feat(test-benchmark): accept float values for --gas-benchmark-values flag

docs/running_tests/releases.md

@@ -52,8 +52,8 @@ follow the same layout.
 
 When filling with `--gas-benchmark-values`, benchmark tests additionally
 include the gas limit in the subdirectory name (`for_{fork}_at_{gas}M`,
-where `{gas}` is in millions, zero-padded to four digits), with one
-subdirectory per gas value:
+where `{gas}` is in millions, zero-padded to at least four digits), with
+one subdirectory per gas value:
 
 ```text
 fixtures/

docs/writing_tests/benchmarks.md

@@ -71,7 +71,7 @@ This mode is designed for gas limit testing, and gas repricing, where it enables
 
 **Note:** For both benchmark modes, users may supply multiple values in a single invocation. For example:
 
-- `--gas-benchmark-values 1,2,3` runs the test with 1M, 2M, and 3M block gas limits
+- `--gas-benchmark-values 0.5,1,10` runs the test with 0.5M, 1M, and 10M block gas limits
 - `--fixed-opcode-count 4,5` runs the test with approximately 4K and 5K opcode executions
 
 **Output layout with gas benchmark values:** When `--gas-benchmark-values` is provided and benchmark tests are filled, fixtures are written into per‑fork, per‑gas‑limit subdirectories under each format directory:

packages/testing/src/execution_testing/cli/pytest_commands/plugins/filler/tests/test_benchmarking.py

@@ -16,6 +16,7 @@
     FORK_SUBDIR_PREFIX,
     SUBFOLDER_LEVEL_SEPARATOR,
     format_fork_subdir,
+    format_gas_limit_prefix,
 )
 
 # EVM binary for fill tests; defaults to geth evm
@@ -200,6 +201,33 @@ def test_benchmarking_mode_configured_with_option(
     assert any("benchmark-gas-value_30M" in line for line in result.outlines)
 
 
+def test_benchmarking_mode_with_float_gas_values(
+    pytester: pytest.Pytester,
+) -> None:
+    """Test that float --gas-benchmark-values are accepted and formatted."""
+    setup_test_directory_structure(
+        pytester, test_module_dummy, "test_dummy_benchmark.py"
+    )
+
+    result = pytester.runpytest(
+        "-c",
+        "pytest-fill.ini",
+        "--fork",
+        "Prague",
+        "--gas-benchmark-values",
+        "0.5,1.5,10",
+        "tests/benchmark/dummy_test_module/",
+        "--collect-only",
+        "-q",
+    )
+
+    assert result.ret == 0
+    assert any("6 tests collected" in line for line in result.outlines)
+    assert any("benchmark-gas-value_0.5M" in line for line in result.outlines)
+    assert any("benchmark-gas-value_1.5M" in line for line in result.outlines)
+    assert any("benchmark-gas-value_10M" in line for line in result.outlines)
+
+
 def test_benchmark_gas_values_split_into_subdirs(
     pytester: pytest.Pytester, tmp_path: Path
 ) -> None:
@@ -1209,3 +1237,20 @@ def test_fork_split_example(state_test, pre) -> None:
             assert "fork_Prague" not in key, (
                 f"Unexpected fork_Prague in key {key} ({file_path})"
             )
+
+
+@pytest.mark.parametrize(
+    "values,expected",
+    [
+        ([1.0, 10.0, 60.0], ["0001M", "0010M", "0060M"]),
+        ([0.5, 1.0, 10.0], ["0000.5M", "0001.0M", "0010.0M"]),
+        ([0.25, 0.5, 1.0], ["0000.25M", "0000.50M", "0001.00M"]),
+    ],
+    ids=["integers", "mixed", "fractional"],
+)
+def test_format_gas_limit_prefix(
+    values: List[float], expected: List[str]
+) -> None:
+    """Test gas limit prefix formatting for int and float values."""
+    result = [format_gas_limit_prefix(v, values) for v in values]
+    assert result == expected

packages/testing/src/execution_testing/cli/pytest_commands/plugins/shared/benchmarking.py

@@ -3,9 +3,10 @@
 import re
 from abc import ABC, abstractmethod
 from pathlib import Path
-from typing import ClassVar, Dict, List, Self
+from typing import Annotated, ClassVar, Dict, List, Self
 
 import pytest
+from annotated_types import Gt
 from pydantic import BaseModel, Field, RootModel
 
 from execution_testing.test_types import Environment, EnvironmentDefaults
@@ -33,7 +34,7 @@ def pytest_addoption(parser: pytest.Parser) -> None:
         default=None,
         help=(
             "Gas limits (in millions) for benchmark tests. "
-            "Example: '100,500' runs tests with 100M and 500M gas. "
+            "Example: '0.5,1,10' runs tests with 0.5M, 1M and 10M gas. "
             "Benchmark outputs are grouped under "
             f"{FORK_SUBDIR_PREFIX}{{fork}}"
             f"{SUBFOLDER_LEVEL_SEPARATOR}XXXXM/ subdirectories. "
@@ -174,7 +175,7 @@ def parametrize(self, metafunc: pytest.Metafunc) -> None:
 class GasBenchmarkValues(RootModel, BenchmarkParametrizer):
     """Gas benchmark values configuration object."""
 
-    root: List[int]
+    root: List[Annotated[float, Gt(0)]]
 
     flag: ClassVar[str] = "--gas-benchmark-values"
     config_field: ClassVar[str] = "_gas_benchmark_values_config"
@@ -191,8 +192,8 @@ def get_test_parameters(self, _test_name: str) -> list[ParameterSet]:
         """Get benchmark values. All tests have the same list."""
         return [
             pytest.param(
-                gas_value * 1_000_000,
-                id=f"benchmark-gas-value_{gas_value}M",
+                int(gas_value * 1_000_000),
+                id=f"benchmark-gas-value_{gas_value:g}M",
                 marks=[
                     pytest.mark.fixture_subfolder(
                         level=1,

packages/testing/src/execution_testing/cli/pytest_commands/plugins/shared/fixture_output.py

@@ -315,12 +315,22 @@ def from_config(cls, config: pytest.Config) -> "FixtureOutput":
 
 
 def format_gas_limit_prefix(
-    gas_value_millions: int, all_values_millions: list[int]
+    gas_value_millions: float, all_values_millions: list[float]
 ) -> str:
     """Return a stable, sortable gas-limit prefix for a fixture subfolder."""
     max_value = max(all_values_millions) if all_values_millions else 0
-    width = max(4, len(str(max_value)))
-    return f"{gas_value_millions:0{width}d}M"
+    int_width = max(4, len(str(int(max_value))))
+    # Find max decimal places across all fractional values for
+    # consistent padding (ensures lexicographic sortability).
+    max_decimals = 0
+    for v in all_values_millions:
+        label = f"{v:g}"
+        if "." in label:
+            max_decimals = max(max_decimals, len(label.split(".")[1]))
+    if gas_value_millions == int(gas_value_millions) and max_decimals == 0:
+        return f"{int(gas_value_millions):0{int_width}d}M"
+    total_width = int_width + 1 + max_decimals  # digits + dot + decimals
+    return f"{gas_value_millions:0{total_width}.{max_decimals}f}M"
 
 
 def format_fork_subdir(