wip: add cmake to tests

test/Ne16Weight.py

@@ -100,7 +100,9 @@ def decode(
         return weight
 
     @staticmethod
-    def source_generate(wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter) -> None:
+    def source_generate(
+        wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter
+    ) -> None:
         assert wmem == "tcdm", f"Invalid wmem source provided: {wmem}"
         section = "PI_L1"
 

test/NeurekaV2TestConf.py

@@ -103,7 +103,8 @@ def check_valid_out_type_with_norm_quant(self) -> NeurekaV2TestConf:
     @field_validator("wmem")
     @classmethod
     def check_valid_wmem(cls, v: WmemLiteral) -> WmemLiteral:
-        _supported_wmem = ["tcdm", "sram", "mram"]
+        breakpoint()
+        _supported_wmem = ["sram", "mram"]
         assert (
             v in _supported_wmem
         ), f"Unsupported wmem {v}. Supported {_supported_wmem}."

test/NeurekaV2Weight.py

@@ -132,7 +132,9 @@ def decode(
         return weight
 
     @staticmethod
-    def source_generate(wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter) -> None:
+    def source_generate(
+        wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter
+    ) -> None:
         if wmem == "sram":
             section = '__attribute__((section(".weightmem_sram")))'
         elif wmem == "mram":

test/NeurekaWeight.py

@@ -158,7 +158,9 @@ def decode(
         return weight
 
     @staticmethod
-    def source_generate(wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter) -> None:
+    def source_generate(
+        wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter
+    ) -> None:
         if wmem == "sram":
             section = '__attribute__((section(".weightmem_sram")))'
         elif wmem == "mram":

test/NnxTestClasses.py

@@ -21,12 +21,12 @@
 import os
 from abc import ABC, abstractmethod
 from enum import Enum
-from typing import Callable, Literal, Optional, Set, Tuple, Type, Union
+from typing import Literal, Optional, Set, Tuple, Type, Union
 
 import numpy as np
 import numpy.typing as npt
 import torch
-from pydantic import BaseModel, PositiveInt, field_validator, model_validator
+from pydantic import BaseModel, PositiveInt, model_validator
 
 from HeaderWriter import HeaderWriter
 from NeuralEngineFunctionalModel import NeuralEngineFunctionalModel
@@ -226,19 +226,25 @@ def incr_generator():
                 x += 1
                 if x > _type.max:
                     x = 0
-        return torch.from_numpy(
-            np.fromiter(incr_generator(), count=np.prod(shape), dtype=np.int64)
-        ).reshape(
-            (shape[0], shape[2], shape[3], shape[1])
-        ).permute((0, 3, 1, 2)).type(torch.int64)
+
+        return (
+            torch.from_numpy(
+                np.fromiter(incr_generator(), count=np.prod(shape), dtype=np.int64)
+            )
+            .reshape((shape[0], shape[2], shape[3], shape[1]))
+            .permute((0, 3, 1, 2))
+            .type(torch.int64)
+        )
 
     class DataGenerationMethod(Enum):
         RANDOM = 0
         ONES = 1
         INCREMENTED = 2
 
     @staticmethod
-    def _generate_data(_type: IntegerType, shape: Tuple, method: NnxTestGenerator.DataGenerationMethod):
+    def _generate_data(
+        _type: IntegerType, shape: Tuple, method: NnxTestGenerator.DataGenerationMethod
+    ):
         if method == NnxTestGenerator.DataGenerationMethod.RANDOM:
             return NnxTestGenerator._generate_random(_type, shape)
         elif method == NnxTestGenerator.DataGenerationMethod.ONES:
@@ -287,13 +293,15 @@ def from_conf(
             if scale is None:
                 assert conf.scale_type is not None
                 scale = NnxTestGenerator._generate_data(
-                    conf.scale_type, shape=scale_shape,
+                    conf.scale_type,
+                    shape=scale_shape,
                     method=data_generation_method,
                 )
             if conf.has_bias and bias is None:
                 assert conf.bias_type is not None
                 bias = NnxTestGenerator._generate_data(
-                    conf.bias_type, shape=bias_shape,
+                    conf.bias_type,
+                    shape=bias_shape,
                     method=data_generation_method,
                 ).type(torch.int32)
             if global_shift is None:
@@ -342,9 +350,7 @@ class NnxWeight(ABC):
     @staticmethod
     @abstractmethod
     def encode(
-        weight: npt.NDArray[np.uint8],
-        bits: int,
-        depthwise: bool = False
+        weight: npt.NDArray[np.uint8], bits: int, depthwise: bool = False
     ) -> npt.NDArray[np.uint8]:
         """Unroll weight into expected memory format
 
