Use the effective number of threads for the Scheduler

One can now use `taskset(1)` and limit the number of workers
accordingly.

kdtree/Makefile

 INCLUDE_DIRS = -I/usr/include/python3.9
 EXE = kdtree
 CXX = g++
-CPPFLAGS = -O2 -g -Wno-unused-result -Wsign-compare -pthread $(INCLUDE_DIRS)
+CPPFLAGS = -O2 -g -Wno-unused-result -Wsign-compare -pthread $(INCLUDE_DIRS) -fopenmp
 LDFLAGS += -Wl,--unresolved-symbols=ignore-all
 MACROS = -DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION
 EXT_SUFFIX := $(shell python3 -c "import sysconfig; print(sysconfig.get_config_var('EXT_SUFFIX'))")

kdtree/kdtree.pyx

@@ -8,6 +8,8 @@ np.import_array()
 from runtime.runtime cimport BatchMailBox, NullResult, Scheduler, WaitResult
 from libc.stdio cimport printf
 from libc.stdlib cimport malloc, free
+from openmp cimport omp_get_max_threads
+
 
 from cython.parallel import prange
 
@@ -479,6 +481,12 @@ cdef cypclass KDTree:
         cdef I_t n = X.shape[0]
         cdef I_t d = X.shape[1]
         cdef I_t initialised = 0
+        # Accessing _SC_NPROCESSORS_ONLN does not return the 
+        # effective number of threads which were assigned to
+        # this task using `taskset(1)` for instance. 
+        # 
+        # This OpenMP API is a workable way to access it.
+        cdef I_t num_workers = omp_get_max_threads()
 
         self._n = n
         self._d = d
@@ -491,7 +499,7 @@ cdef cypclass KDTree:
 
         # Recurvisely building the tree here
         global scheduler
-        scheduler = Scheduler()
+        scheduler = Scheduler(num_workers)
         cdef active Counter counter = consume Counter()
 
         self._root = consume Node()