Spaces:
Sleeping
Sleeping
File size: 8,181 Bytes
eb82b91 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 | # UVM Verification Framework — User Guide
## Overview
Automated UVM testbench generation from YAML / FuseSoC `.core` specifications. Outputs a complete, ready-to-compile UVM environment for protocol interfaces (UART, SPI, I2C, APB, AXI4-Lite, Wishbone) with AI/ML-powered coverage optimization.
## Architecture
```
Spec (.core / .yaml)
|
v
Generation Engine (Jinja2 + ML)
|
v
+----------------------------+
| Generated UVM Testbench |
| +----------------------+ |
| | testbench.sv | |
| | interface_{name}.sv | |
| | sequence_item_{n}.sv | |
| | driver_{name}.sv | |
| | monitor_{name}.sv | |
| | agent_{name}.sv | |
| | env_{name}.sv | |
| | scoreboard_{name}.sv | |
| | coverage_collector.sv | |
| | protocol_checker.sv | |
| | ral_model_{name}.sv | |
| | base_sequence_{n}.sv | |
| | test_{name}.sv | |
| +----------------------+ |
+----------------------------+
```
## Quick Start
### 1. Install
```bash
pip install -r requirements.txt
pip install -r requirements-dev.txt # optional: dev/lint
```
### 2. Generate a UVM testbench
```bash
# From YAML spec
python -m src.main --spec configs/uart_demo.yaml
# From FuseSoC .core spec
python -m src.main --spec configs/uart16550-1.5.core
# With auto-training (AI coverage optimization)
python -m src.main --spec configs/uart16550-1.5.core --auto-train --max-iterations 3
```
### 3. Output
```
output/{design_name}_tb/
testbench.sv # Top-level module
interface_{name}.sv # Clocking + modport
sequence_item_{name}.sv # Transaction object
driver_{name}.sv # Bus driver
monitor_{name}.sv # Bus monitor
agent_{name}.sv # Agent (sequencer + driver + monitor)
environment_{name}.sv # Env (agent + scoreboard + coverage + RAL)
scoreboard_{name}.sv # Scoreboard (TX/RX compare, error check)
coverage_collector_{name}.sv # Functional coverage groups
protocol_checker_{name}.sv # SVA assertions
ral_model_{name}.sv # RAL model + adapter + predictor
base_sequence_{name}.sv # Sequence library
test_{name}.sv # Test library
compile.f # Compile file list
sim_{name}.tcl # Simulation script
```
## Spec Format
### YAML format
```yaml
design_name: uart16550
protocol: uart
interfaces:
- name: bus
direction: slave
protocol: wishbone
signals:
- {name: addr, direction: input, width: 3}
- {name: data_in, direction: input, width: 8}
- {name: data_out, direction: output, width: 8}
- name: serial
direction: master
protocol: uart
signals:
- {name: tx, direction: output, width: 1}
- {name: rx, direction: input, width: 1}
registers:
- name: RBR
address: 0x00
access: ro
size: 8
description: Receiver Buffer Register
- name: THR
address: 0x00
access: wo
size: 8
- name: IER
address: 0x01
access: rw
size: 8
- name: LCR
address: 0x03
access: rw
size: 8
reset: 0x03
clocks:
clk: 50MHz
reset: {name: rst_n, polarity: active_low}
```
### FuseSoC .core format
See `configs/uart16550-1.5.core` for a complete example.
## UVM VIP Integration
### Packaging as a VIP
The generated output is structured as a self-contained UVM Verification IP:
```
{design_name}_vip/
pkg/
{design_name}_vip_pkg.sv # UVM package (all files)
{design_name}_reg_pkg.sv # RAL package
src/
{design_name}_if.sv # Interface
{design_name}_agent.sv # Agent
{design_name}_env.sv # Environment
{design_name}_driver.sv # Driver
{design_name}_monitor.sv # Monitor
{design_name}_scoreboard.sv # Scoreboard
{design_name}_coverage.sv # Coverage collector
{design_name}_checker.sv # Protocol checker
{design_name}_ral.sv # RAL model
{design_name}_sequences.sv # Sequence library
{design_name}_tests.sv # Test library
sim/
compile.f # Compile list
sim.do # Questa script
sim.tcl # Generic TCL script
examples/
smoke_test.sv # Standalone testbench
```
### Integration into a larger SoC environment
```systemverilog
// 1. Import the VIP
import uart_vip_pkg::*;
// 2. Instantiate interface
uart_if uart_if_inst (
.clk (sys_clk),
.rst_n (sys_rst_n),
.tx (uart_tx),
.rx (uart_rx)
);
// 3. Configure via config_db
initial begin
uvm_config_db#(virtual uart_if)::set(
null, "uvm_test_top", "vif", uart_if_inst
);
end
// 4. Run test
initial begin
run_test("uart_smoke_test");
end
```
### Configuration parameters
| Parameter | Type | Default | Description |
|-----------|------|---------|-------------|
| `protocol` | string | `"uart"` | Protocol selection |
| `model_type` | string | `"v2"` | Generation model (template / v2) |
| `rl_strategy` | string | `"ucb"` | RL exploration strategy |
| `enable_learning` | bool | true | Enable RL + coverage feedback |
| `strict_uvm` | bool | true | Generate IEEE 1800.2 compliant code |
| `max_iterations` | int | 1 | Auto-training iterations |
## AI/ML Features
### Coverage Prediction
The coverage predictor uses a 3-model ensemble (RandomForest + GradientBoosting + LinearRegression) with Ridge meta-blender to predict functional coverage gaps and suggest targeted sequences.
### Reinforcement Learning
Q-learning with:
- Double Q-learning (two independent Q-tables)
- Prioritized experience replay
- N-step returns
- Eligibility traces
- 5 exploration strategies (epsilon-greedy, softmax, UCB, Thompson, NoisyNet)
States are encoded as `{protocol}:{file_type}:{complexity}`.
### Auto-Training Loop
```
1. Generate UVM testbench
2. Run simulation (or stub)
3. Predict coverage gaps
4. Generate targeted sequences
5. Re-train RL model
6. Repeat until coverage target met
```
## Regression Management
### Running regressions
```bash
# Single test
python -m src.main --spec configs/uart16550-1.5.core \
--test smoke
# All tests
python regression/run_regression.py \
--spec configs/uart16550-1.5.core
# Multi-seed regression
python regression/run_regression.py \
--spec configs/uart16550-1.5.core \
--seeds 100 \
--tests smoke,loopback,interrupt
```
### YAML regression spec
```yaml
regression:
name: uart_full_regression
spec: configs/uart16550-1.5.core
tests:
- uart_smoke_test
- uart_reg_access_test
- uart_loopback_test
- uart_interrupt_test
- uart_fifo_test
- uart_random_test
seeds: [10, 20, 30, 50, 100]
simulator: questa
coverage: true
output: results/
```
## Simulator Support
| Simulator | Status | Notes |
|-----------|--------|-------|
| Questa/ModelSim | ✅ | Full support via .do / .tcl script |
| VCS | ✅ | Compatible (IEEE 1800.2 compliant) |
| Xcelium | ✅ | Compatible |
| Icarus Verilog | ✅ | Basic support (stub mode) |
## Project Structure
```
UVM-verification/
configs/ # Spec files (.yaml, .core)
docs/ # Documentation
frontend/ # React UI
protocols/ # Protocol definitions (UART, SPI, I2C, etc.)
regression/ # Regression scripts
src/ # Core engine
features/ # Spec feature extraction
generation/ # Template engine
templates/ # Jinja2 UVM templates
models/ # ML models (RL, coverage predictor, etc.)
evaluation/ # Quality scoring, SV checking
pipeline.py # Auto-training pipeline
vip/ # Packaged VIP files
output/ # Generated testbenches
```
## Testing
```bash
# Run unit tests
python -m pytest tests/
# Test coverage prediction
python -m pytest tests/test_coverage_predictor.py -v
# Test RL learner
python -m pytest tests/test_rl_learner.py -v
# Test pipeline end-to-end
python test_pipeline.py
```
## Docker
```bash
docker-compose up --build
# Frontend: http://localhost:7860
# API: http://localhost:8000/docs
```
## License
MIT License — see LICENSE file.
|