Files
FPGA-Neural/hardware/v1/sim/neuron_parallel_saturation_bounds_tb.v
micheleandClaude Sonnet 5 dc0b331d3e feat(v2): scaffold hardware/v1 frozen baseline + M1 Neural Processor
Begins the V2 Neural Multiprocessor / Dataflow architecture per
docs/v2-description.md, per explicit user request to freeze V1 and
start V2 development, copying from V1 what's needed.

Scaffold:
- hardware/v1/: byte-exact, read-only copy of the current V1 codebase
  (rtl, testbenches, tools, constraints, a representative subset of
  synthesis results, and reference docs) -- verified identical via
  diff/cmp against the live top-level tree before being made
  filesystem-read-only. The live top-level tree is untouched and
  remains the project's "production" V1 (see hardware/v1/README.md
  and hardware/v2/logs/decisions.log DEC-0001 for why copy-not-move).
- hardware/v2/: mandatory structure (rtl/sim/constraints/synthesis/
  reports/scripts/logs/docs) plus the full logging system required by
  the spec (development/architecture/simulation/synthesis/timing/
  benchmark/decisions/experiments/errors.log).

M1 -- Neural Processor (hardware/v2/rtl/neural_processor.v):
- 8-stage pipelined perceptron unit (P_IN=8): input align, 8
  multipliers, 3-level adder tree, accumulator, bias+activation, INT8
  saturation. Genuine 1-tile/cycle throughput, not just a wider
  combinational datapath.
- 7-state FSM (NP_IDLE..NP_ERROR per docs/v2-description.md §6, with
  4 baseline states merged into NP_WAIT_OPERANDS -- see
  decisions.log DEC-0002); valid/ready/data/last stream interfaces
  per §7.
- Bit-exact vs the frozen hardware/v1/rtl/neuron_parallel.v + mac8.v
  + mac_unit.v: 7/7 tests pass (hardware/v2/sim/tb_neural_processor.v),
  covering regular/mixed-sign/extreme-INT8 vectors, both activations,
  a zero-idle-gap back-to-back-tiles throughput check, and an 8-tile
  job -- verified with Verilator (see below for why).
- Real synthesis + place&route (Yosys + nextpnr-ecp5): 0 CHECK
  problems, Fmax 183.12 MHz at ACC_WIDTH=32 (PASS at 80MHz, ~3x V1's
  isolated PARALLEL=8 Fmax of 61.71 MHz) and 176.21 MHz at ACC_WIDTH=24
  (a user-requested comparison experiment, also bit-exact-verified;
  see experiments.log EXP-0001/EXP-0002 and benchmark.log).

Three real bugs found and resolved during M1 development (full
diagnostic record in errors.log):
- Two independent, reproducible Icarus Verilog v13.0 scheduling
  defects (ERR-0001, ERR-0002) that silently produced wrong simulation
  results for standard sequential Verilog -- confirmed via Verilator
  5.050 giving correct results on the same minimal repros. Verilator
  is now the trusted simulator for hardware/v2/ (decisions.log
  DEC-0004); Icarus's affected protocol-violation check was removed
  from the RTL and deferred architecturally to the Neural Director
  (DEC-0003) rather than chased further.
- One real RTL bug (ERR-0003): last0 wasn't gated like valid0,
  letting a "last tile" tag leak into the pipeline ahead of its
  actual valid tile on back-to-back jobs. Fixed and verified.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013xXuuRUWZScuo1DeYJxs3v
2026-09-05 14:06:53 +02:00

