Edit

Share via

/forceInterlockedFunctions

  • Article

Dynamically selects between Armv8.0 load, store exclusive instructions or Armv8.1 Large System Extension (LSE) atomic instructions based on CPU capability at runtime.

Syntax

/forceInterlockedFunctions[-]

Remarks

When possible, this flag avoids using Armv8.0 load and store exclusive instructions, as these instructions can result in livelocks. This flag forces the following interlocked intrinsics to be generated as out-of-line functions:

Operation 8 16 32 64 128 Pointer
Add None None Full Full None None
And Full Full Full Full None None
CompareExchange Full Full Full Full Full Full
Decrement None Full Full Full None None
Exchange Full Full Full Full None Full
ExchangeAdd Full Full Full Full None None
Increment None Full Full Full None None
Or Full Full Full Full None None
Xor Full Full Full Full None None
bittestandreset None None Full Full None None
bittestandset None None Full Full None None

Key:

  • Full: supports plain, _acq, _rel, and _nf forms.

  • None: Not supported

For more information about interlocked intrinsics, see the "Interlocked intrinsics" section in Arm64 Intrinsics.

To set the /forceInterlockedFunctions compiler option in Visual Studio

  1. Open the Property Pages dialog box for the project. For more information, see Set C++ compiler and build properties in Visual Studio.

  2. Select the Configuration Properties > C/C++ > Command Line property page.

  3. In the Additional options box, add /forceInterlockedFunctions to enable. Choose OK to save your changes.

To set this compiler option programmatically

See also

Arm64 Intrinsics
MSVC compiler options
MSVC compiler command-line syntax