
|
|
(B => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 162;
|
|
(B => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = LUT >> {B, A};
|
|
|
|
endmodule
|
|
|
|
|
|
(* abc9_lut=1, lib_whitebox *)
|
|
module MISTRAL_NOT(input A, output Q);
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => Q) = 97;
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => Q) = 53;
|
|
endspecify
|
|
`endif
|
|
|
|
assign Q = ~A;
|
|
|
|
endmodule
|
|
|
|
// Despite the abc9_carry attributes, this doesn't seem to stop ABC9 adding illegal fanout to the carry chain that nextpnr cannot handle.
|
|
// So we treat it as a total blackbox from ABC9's perspective for now.
|
|
// (* abc9_box, lib_whitebox *)
|
|
module MISTRAL_ALUT_ARITH(input A, B, C, D0, D1, /* (* abc9_carry *) */ input CI, output SO, /* (* abc9_carry *) */ output CO);
|
|
|
|
parameter LUT0 = 16'h0000;
|
|
parameter LUT1 = 16'h0000;
|
|
|
|
`ifdef cyclonev
|
|
specify
|
|
(A => SO) = 1342;
|
|
(B => SO) = 1323;
|
|
(C => SO) = 927;
|
|
(D0 => SO) = 887;
|
|
(D1 => SO) = 785;
|
|
(CI => SO) = 368;
|
|
|
|
(A => CO) = 1082;
|
|
(B => CO) = 1062;
|
|
(C => CO) = 813;
|
|
(D0 => CO) = 866;
|
|
(D1 => CO) = 1198;
|
|
(CI => CO) = 36; // Divided by 2 to account for there being two ALUT_ARITHs in an ALM)
|
|
endspecify
|
|
`endif
|
|
`ifdef cyclone10gx
|
|
specify
|
|
(A => SO) = 644;
|
|
(B => SO) = 477;
|
|
(C => SO) = 416;
|
|
(D0 => SO) = 380;
|
|
(D1 => SO) = 431;
|
|
(CI => SO) = 276;
|
|
|
|
(A => CO) = 525;
|
|
(B => CO) = 433;
|
|
(C => CO) = 712;
|
|
(D0 => CO) = 653;
|
|
(D1 => CO) = 593;
|
|
(CI => CO) = 16;
|
|
endspecify
|
|
`endif
|
|
|
|
wire q0, q1;
|
|
|
|
assign q0 = LUT0 >> {D0, C, B, A};
|
|
assign q1 = LUT1 >> {D1, C, B, A};
|
|
|
|
assign {CO, SO} = q0 + !q1 + CI;
|
|
|
|
endmodule
|
|
|
|
|
|
/*
|
|
// A, B, C0, C1, E0, E1, F0, F1: data inputs
|
|
// CARRYIN: carry input
|
|
// SHAREIN: shared-arithmetic input
|
|
// CLK0, CLK1, CLK2: clock inputs
|
|
//
|
|
// COMB0, COMB1: combinational outputs
|
|
// FF0, FF1, FF2, FF3: DFF outputs
|
|
// SUM0, SUM1: adder outputs
|
|
// CARRYOUT: carry output
|
|
// SHAREOUT: shared-arithmetic output
|
|
module MISTRAL_ALM(
|
|
input A, B, C0, C1, E0, E1, F0, F1, CARRYIN, SHAREIN, // LUT path
|
|
input CLK0, CLK1, CLK2, AC0, AC1, // FF path
|
|
output COMB0, COMB1, SUM0, SUM1, CARRYOUT, SHAREOUT,
|
|
output FF0, FF1, FF2, FF3
|
|
);
|
|
|
|
parameter LUT0 = 16'b0000;
|
|
parameter LUT1 = 16'b0000;
|
|
parameter LUT2 = 16'b0000;
|
|
parameter LUT3 = 16'b0000;
|
|
|
|
parameter INIT0 = 1'b0;
|
|
parameter INIT1 = 1'b0;
|
|
parameter INIT2 = 1'b0;
|
|
parameter INIT3 = 1'b0;
|
|
|
|
parameter C0_MUX = "C0";
|
|
parameter C1_MUX = "C1";
|
|
|
|
parameter F0_MUX = "VCC";
|
|
parameter F1_MUX = "GND";
|
|
|
|
parameter FEEDBACK0 = "FF0";
|
|
parameter FEEDBACK1 = "FF2";
|
|
|
|
parameter ADD_MUX = "LUT";
|
|
|
|
parameter DFF01_DATA_MUX = "COMB";
|
|
parameter DFF23_DATA_MUX = "COMB";
|
|
|
|
parameter DFF0_CLK = "CLK0";
|
|
parameter DFF1_CLK = "CLK0";
|
|
parameter DFF2_CLK = "CLK0";
|
|
parameter DFF3_CLK = "CLK0";
|
|
|
|
parameter DFF0_AC = "AC0";
|
|
parameter DFF1_AC = "AC0";
|
|
parameter DFF2_AC = "AC0";
|
|
parameter DFF3_AC = "AC0";
|
|
|
|
// Feedback muxes from the flip-flop outputs.
