Functions
	generate_submission_script_fragment (bool use_cpp, int patch_size, list[str] number_of_patches, list[str] passes, str output_directory)

	generate_compilation_script_fragment (bool use_cpp, int patch_size, list[int] number_of_patches, list[str] passes, str output_directory)

	generate_running_script_fragment (bool use_cpp, int patch_size, list[int] number_of_patches, list[str] passes, str output_directory)

Variables
str	application = "euler"

str	solver = "fd4"

str	machine = "grace_hopper"

int	runtime_hours = 24

list	patch_sizes = [3, 6, 16]

int	number_of_samples = 10

dict	numbers_of_patches

str	partition = "gracehopper"

str	account = "do016"

str	set_environment = f"source {Path.home()}/set_env_arm.sh"

dict	optional_passes

str	output_name = f"{application}_kernel_benchmark"

str	job_name = f"{application}_kernel_benchmark"

str	output_root = f"{Path.home()}/Peano/benchmarks/exahype2/kernel-benchmarks/"

str	script_name = f"submit_kernel_benchmark-gpu-parallel-{application}-{solver}.sh"

str	compile_script_name = f"compile_kernel_benchmark-gpu-parallel-{application}-{solver}.sh"

str	run_script_name = f"run_kernel_benchmark-gpu-parallel-{application}-{solver}.sh"

str	benchmark_script_name = "kernel-benchmarks-fd4-dsl.py"

str	benchmark_app_name = "KernelBenchmarksFD4DSL.Release"

list	compile_scripts = []

list	run_scripts = []

list	submission_scripts = []

dict	number_of_patches = numbers_of_patches[patch_size]

list	pass_list = []

str	output_directory = output_root + f"results/{application}/{solver}/gpu_parallel/{machine}/cpp"

Function Documentation

◆ generate_compilation_script_fragment()

generate-submission-script-gpu-parallel.generate_compilation_script_fragment	(	bool	use_cpp,
		int	patch_size,
		list[int]	number_of_patches,
		list[str]	passes,
		str	output_directory
	)

Definition at line 61 of file generate-submission-script-gpu-parallel.py.

◆ generate_running_script_fragment()

generate-submission-script-gpu-parallel.generate_running_script_fragment	(	bool	use_cpp,
		int	patch_size,
		list[int]	number_of_patches,
		list[str]	passes,
		str	output_directory
	)

Definition at line 76 of file generate-submission-script-gpu-parallel.py.

◆ generate_submission_script_fragment()

generate-submission-script-gpu-parallel.generate_submission_script_fragment	(	bool	use_cpp,
		int	patch_size,
		list[str]	number_of_patches,
		list[str]	passes,
		str	output_directory
	)

Definition at line 50 of file generate-submission-script-gpu-parallel.py.

Variable Documentation

◆ account

str generate-submission-script-gpu-parallel.account = "do016"

Definition at line 21 of file generate-submission-script-gpu-parallel.py.

◆ application

str generate-submission-script-gpu-parallel.application = "euler"

Definition at line 6 of file generate-submission-script-gpu-parallel.py.

◆ benchmark_app_name

str generate-submission-script-gpu-parallel.benchmark_app_name = "KernelBenchmarksFD4DSL.Release"

Definition at line 45 of file generate-submission-script-gpu-parallel.py.

◆ benchmark_script_name

str generate-submission-script-gpu-parallel.benchmark_script_name = "kernel-benchmarks-fd4-dsl.py"

Definition at line 44 of file generate-submission-script-gpu-parallel.py.

◆ compile_script_name

str generate-submission-script-gpu-parallel.compile_script_name = f"compile_kernel_benchmark-gpu-parallel-{application}-{solver}.sh"

Definition at line 40 of file generate-submission-script-gpu-parallel.py.

◆ compile_scripts

list generate-submission-script-gpu-parallel.compile_scripts = []

Definition at line 90 of file generate-submission-script-gpu-parallel.py.

◆ job_name

str generate-submission-script-gpu-parallel.job_name = f"{application}_kernel_benchmark"

Definition at line 37 of file generate-submission-script-gpu-parallel.py.

◆ machine

str generate-submission-script-gpu-parallel.machine = "grace_hopper"

Definition at line 8 of file generate-submission-script-gpu-parallel.py.

◆ number_of_patches

dict generate-submission-script-gpu-parallel.number_of_patches = numbers_of_patches[patch_size]

Definition at line 95 of file generate-submission-script-gpu-parallel.py.

◆ number_of_samples

int generate-submission-script-gpu-parallel.number_of_samples = 10

Definition at line 11 of file generate-submission-script-gpu-parallel.py.

◆ numbers_of_patches

dict generate-submission-script-gpu-parallel.numbers_of_patches

Initial value:

=  {
    3: ["1", "2", "4", "8", "16", "32", "64", "128", "256", "512", "1024", "2048", "4096", "8192"],
    6: ["1", "2", "4", "8", "16", "32", "64", "128", "256", "512", "1024", "2048", "4096", "8192"],
    16: ["1", "2", "4", "8", "16", "32", "64", "128", "256", "512", "1024", "2048", "4096", "8192"]
}

Definition at line 13 of file generate-submission-script-gpu-parallel.py.

◆ optional_passes

dict generate-submission-script-gpu-parallel.optional_passes

Initial value:

=  {
    "memory-reuse": "--enable-memory-reuse",
    "perfect-loop-nesting": "--enable-imperfect-nested-loops-to-perfect",
    "memory-fusion": "--enable-memory-fusion",
    "loop-fusion": "--enable-loop-fusion"
}