150 lines
5.3 KiB
Verilog

`timescale 1ns/1ps
// ================================================================
// NEURON_PARALLEL SATURATION/ACTIVATION BOUNDARY TESTBENCH
//
// Timing-closure task: the saturation/ReLU comparisons in
// rtl/neuron_parallel.v were rewritten from arithmetic comparisons
// (`> 127`, `< -128`, `<= 0` -- a second 32-bit carry chain in
// series with the accumulator's own adder, the measured critical
// path) to bit-test form (wide AND/OR reductions). This testbench
// exists specifically to prove bit-exact equivalence at every
// saturation boundary called out in the task brief: final_acc =
// 126, 127, 128, 129, -128, -129, -1, 0, for BOTH ACT_NONE and
// ACT_RELU.
//
// N_INPUTS=PARALLEL=1 (single MAC lane) so final_acc = x*w + bias
// exactly, letting each test case hit its target accumulator value
// directly via a hand-picked (x, w, bias) triple -- no need to sum
// multiple groups.
// ================================================================
module tb;
localparam DATA_WIDTH = 8;
localparam N_INPUTS = 1;
localparam PARALLEL = 1;
localparam ACC_WIDTH = 32;
localparam CLK_PERIOD = 10.0;
localparam ACT_NONE = 2'd0;
localparam ACT_RELU = 2'd1;
reg clk;
initial begin
clk = 1'b0;
forever #(CLK_PERIOD / 2.0) clk = ~clk;
end
reg rst;
reg start;
reg signed [DATA_WIDTH-1:0] x_val, w_val, bias_val;
reg [1:0] activation;
wire signed [DATA_WIDTH-1:0] y;
wire busy, done;
neuron_parallel #(
.DATA_WIDTH(DATA_WIDTH),
.N_INPUTS(N_INPUTS),
.PARALLEL(PARALLEL),
.ACC_WIDTH(ACC_WIDTH)
) dut (
.clk(clk), .rst(rst), .start(start),
.x_bus(x_val), .w_bus(w_val), .bias(bias_val),
.activation(activation),
.n_inputs_real(N_INPUTS[15:0]),
.y(y), .busy(busy), .done(done)
);
integer errors;
task run_case(
input signed [DATA_WIDTH-1:0] x,
input signed [DATA_WIDTH-1:0] w,
input signed [DATA_WIDTH-1:0] b,
input [1:0] act,
input signed [DATA_WIDTH-1:0] expected,
input [255:0] name
);
begin
@(negedge clk);
x_val = x; w_val = w; bias_val = b; activation = act;
start = 1'b1;
@(negedge clk);
start = 1'b0;
wait (done);
if (y !== expected) begin
$display("FAIL %0s: x=%0d w=%0d bias=%0d act=%0d expected=%0d got=%0d",
name, x, w, b, act, expected, y);
errors = errors + 1;
end else begin
$display("PASS %0s: x=%0d w=%0d bias=%0d act=%0d -> y=%0d",
name, x, w, b, act, y);
end
@(negedge clk);
end
endtask
initial begin
errors = 0;
rst = 1'b1; start = 1'b0;
x_val = 0; w_val = 0; bias_val = 0; activation = ACT_RELU;
repeat (3) @(negedge clk);
rst = 1'b0;
@(negedge clk);
$display("");
$display("========================================");
$display("NEURON_PARALLEL SATURATION BOUNDARY TEST");
$display("========================================");
// final_acc = 126 (x=1,w=126,bias=0)
run_case(8'sd1, 8'sd126, 8'sd0, ACT_NONE, 8'sd126, "acc=126 NONE");
run_case(8'sd1, 8'sd126, 8'sd0, ACT_RELU, 8'sd126, "acc=126 RELU");
// final_acc = 127 (exact positive boundary, must NOT saturate)
run_case(8'sd1, 8'sd127, 8'sd0, ACT_NONE, 8'sd127, "acc=127 NONE");
run_case(8'sd1, 8'sd127, 8'sd0, ACT_RELU, 8'sd127, "acc=127 RELU");
// final_acc = 128 (one past positive boundary, must saturate to 127)
run_case(8'sd2, 8'sd64, 8'sd0, ACT_NONE, 8'sd127, "acc=128 NONE");
run_case(8'sd2, 8'sd64, 8'sd0, ACT_RELU, 8'sd127, "acc=128 RELU");
// final_acc = 129
run_case(8'sd2, 8'sd64, 8'sd1, ACT_NONE, 8'sd127, "acc=129 NONE");
run_case(8'sd2, 8'sd64, 8'sd1, ACT_RELU, 8'sd127, "acc=129 RELU");
// final_acc = -128 (exact negative boundary, must NOT saturate under NONE)
run_case(8'sd1, -8'sd128, 8'sd0, ACT_NONE, -8'sd128, "acc=-128 NONE");
run_case(8'sd1, -8'sd128, 8'sd0, ACT_RELU, 8'sd0, "acc=-128 RELU");
// final_acc = -129 (one past negative boundary, must saturate to -128 under NONE)
run_case(8'sd2, -8'sd64, -8'sd1, ACT_NONE, -8'sd128, "acc=-129 NONE");
run_case(8'sd2, -8'sd64, -8'sd1, ACT_RELU, 8'sd0, "acc=-129 RELU");
// final_acc = -1
run_case(8'sd1, -8'sd1, 8'sd0, ACT_NONE, -8'sd1, "acc=-1 NONE");
run_case(8'sd1, -8'sd1, 8'sd0, ACT_RELU, 8'sd0, "acc=-1 RELU");
// final_acc = 0
run_case(8'sd0, 8'sd0, 8'sd0, ACT_NONE, 8'sd0, "acc=0 NONE");
run_case(8'sd0, 8'sd0, 8'sd0, ACT_RELU, 8'sd0, "acc=0 RELU");
$display("");
if (errors == 0) begin
$display("========================================");
$display("SATURATION BOUNDARY TEST PASSED (0 errors)");
$display("========================================");
end else begin
$display("========================================");
$display("SATURATION BOUNDARY TEST FAILED (%0d errors)", errors);
$display("========================================");
$fatal;
end
$finish;
end
endmodule