Merge remote-tracking branch 'origin/master' into parallel_runner

Author: stijnh

doc/requirements.txt

@@ -34,7 +34,7 @@ matplotlib-inline==0.1.7 ; python_version >= "3.9" and python_version < "3.15"
 mistune==3.1.2 ; python_version >= "3.9" and python_version < "3.15"
 natsort==8.4.0 ; python_version >= "3.9" and python_version < "3.15"
 nbclient==0.10.2 ; python_version >= "3.9" and python_version < "3.15"
-nbconvert==7.16.6 ; python_version >= "3.9" and python_version < "3.15"
+nbconvert==7.17.0 ; python_version >= "3.9" and python_version < "3.15"
 nbformat==5.10.4 ; python_version >= "3.9" and python_version < "3.15"
 nbsphinx==0.9.7 ; python_version >= "3.9" and python_version < "3.15"
 numpy==1.26.4 ; python_version >= "3.9" and python_version < "3.15"
@@ -49,15 +49,15 @@ prompt-toolkit==3.0.50 ; python_version >= "3.9" and python_version < "3.15"
 ptyprocess==0.7.0 ; python_version >= "3.9" and python_version < "3.15" and sys_platform != "win32"
 pure-eval==0.2.3 ; python_version >= "3.9" and python_version < "3.15"
 pycparser==2.22 ; python_version >= "3.9" and python_version < "3.15" and implementation_name == "pypy"
-pygments==2.19.1 ; python_version >= "3.9" and python_version < "3.15"
+pygments==2.20.0 ; python_version >= "3.9" and python_version < "3.15"
 pytest==8.3.5 ; python_version >= "3.9" and python_version < "3.15"
 python-constraint2==2.1.0 ; python_version >= "3.9" and python_version < "3.15"
 python-dateutil==2.9.0.post0 ; python_version >= "3.9" and python_version < "3.15"
 pytz==2025.1 ; python_version >= "3.9" and python_version < "3.15"
 pywin32==308 ; sys_platform == "win32" and platform_python_implementation != "PyPy" and python_version >= "3.9" and python_version < "3.15"
 pyzmq==26.2.1 ; python_version >= "3.9" and python_version < "3.15"
 referencing==0.36.2 ; python_version >= "3.9" and python_version < "3.15"
-requests==2.32.4 ; python_version >= "3.9" and python_version < "3.15"
+requests==2.33.0 ; python_version >= "3.9" and python_version < "3.15"
 rpds-py==0.23.1 ; python_version >= "3.9" and python_version < "3.15"
 scikit-learn==1.6.1 ; python_version >= "3.9" and python_version < "3.15"
 scipy==1.13.1 ; python_version >= "3.9" and python_version < "3.15"
@@ -78,7 +78,7 @@ stack-data==0.6.3 ; python_version >= "3.9" and python_version < "3.15"
 threadpoolctl==3.5.0 ; python_version >= "3.9" and python_version < "3.15"
 tinycss2==1.4.0 ; python_version >= "3.9" and python_version < "3.15"
 tomli==2.2.1 ; python_version >= "3.9" and python_version < "3.15"
-tornado==6.5.1 ; python_version >= "3.9" and python_version < "3.15"
+tornado==6.5.5 ; python_version >= "3.9" and python_version < "3.15"
 traitlets==5.14.3 ; python_version >= "3.9" and python_version < "3.15"
 typing-extensions==4.12.2 ; python_version >= "3.9" and python_version < "3.15"
 tzdata==2025.1 ; python_version >= "3.9" and python_version < "3.15"

doc/requirements_test.txt

@@ -311,9 +311,9 @@ ptyprocess==0.7.0 ; python_version >= "3.10" and python_version < "4" and (os_na
 pure-eval==0.2.3 ; python_version >= "3.10" and python_version < "4" \
     --hash=sha256:1db8e35b67b3d218d818ae653e27f06c3aa420901fa7b081ca98cbedc874e0d0 \
     --hash=sha256:5f4e983f40564c576c7c8635ae88db5956bb2229d7e9237d03b3c0b0190eaf42
-pygments==2.19.1 ; python_version >= "3.10" and python_version < "4" \
-    --hash=sha256:61c16d2a8576dc0649d9f39e089b5f02bcd27fba10d8fb4dcc28173f7a45151f \
-    --hash=sha256:9ea1544ad55cecf4b8242fab6dd35a93bbce657034b0611ee383099054ab6d8c
+pygments==2.20.0 ; python_version >= "3.10" and python_version < "4" \
+    --hash=sha256:6757cd03768053ff99f3039c1a36d6c0aa0b263438fcab17520b30a303a82b5f \
+    --hash=sha256:81a9e26dd42fd28a23a2d169d86d7ac03b46e2f8b59ed4698fb4785f946d0176
 pyproject-hooks==1.2.0 ; python_version >= "3.10" and python_version < "4" \
     --hash=sha256:1e859bd5c40fae9448642dd871adf459e5e2084186e8d2c2a79a824c970da1f8 \
     --hash=sha256:9e5c6bfa8dcc30091c74b0cf803c81fdd29d94f01992a7707bc97babb1141913

kernel_tuner/accuracy.py

@@ -96,7 +96,7 @@ def _find_bfloat16_if_available():
             + "please install either the package `ml_dtypes`, `jax`, or `tensorflow`"
         )
 
-    return None
+    return dtype
 
 
 def _to_float_dtype(x: str) -> np.dtype:

kernel_tuner/strategies/dual_annealing.py

@@ -1,5 +1,6 @@
 """The strategy that uses the dual annealing optimization method."""
 import scipy.optimize
+import numpy as np
 
 from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
@@ -16,7 +17,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     method, max_fevals = common.get_options(tuning_options.strategy_options, _options)
 
     #scale variables in x to make 'eps' relevant for multiple variables
-    cost_func = CostFunc(searchspace, tuning_options, runner, scaling=True)
+    cost_func = CostFunc(searchspace, tuning_options, runner, scaling=True, invalid_value=np.inf)
 
     bounds, x0, _ = cost_func.get_bounds_x0_eps()
 

kernel_tuner/strategies/simulated_annealing.py

@@ -96,7 +96,7 @@ def acceptance_prob(old_cost, new_cost, T):
     # if start pos is not valid, always move
     if isinstance(old_cost, ErrorConfig):
         res = 1.0
-    # if we have found a valid ps before, never move to nonvalid pos
+    # if we have found a valid pos before, never move to nonvalid pos
     elif isinstance(new_cost, ErrorConfig):
         res = 0.0
     # always move if new cost is better
@@ -108,7 +108,7 @@ def acceptance_prob(old_cost, new_cost, T):
         abs_diff = old_cost - new_cost
 
         # relative to abs(old_cost), as the cost might be negative
-        rel_diff = abs_diff / np.abs(old_cost)
+        rel_diff = abs_diff / (np.abs(old_cost) if old_cost != 0.0 else 1e-20)
 
         # exponential decay
         res = np.exp(rel_diff / T)

kernel_tuner/utils/nvcuda.py

@@ -56,6 +56,10 @@ def cuda_error_check(error):
         if error != nvrtc.nvrtcResult.NVRTC_SUCCESS:
             _, desc = nvrtc.nvrtcGetErrorString(error)
             raise RuntimeError(f"NVRTC error: {desc.decode()}")
+    elif isinstance(error, tuple) and len(error) > 0:
+        cuda_error_check(error[0])
+    else:
+        raise RuntimeError(f"unknown error type returned by CUDA: {error!r} (type: {type(error).__name__})")
 
 
 def to_valid_nvrtc_gpu_arch_cc(compute_capability: str) -> str: