xobjects 0.5.12__tar.gz → 0.5.14__tar.gz

{xobjects-0.5.12/xobjects.egg-info → xobjects-0.5.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xobjects
-Version: 0.5.12
+Version: 0.5.14
 Summary: In-memory serialization and code generator for CPU and GPU
 Home-page: https://xsuite.readthedocs.io/
 Download-URL: https://pypi.python.org/pypi/xobjects

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_capi.py

@@ -179,9 +179,7 @@ def test_array_get_shape(test_context, array_type):
             *out_nd = ARRAY_TYPE_nd(arr);
             ARRAY_TYPE_shape(arr, out_shape);
         }
-    """.replace(
-        "ARRAY_TYPE", array_type.__name__
-    )
+    """.replace("ARRAY_TYPE", array_type.__name__)
 
     kernels = {
         "get_nd_and_shape": xo.Kernel(

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_common.py

@@ -87,8 +87,7 @@ def test_atomic(overload, ctype, test_context):
 
     class TestAtomic(xo.Struct):
         val = ctype
-        _extra_c_sources = [
-            f"""
+        _extra_c_sources = [f"""
             #include "xobjects/headers/common.h"
             #include "xobjects/headers/atomicadd.h"
 
@@ -101,8 +100,7 @@ def test_atomic(overload, ctype, test_context):
                     retvals[ii] = ret;
                 END_VECTORIZE;
             }}
-            """
-        ]
+            """]
 
     kernels = {
         "run_atomic_test": xo.Kernel(

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_kernel.py

@@ -216,15 +216,13 @@ def test_kernels_save_load_with_classes(tmp_path, mocker):
             "x": xo.Float64,
             "y": xo.Float64,
         }
-        _extra_c_sources = [
-            """
+        _extra_c_sources = ["""
             /*gpufun*/ double myfun(TestClassData tc){
                 double x = TestClassData_get_x(tc);
                 double y = TestClassData_get_y(tc);
                 return x * y;
             }
-        """
-        ]
+        """]
         _kernels = {
             "myfun": xo.Kernel(
                 args=[

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_ref.py

@@ -220,29 +220,25 @@ def test_unionref(test_context):
         b = xo.Float64
         h = xo.Float64
 
-        _extra_c_sources = [
-            """
+        _extra_c_sources = ["""
             /*gpufun*/
             double Triangle_compute_area(Triangle tr, double scale){
                 double b = Triangle_get_b(tr);
                 double h = Triangle_get_h(tr);
                 return 0.5*b*h*scale;
             }
-            """
-        ]
+            """]
 
     class Square(xo.Struct):
         a = xo.Float64
 
-        _extra_c_sources = [
-            """
+        _extra_c_sources = ["""
             /*gpufun*/
             double Square_compute_area(Square sq, double scale){
                 double a = Square_get_a(sq);
                 return a*a*scale;
             }
-            """
-        ]
+            """]
 
     class Base(xo.UnionRef):
         _reftypes = (Triangle, Square)
@@ -259,8 +255,7 @@ def test_unionref(test_context):
         height = xo.Float64
         volume = xo.Float64
 
-        _extra_c_sources = [
-            """
+        _extra_c_sources = ["""
             /*gpukern*/
             void Prism_compute_volume(Prism pr){
                 Base base = Prism_getp_base(pr);
@@ -269,8 +264,7 @@ def test_unionref(test_context):
                 printf("base_area = %e", base_area);
                 Prism_set_volume(pr, base_area*height);
             }
-            """
-        ]
+            """]
 
     test_context.add_kernels(
         kernels={

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_shared_memory.py

@@ -35,8 +35,7 @@ def test_shared_memory():
 
         class TestElement(xo.HybridClass):
             _xofields = {}
-            _extra_c_sources = [
-                """
+            _extra_c_sources = ["""
                 __global__ void test_shared_memory(const double* input_arr, double* result, const int n) {
                   // simple kernel to test shared memory
                   // reduction with an array of 4 doubles using 2 blocks each 2 threads
@@ -47,7 +46,7 @@ def test_shared_memory():
                   unsigned int gid = blockIdx.x*blockDim.x + threadIdx.x;  // global thread ID: 0,1,2,3
 
                   // init shared memory with chunk of input array
-                  extern __shared__ double sdata[2];
+                  extern __shared__ double sdata[];
                   sdata[tid] = input_arr[gid];
                   __syncthreads();
 
@@ -59,8 +58,7 @@ def test_shared_memory():
                     atomicAdd(&result[tid], sdata[tid]);
                   }
                 }
-                """
-            ]
+                """]
             _kernels = _test_shared_memory_kernels
 
             def __init__(

{xobjects-0.5.12 → xobjects-0.5.14}/tests/test_struct.py

@@ -222,8 +222,7 @@ def test_kernel_namings():
         var_mult_3 = xo.Float64[:]
         var_mult_4 = xo.Float64[:]
 