|
|
wire ff_feedback_mux0, ff_feedback_mux1;
|
|
|
|
// C-input muxes which can be set to also use the F-input.
|
|
wire c0_input_mux, c1_input_mux;
|
|
|
|
// F-input muxes which can be set to a constant to allow LUT5 use.
|
|
wire f0_input_mux, f1_input_mux;
|
|
|
|
// Adder input muxes to select between shared-arithmetic mode and arithmetic mode.
|
|
wire add0_input_mux, add1_input_mux;
|
|
|
|
// Combinational-output muxes for LUT #1 and LUT #3
|
|
wire lut1_comb_mux, lut3_comb_mux;
|
|
|
|
// Sum-output muxes for LUT #1 and LUT #3
|
|
wire lut1_sum_mux, lut3_sum_mux;
|
|
|
|
// DFF data-input muxes
|
|
wire dff01_data_mux, dff23_data_mux;
|
|
|
|
// DFF clock selectors
|
|
wire dff0_clk, dff1_clk, dff2_clk, dff3_clk;
|
|
|
|
// DFF asynchronous-clear selectors
|
|
wire dff0_ac, dff1_ac, dff2_ac, dff3_ac;
|
|
|
|
// LUT, DFF and adder output wires for routing.
|
|
wire lut0_out, lut1a_out, lut1b_out, lut2_out, lut3a_out, lut3b_out;
|
|
wire dff0_out, dff1_out, dff2_out, dff3_out;
|
|
wire add0_sum, add1_sum, add0_carry, add1_carry;
|
|
|
|
generate
|
|
if (FEEDBACK0 === "FF0")
|
|
assign ff_feedback_mux0 = dff0_out;
|
|
else if (FEEDBACK0 === "FF1")
|
|
assign ff_feedback_mux0 = dff1_out;
|
|
else
|
|
$error("Invalid FEEDBACK0 setting!");
|
|
|
|
if (FEEDBACK1 == "FF2")
|
|
assign ff_feedback_mux1 = dff2_out;
|
|
else if (FEEDBACK1 == "FF3")
|
|
assign ff_feedback_mux1 = dff3_out;
|
|
else
|
|
$error("Invalid FEEDBACK1 setting!");
|
|
|
|
if (C0_MUX === "C0")
|
|
assign c0_input_mux = C0;
|
|
else if (C0_MUX === "F1")
|
|
assign c0_input_mux = F1;
|
|
else if (C0_MUX === "FEEDBACK1")
|
|
assign c0_input_mux = ff_feedback_mux1;
|
|
else
|
|
$error("Invalid C0_MUX setting!");
|
|
|
|
if (C1_MUX === "C1")
|
|
assign c1_input_mux = C1;
|
|
else if (C1_MUX === "F0")
|
|
assign c1_input_mux = F0;
|
|
else if (C1_MUX === "FEEDBACK0")
|
|
assign c1_input_mux = ff_feedback_mux0;
|
|
else
|
|
$error("Invalid C1_MUX setting!");
|
|
|
|
// F0 == VCC is LUT5
|
|
// F0 == F0 is LUT6
|
|
// F0 == FEEDBACK is unknown
|
|
if (F0_MUX === "VCC")
|
|
assign f0_input_mux = 1'b1;
|
|
else if (F0_MUX === "F0")
|
|
assign f0_input_mux = F0;
|
|
else if (F0_MUX === "FEEDBACK0")
|
|
assign f0_input_mux = ff_feedback_mux0;
|
|
else
|
|
$error("Invalid F0_MUX setting!");
|
|
|
|
// F1 == GND is LUT5
|
|
// F1 == F1 is LUT6
|
|
// F1 == FEEDBACK is unknown
|
|
if (F1_MUX === "GND")
|
|
assign f1_input_mux = 1'b0;
|
|
else if (F1_MUX === "F1")
|
|
assign f1_input_mux = F1;
|
|
else if (F1_MUX === "FEEDBACK1")
|
|
assign f1_input_mux = ff_feedback_mux1;
|
|
else
|
|
$error("Invalid F1_MUX setting!");
|
|
|
|
if (ADD_MUX === "LUT") begin
|
|
assign add0_input_mux = ~lut1_sum_mux;
|
|
assign add1_input_mux = ~lut3_sum_mux;
|
|
end else if (ADD_MUX === "SHARE") begin
|
|
assign add0_input_mux = SHAREIN;
|
|
assign add1_input_mux = lut1_comb_mux;
|
|
end else
|
|
$error("Invalid ADD_MUX setting!");
|
|
|
|
if (DFF01_DATA_MUX === "COMB")
|
|
assign dff01_data_mux = COMB0;
|
|
else if (DFF01_DATA_MUX === "SUM")
|
|
assign dff01_data_mux = SUM0;
|
|
else
|
|
$error("Invalid DFF01_DATA_MUX setting!");
|
|
|
|
if (DFF23_DATA_MUX === "COMB")
|
|
assign dff23_data_mux = COMB0;
|
|
else if (DFF23_DATA_MUX === "SUM")
|
|
assign dff23_data_mux = SUM0;
|
|
else
|
|
$error("Invalid DFF23_DATA_MUX setting!");
|
|
|
|
if (DFF0_CLK === "CLK0")
|
|
assign dff0_clk = CLK0;
|
|
else if (DFF0_CLK === "CLK1")
|
|
assign dff0_clk = CLK1;
|
|
else if (DFF0_CLK === "CLK2")
|
|
assign dff0_clk = CLK2;
|
|
else
|
|
$error("Invalid DFF0_CLK setting!");
|
|
|
|
if (DFF1_CLK === "CLK0")
|
|
assign dff1_clk = CLK0;
|
|
else if (DFF1_CLK === "CLK1")
|
|
assign dff1_clk = CLK1;
|
|
else if (DFF1_CLK === "CLK2")
|
|
assign dff1_clk = CLK2;
|
|
else
|
|
$error("Invalid DFF1_CLK setting!");
|
|
|
|
if (DFF2_CLK === "CLK0")
|
|
assign dff2_clk = CLK0;
|
|
else if (DFF2_CLK === "CLK1")
|
|
assign dff2_clk = CLK1;
|
|
else if (DFF2_CLK === "CLK2")
|
|
assign dff2_clk = CLK2;
|
|
else
|
|
$error("Invalid DFF2_CLK setting!");
|
|
|
|
if (DFF3_CLK === "CLK0")
|
|
assign dff3_clk = CLK0;
|
|
else if (DFF3_CLK === "CLK1")
|
|
assign dff3_clk = CLK1;
|
|
else if (DFF3_CLK === "CLK2")
|
|
assign dff3_clk = CLK2;
|
|
else
|
|
$error("Invalid DFF3_CLK setting!");
|
|
|
|
if (DFF0_AC === "AC0")
|
|
assign dff0_ac = AC0;
|
|
else if (DFF0_AC === "AC1")
|
|
assign dff0_ac = AC1;
|
|
else
|
|
$error("Invalid DFF0_AC setting!");
|
|
|
|
if (DFF1_AC === "AC0")
|
|
assign dff1_ac = AC0;
|
|
else if (DFF1_AC === "AC1")
|
|
assign dff1_ac = AC1;
|
|
else
|
|
$error("Invalid DFF1_AC setting!");
|
|
|
|
if (DFF2_AC === "AC0")
|
|
assign dff2_ac = AC0;
|
|
else if (DFF2_AC === "AC1")
|
|
assign dff2_ac = AC1;
|
|
else
|
|
$error("Invalid DFF2_AC setting!");
|
|
|
|
if (DFF3_AC === "AC0")
|
|
assign dff3_ac = AC0;
|
|
else if (DFF3_AC === "AC1")
|
|
assign dff3_ac = AC1;
|
|
else
|
|
$error("Invalid DFF3_AC setting!");
|
|
|
|
endgenerate
|
|
|
|
// F0 on the Quartus diagram
|
|
MISTRAL_ALUT4 #(.LUT(LUT0)) lut0 (.A(A), .B(B), .C(C0), .D(c1_input_mux), .Q(lut0_out));
|
|
|
|
// F2 on the Quartus diagram
|
|
MISTRAL_ALUT4 #(.LUT(LUT1)) lut1_comb (.A(A), .B(B), .C(C0), .D(c1_input_mux), .Q(lut1_comb_mux));
|
|
MISTRAL_ALUT4 #(.LUT(LUT1)) lut1_sum (.A(A), .B(B), .C(C0), .D(E0), .Q(lut1_sum_mux));
|
|
|
|
// F1 on the Quartus diagram
|
|
MISTRAL_ALUT4 #(.LUT(LUT2)) lut2 (.A(A), .B(B), .C(C1), .D(c0_input_mux), .Q(lut2_out));
|
|
|
|
// F3 on the Quartus diagram
|
|
MISTRAL_ALUT4 #(.LUT(LUT3)) lut3_comb (.A(A), .B(B), .C(C1), .D(c0_input_mux), .Q(lut3_comb_mux));
|
|
MISTRAL_ALUT4 #(.LUT(LUT3)) lut3_sum (.A(A), .B(B), .C(C1), .D(E1), .Q(lut3_sum_mux));
|
|
|
|
MISTRAL_FF #(.INIT(INIT0)) dff0 (.D(dff01_data_mux), .CLK(dff0_clk), .ACn(dff0_ac), .Q(dff0_out));
|
|
MISTRAL_FF #(.INIT(INIT1)) dff1 (.D(dff01_data_mux), .CLK(dff1_clk), .ACn(dff1_ac), .Q(dff1_out));
|
|
MISTRAL_FF #(.INIT(INIT2)) dff2 (.D(dff23_data_mux), .CLK(dff2_clk), .ACn(dff2_ac), .Q(dff2_out));
|
|
MISTRAL_FF #(.INIT(INIT3)) dff3 (.D(dff23_data_mux), .CLK(dff3_clk), .ACn(dff3_ac), .Q(dff3_out));
|
|
|
|
// Adders
|
|
assign {add0_carry, add0_sum} = CARRYIN + lut0_out + lut1_sum_mux;
|
|
assign {add1_carry, add1_sum} = add0_carry + lut2_out + lut3_sum_mux;
|
|
|
|
// COMBOUT outputs on the Quartus diagram
|
|
assign COMB0 = E0 ? (f0_input_mux ? lut3_comb_mux : lut1_comb_mux)
|
|
: (f0_input_mux ? lut2_out : lut0_out);
|
|
|
|
assign COMB1 = E1 ? (f1_input_mux ? lut3_comb_mux : lut1_comb_mux)
|
|
: (f1_input_mux ? lut2_out : lut0_out);
|
|
|
|
// SUMOUT output on the Quartus diagram
|
|
assign SUM0 = add0_sum;
|
|
assign SUM1 = add1_sum;
|
|
|
|
// COUT output on the Quartus diagram
|
|
assign CARRYOUT = add1_carry;
|
|
|
|
// SHAREOUT output on the Quartus diagram
|
|
assign SHAREOUT = lut3_comb_mux;
|
|
|
|
// REGOUT outputs on the Quartus diagram
|
|
assign FF0 = dff0_out;
|
|
assign FF1 = dff1_out;
|
|
assign FF2 = dff2_out;
|
|
assign FF3 = dff3_out;
|
|
|
|
endmodule
|
|
*/
|