def create_kernel_instance(self, kernel_options, params, verbose):
"""create kernel instance from kernel source, parameters, problem size, grid divisors, and so on"""
instance_string = util.get_instance_string(params)
grid_div = (kernel_options.grid_div_x, kernel_options.grid_div_y, kernel_options.grid_div_z)
#insert default block_size_names if needed
if not kernel_options.block_size_names:
kernel_options.block_size_names = util.default_block_size_names
#setup thread block and grid dimensions
threads, grid = util.setup_block_and_grid(kernel_options.problem_size, grid_div, params, kernel_options.block_size_names)
if numpy.prod(threads) > self.dev.max_threads:
if verbose:
print("skipping config", instance_string, "reason: too many threads per block")
return None
#obtain the kernel_string and prepare additional files, if any
temp_files = dict()
kernel_source = kernel_options.kernel_string
if not isinstance(kernel_source, list):
kernel_source = [kernel_source]
name, kernel_string, temp_files = util.prepare_list_of_files(kernel_options.kernel_name, kernel_source, params, grid, threads, kernel_options.block_size_names)
#collect everything we know about this instance and return it
return KernelInstance(name, kernel_string, temp_files, threads, grid, params, kernel_options.arguments)
def create_kernel_instance(self, kernel_source, kernel_options, params,
verbose):
"""create kernel instance from kernel source, parameters, problem size, grid divisors, and so on"""
grid_div = (kernel_options.grid_div_x, kernel_options.grid_div_y,
kernel_options.grid_div_z)
#insert default block_size_names if needed
if not kernel_options.block_size_names:
kernel_options.block_size_names = util.default_block_size_names
#setup thread block and grid dimensions
threads, grid = util.setup_block_and_grid(
kernel_options.problem_size, grid_div, params,
kernel_options.block_size_names)
if np.prod(threads) > self.dev.max_threads:
if verbose:
print(
f"skipping config {util.get_instance_string(params)} reason: too many threads per block"
)
return None
#obtain the kernel_string and prepare additional files, if any
name, kernel_string, temp_files = kernel_source.prepare_list_of_files(
kernel_options.kernel_name, params, grid, threads,
kernel_options.block_size_names)
#check for templated kernel
if kernel_source.lang == "CUDA" and "<" in name and ">" in name:
kernel_string, name = wrap_templated_kernel(kernel_string, name)
#collect everything we know about this instance and return it
return KernelInstance(name, kernel_source, kernel_string, temp_files,
threads, grid, params, kernel_options.arguments)
def create_kernel_instance(self, kernel_options, params, verbose):
"""create kernel instance from kernel source, parameters, problem size, grid divisors, and so on"""
instance_string = util.get_instance_string(params)
grid_div = (kernel_options.grid_div_x, kernel_options.grid_div_y,
kernel_options.grid_div_z)
#setup thread block and grid dimensions
threads, grid = util.setup_block_and_grid(
kernel_options.problem_size, grid_div, params,
kernel_options.block_size_names)
if numpy.prod(threads) > self.dev.max_threads:
if verbose:
print("skipping config", instance_string,
"reason: too many threads per block")
return None
#obtain the kernel_string and prepare additional files, if any
temp_files = dict()
original_kernel = kernel_options.kernel_string
if isinstance(original_kernel, list):
kernel_string, temp_files = util.prepare_list_of_files(
original_kernel, params, grid)
else:
kernel_string = util.get_kernel_string(original_kernel, params)
#prepare kernel_string for compilation
name, kernel_string = util.setup_kernel_strings(
kernel_options.kernel_name, kernel_string, params, grid)
#collect everything we know about this instance and return it
return KernelInstance(name, kernel_string, temp_files, threads, grid,
params, kernel_options.arguments)