@@ -368,9 +374,7 @@ def decode(
     @staticmethod
     @abstractmethod
     def source_generate(
-        wmem: WmemLiteral,
-        init: npt.NDArray[np.uint8],
-        header_writer: HeaderWriter
+        wmem: WmemLiteral, init: npt.NDArray[np.uint8], header_writer: HeaderWriter
     ) -> None:
         """Function implementing generation of weight's sources"""
         ...
@@ -429,7 +433,9 @@ def generate(self, test_name: str, test: NnxTest):
             test.conf.depthwise,
         )
 
-        self.nnxWeightCls.source_generate(test.conf.wmem, weight_init, self.header_writer)
+        self.nnxWeightCls.source_generate(
+            test.conf.wmem, weight_init, self.header_writer
+        )
 
         # Render scale
         if test.scale is not None:

test/conftest.py

@@ -17,6 +17,7 @@
 # SPDX-License-Identifier: Apache-2.0
 
 import os
+import subprocess
 from typing import Union
 
 import pydantic
@@ -28,7 +29,7 @@
 from NeurekaV2TestConf import NeurekaV2TestConf
 from NeurekaV2Weight import NeurekaV2Weight
 from NeurekaWeight import NeurekaWeight
-from NnxTestClasses import NnxTest, NnxTestGenerator, NnxWeight
+from NnxTestClasses import NnxTest, NnxTestGenerator
 
 _SUPPORTED_ACCELERATORS = ["ne16", "neureka", "neureka_v2"]
 
@@ -70,6 +71,13 @@ def pytest_addoption(parser):
         default=120,
         help="Execution timeout in seconds. Default: 120s",
     )
+    parser.addoption(
+        "--build-flow",
+        dest="build_flow",
+        choices=["make", "cmake"],
+        default="make",
+        help="Choose the build flow. Default: make",
+    )
 
 
 def _find_test_dirs(path: Union[str, os.PathLike]):
@@ -82,6 +90,7 @@ def pytest_generate_tests(metafunc):
     regenerate = metafunc.config.getoption("regenerate")
     timeout = metafunc.config.getoption("timeout")
     nnxName = metafunc.config.getoption("accelerator")
+    build_flow = metafunc.config.getoption("build_flow")
 
     if nnxName == "ne16":
         nnxWeightCls = Ne16Weight
@@ -127,7 +136,16 @@ def pytest_generate_tests(metafunc):
                 )
             )
 
+    if build_flow == "cmake":
+        os.makedirs("app/build/gvsoc_workdir", exist_ok=True)
+        assert "GVSOC" in os.environ, "The GVSOC environment variable is not set."
+        subprocess.run(
+            f"cmake -Sapp -Bapp/build -GNinja -DCMAKE_TOOLCHAIN_FILE=cmake/toolchain_llvm.cmake -DACCELERATOR={nnxName}".split(),
+            check=True,
+        )
+
     metafunc.parametrize("nnxTestAndName", nnxTestAndNames)
     metafunc.parametrize("timeout", [timeout])
     metafunc.parametrize("nnxName", [nnxName])
     metafunc.parametrize("nnxWeightCls", [nnxWeightCls])
+    metafunc.parametrize("build_flow", [build_flow])

test/requirements.txt

@@ -3,3 +3,4 @@ pydantic
 pytest
 pytorch==1.11.0
 toml
+ninja

test/test.py

@@ -21,7 +21,7 @@
 import re
 import subprocess
 from pathlib import Path
-from typing import Dict, Optional, Tuple, Type, Union
+from typing import Dict, Literal, Optional, Tuple, Type, Union
 
 from NnxTestClasses import NnxTest, NnxTestConf, NnxTestHeaderGenerator, NnxWeight
 
