GeneralRules

General CUDA rules

Nested Rules

CUDA-1.1	Use CUDA after host program initiation and before host program termination
CUDA-1.2	Synchronize with all devices before termination
CUDA-1.3	Check for errors after calling any CUDA library interface
CUDA-1.4	Check for errors after launching a kernel by calling cudaGetLastError
CUDA-1.5	Treat CUdevice objects as handles not integers
CUDA-1.6	A call to fork must be immediately followed to a call to exec
CUDA-1.7	Be consistent with execution space and kernel function specifiers across all declarations
CUDA-1.8	Kernel parameters passed by value should be trivial
CUDA-1.9	Kernel parameters passed by reference should be managed storage duration objects
CUDA-1.10	Never use CUDA_ARCH to guard CUDA declarations
CUDA-1.11	Do not modify CUDA language specific macro names
CUDA-1.12	Constant memory objects shall not be modified from the device program
CUDA-1.13	No thread_local storage duration objects in device code
CUDA-1.14	A implicit thread block synchronization function should be called by all active threads

Options

• Setting an option for this rule means setting the default for all nested rules.

• This rule shares the following common options: exclude_in_macros, exclude_messages_in_system_headers, excludes, extend_exclude_to_macro_invocations, includes, justification_checker, languages, post_processing, provider, report_at, severity

• The following places define options that affect this rule: Stylechecks, Analysis-GlobalOptions

• This rule has no individual options.