CUDASafety-1.4¶
Device code should not contain unsupported types
Required inputs: IR
CUDA SAFETY 1.4 [safety.unsupported_type] Device code should not contain unsupported types
The long double type shall not be used in device code.
Scope: Device.
Audience: CUDA C++.
Category: Mandatory.
Hardware Applicability: All Compute Capabilities.
Rationale
The long double type is not supported in the CUDA safety subset.
Example 1 (Bad)
template<typename T> __global__ void kernel(long double) { // non-compliant: use of long double in // kernel arguments. unsigned i = sizeof(long double); // non-compliant: use of long double in // a constant expression. } int main() { kernel<long double><<<1,1>>>(0); // non-compliant: instantiation of // a device function is dependent on // long double. }Excerpt from NVIDIA CUDA C++ Guidelines for robust and safety-critical programming, Version 3.0.1, Copyright (C) 2018-2023 NVIDIA Corporation.
Possible Messages
Key |
Text |
Severity |
Disabled |
|---|---|---|---|
unsupported_type_in_device_code |
Do not use unsupported types in device code |
None |
False |
Options¶
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
forbidden_types¶
forbidden_types : set[bauhaus.ir.LIR_Class_Name] = {'C_Long_Double_Type'}