-        _extra_c_sources = [
-            r"""
+        _extra_c_sources = [r"""
 double mul(MyStruct stru) {
     int32_t n = MyStruct_get_n(stru);
     double* var_mult_1 = MyStruct_getp1_var_mult_1(stru, 0);
@@ -258,8 +257,7 @@ double mult_four(MyStruct stru) {
         y+= var_mult_1[tid] * var_mult_2[tid] * var_mult_3[tid] * var_mult_4[tid];
         }
     return y;
-    }"""
-        ]
+    }"""]
 
     kernel_descriptions = {
         "mul": xo.Kernel(
@@ -316,15 +314,13 @@ def test_compile_kernels_only_if_needed(tmp_path, mocker):
             "x": xo.Float64,
             "y": xo.Float64,
         }
-        _extra_c_sources = [
-            """
+        _extra_c_sources = ["""
             /*gpufun*/ double myfun(TestClassData tc){
                 double x = TestClassData_get_x(tc);
                 double y = TestClassData_get_y(tc);
                 return x * y;
             }
-        """
-        ]
+        """]
         _kernels = {
             "myfun": xo.Kernel(
                 args=[

xobjects-0.5.14/xobjects/_version.py

@@ -0,0 +1 @@
+__version__ = "0.5.14"

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/capi.py

@@ -525,14 +525,12 @@ def gen_method_switch(cls, path, conf, method):
         for arg in kernel.args[1:]:
             targs.append(f"{arg.name}")
         targs = ",".join(targs)
-        lst.append(
-            f"""\
+        lst.append(f"""\
         #ifndef {refname.upper()}_SKIP_{atname.upper()}
         case {refname}_{atname}_t:
             return {atname}_{method.c_name}({targs});
             break;
-        #endif"""
-        )
+        #endif""")
     lst.append("  }")
     lst.append(f"  return{'' if method.ret is None else ' 0'};")
     lst.append("}")

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/context_cpu.py

@@ -316,6 +316,12 @@ class ContextCpu(XContext):
             if _forbid_compile:
                 raise RuntimeError("Compilation is forbidden")
 
+            if os.environ.get('XOBJECTS_FORBID_COMPILE'):
+                raise RuntimeError(
+                    "Compilation is forbidden by the environment variable "
+                    "XOBJECTS_FORBID_COMPILE"
+                )
+
             so_file = self.compile_kernel(
                 module_name,
                 kernel_descriptions,
@@ -376,8 +382,9 @@ class ContextCpu(XContext):
         )
         out_kernels = {}
         for pyname, kernel_desc in kernel_descriptions.items():
+            c_name = kernel_desc.c_name or pyname
             out_kernels[pyname] = KernelCpu(
-                function=getattr(module.lib, kernel_desc.c_name),
+                function=getattr(module.lib, c_name),
                 description=kernel_desc,
                 ffi_interface=module.ffi,
                 context=self,
@@ -459,14 +466,15 @@ class ContextCpu(XContext):
             return Path(output_file)
         finally:
             # Clean temp files
-            files_to_remove = [
-                module_name + ".c",
-                module_name + ".o",
-            ]
-
-            for ff in files_to_remove:
-                if os.path.exists(ff):
-                    os.remove(ff)
+            if 'XOBJECTS_KEEP_BUILD_FILES' not in os.environ:
+                files_to_remove = [
+                    module_name + ".c",
+                    module_name + ".o",
+                ]
+
+                for ff in files_to_remove:
+                    if os.path.exists(ff):
+                        os.remove(ff)
 
     def _build_sources(
         self,

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/context_cupy.py

@@ -350,19 +350,24 @@ if _enabled:
             return cupy.ndarray.__invert__(self._as_cupy())
 
 
-cudaheader: List[SourceType] = [
-    """\
-typedef signed long long   int64_t;  //only_for_context cuda
+cudaheader: List[SourceType] = ["""\
 typedef signed int         int32_t;  //only_for_context cuda
 typedef signed short       int16_t;  //only_for_context cuda
 typedef signed char        int8_t;   //only_for_context cuda
-typedef unsigned long long uint64_t; //only_for_context cuda
 typedef unsigned int       uint32_t; //only_for_context cuda
 typedef unsigned short     uint16_t; //only_for_context cuda
 typedef unsigned char      uint8_t;  //only_for_context cuda
 
-"""
-]
+#if defined(__CUDACC__) || defined(__HIPCC_RTC__)
+typedef signed long long   int64_t;
+typedef unsigned long long uint64_t;
+#endif
+
+#ifndef NULL
+  #define NULL nullptr
+#endif
+
+"""]
 
 
 def nplike_to_cupy(arr):

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/context_pyopencl.py

@@ -44,8 +44,7 @@ except ImportError:
 
 from ._patch_pyopencl_array import _patch_pyopencl_array
 
-openclheader: List[SourceType] = [
-    """\
+openclheader: List[SourceType] = ["""\
 #ifndef XOBJ_STDINT
 typedef long           int64_t;
 typedef int            int32_t;
@@ -59,8 +58,7 @@ typedef unsigned char  uint8_t;
 #ifndef NULL
 #define NULL 0L
 #endif
-"""
-]
+"""]
 
 if _enabled:
     # order of base classes matters as it defines which __setitem__ is used

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/headers/atomicadd.h

@@ -101,8 +101,9 @@ DEF_ATOMIC_ADD(double  , f64)
 // -------------------------------------------
 #if defined(XO_CONTEXT_CUDA)
     // CUDA compiler may not have <stdint.h>, so define the types if needed.
-    #ifdef __CUDACC_RTC__
-        // NVRTC (CuPy RawModule default) can’t see <stdint.h>, so detect it via __CUDACC_RTC__
+    #if defined(__CUDACC_RTC__) || defined(__HIPCC_RTC__)
+        // NVRTC and HIPRTC (CuPy RawModule default) can’t see <stdint.h> 
+        // We detect via __CUDACC_RTC__ (Nvidia) or __HIPCC_RTC__ (ROCm)
         typedef signed char        int8_t;
         typedef short              int16_t;
         typedef int                int32_t;
@@ -111,6 +112,14 @@ DEF_ATOMIC_ADD(double  , f64)
         typedef unsigned short     uint16_t;
         typedef unsigned int       uint32_t;
         typedef unsigned long long uint64_t;
+    #elif defined(__HIPCC__) && !defined(__HIPCC_RTC__)
+        // ROCm-HIPCC compiler appears to have definitions for 64-bit int types
+        typedef signed char        int8_t;
+        typedef short              int16_t;
+        typedef int                int32_t;
+        typedef unsigned char      uint8_t;
+        typedef unsigned short     uint16_t;
+        typedef unsigned int       uint32_t;
     #else
         // Alternatively, NVCC path is fine with host headers
         #include <stdint.h>

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/scalar.py

@@ -14,7 +14,6 @@ import numpy as np
 import logging
 from .typeutils import Info
 
-
 log = logging.getLogger(__name__)
 
 

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/struct.py

@@ -45,6 +45,7 @@ Field instance:
 
 
 """
+
 import logging
 from typing import Callable, Optional
 
@@ -475,6 +476,32 @@ class Struct(metaclass=MetaStruct):
         extra_classes=(),
         extra_compile_args=(),
     ):
+
+        if context.allow_prebuilt_kernels:
+            _print_state = Print.suppress
+            Print.suppress = True
+            try:
+                from xsuite import (
+                    get_suitable_kernel,
+                    XSK_PREBUILT_KERNELS_LOCATION,
+                )
+                kernel_info = get_suitable_kernel(
+                    config={},
+                    tracker_element_classes=[],
+                    classes=list(extra_classes) + [cls],
+                )
+            except ImportError:
+                kernel_info = None
+
+            Print.suppress = _print_state
+            if kernel_info:
+                kernels = context.kernels_from_file(
+                    module_name=kernel_info['module_name'],
+                    containing_dir=XSK_PREBUILT_KERNELS_LOCATION,
+                    kernel_descriptions=cls._kernels,
+                )
+                context.kernels.update(kernels)
+
         if only_if_needed:
             all_found = True
             for kk, kernel_description in cls._kernels.items():

{xobjects-0.5.12 → xobjects-0.5.14}/xobjects/test_helpers.py

@@ -5,6 +5,7 @@
 
 from functools import wraps
 from typing import Callable, Iterable, Union
+import os
 
 import pytest
 
@@ -105,3 +106,10 @@ def fix_random_seed(seed: int):
         return wrapper
 
     return decorator
+
+def skip_if_forbid_compile():
+    if os.environ.get('XOBJECTS_FORBID_COMPILE'):
+        pytest.skip(
+            "Compilation is forbidden by the environment variable "
+            "XOBJECTS_FORBID_COMPILE"
+        )

{xobjects-0.5.12 → xobjects-0.5.14/xobjects.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xobjects
-Version: 0.5.12
+Version: 0.5.14
 Summary: In-memory serialization and code generator for CPU and GPU
 Home-page: https://xsuite.readthedocs.io/
 Download-URL: https://pypi.python.org/pypi/xobjects

xobjects-0.5.12/xobjects/_version.py

@@ -1 +0,0 @@
-__version__ = "0.5.12"