@@ -93,12 +93,11 @@ def execute_command(
     return status, msg, stdout, stderr
 
 
-def assert_message(
-    msg: str, test_name: str, cmd: str, stdout: str, stderr: Optional[str] = None
-):
+def assert_message(msg: str, test_name: str, stdout: str, stderr: Optional[str] = None):
     retval = (
         f"Test {test_name} failed: {msg}\n"
-        f"Command: {cmd}\n" + HORIZONTAL_LINE + f"\nCaptured stdout:\n{stdout}\n"
+        + HORIZONTAL_LINE
+        + f"\nCaptured stdout:\n{stdout}\n"
     )
 
     if stderr is not None:
@@ -107,35 +106,58 @@ def assert_message(
     return retval
 
 
+def build(nnxName: str, flow: Literal["make", "cmake"]) -> None:
+    env = os.environ
+
+    if flow == "make":
+        cmd = "make -C app all platform=gvsoc"
+        env["ACCELERATOR"] = nnxName
+    elif flow == "cmake":
+        cmd = "cmake --build app/build"
+
+    subprocess.run(cmd.split(), check=True, capture_output=True, text=True, env=env)
+
+
+def run(nnxName: str, flow: Literal["make", "cmake"]) -> str:
+    env = os.environ
+
+    if flow == "make":
+        cmd = "make -C app run platform=gvsoc"
+        env["ACCELERATOR"] = nnxName
+    elif flow == "cmake":
+        bin = os.path.abspath("app/build/test-pulp-nnx")
+        gvsoc = os.environ["GVSOC"]
+        cmd = f"{gvsoc} --binary {bin} --work-dir app/build/gvsoc_workdir --target siracusa image flash run"
+
+    proc = subprocess.run(
+        cmd.split(), check=True, capture_output=True, text=True, env=env
+    )
+
+    return proc.stdout
+
+
 def test(
     nnxTestAndName: Tuple[NnxTest, str],
     timeout: int,
     nnxName: str,
     nnxWeightCls: Type[NnxWeight],
+    build_flow: Literal["cmake", "make"],
 ):
     nnxTest, nnxTestName = nnxTestAndName
-    NnxTestHeaderGenerator(nnxWeightCls).generate(
-        nnxTestName, nnxTest
-    )
-
-    Path("app/src/nnx_layer.c").touch()
-    cmd = f"make -C app all run platform=gvsoc"
-    passed, msg, stdout, stderr = execute_command(
-        cmd=cmd, timeout=timeout, envflags={"ACCELERATOR": nnxName}
-    )
+    NnxTestHeaderGenerator(nnxWeightCls).generate(nnxTestName, nnxTest)
 
-    assert passed, assert_message(msg, nnxTestName, cmd, stdout, stderr)
+    build(nnxName, build_flow)
+    stdout = run(nnxName, build_flow)
 
     match_success = re.search(r"> Success! No errors found.", stdout)
     match_fail = re.search(r"> Failure! Found (\d*)/(\d*) errors.", stdout)
 
     assert match_success or match_fail, assert_message(
-        "No regexes matched.", nnxTestName, cmd, stdout
+        "No regexes matched.", nnxTestName, stdout
     )
 
     assert not match_fail, assert_message(
         f"Errors found: {match_fail.group(1)}/{match_fail.group(2)}",
         nnxTestName,
-        cmd,
         stdout,
     )

test/testgen.py

@@ -19,13 +19,10 @@
 import argparse
 import json
 import os
-from typing import Callable, Optional, Set, Type, Union
+from typing import Optional, Set, Type, Union
 
-import numpy as np
-import numpy.typing as npt
 import toml
 
-from HeaderWriter import HeaderWriter
 from Ne16TestConf import Ne16TestConf
 from Ne16Weight import Ne16Weight
 from NeurekaTestConf import NeurekaTestConf
@@ -38,7 +35,6 @@
     NnxTestGenerator,
     NnxTestHeaderGenerator,
     NnxWeight,
-    WmemLiteral,
 )