Module nvptx

Source
🔬This is a nightly-only experimental API. (stdarch_nvptx #111199)
Available on target_arch="nvptx64" only.
Expand description

Platform-specific intrinsics for the NVPTX platform.

See the module documentation for more details.

Structs§

f16x2Experimental
PTX-specific 32-bit wide floating point (f16 x 2) vector type

Functions§

__assert_failâš Experimental
Syscall to be used whenever the assert expression produces a false value.
_block_dim_xâš Experimental
x-th thread-block dimension.
_block_dim_yâš Experimental
y-th thread-block dimension.
_block_dim_zâš Experimental
z-th thread-block dimension.
_block_idx_xâš Experimental
x-th thread-block index.
_block_idx_yâš Experimental
y-th thread-block index.
_block_idx_zâš Experimental
z-th thread-block index.
_grid_dim_xâš Experimental
x-th block-grid dimension.
_grid_dim_yâš Experimental
y-th block-grid dimension.
_grid_dim_zâš Experimental
z-th block-grid dimension.
_syncthreadsâš Experimental
Synchronizes all threads in the block.
_thread_idx_xâš Experimental
x-th thread index.
_thread_idx_yâš Experimental
y-th thread index.
_thread_idx_zâš Experimental
z-th thread index.
f16x2_addâš Experimental
Add two values, round to nearest even
f16x2_fmaâš Experimental
Fused multiply-add, round to nearest even
f16x2_maxâš Experimental
Find the maximum of two values
f16x2_max_nanâš Experimental
Find the maximum of two values, NaNs pass through.
f16x2_minâš Experimental
Find the minimum of two values
f16x2_min_nanâš Experimental
Find the minimum of two values, NaNs pass through.
f16x2_mulâš Experimental
Multiply two values, round to nearest even
f16x2_negâš Experimental
Arithmetic negate
f16x2_subâš Experimental
Subtract two values, round to nearest even
freeâš Experimental
Free previously dynamically allocated memory.
mallocâš Experimental
Allocate memory dynamically from a fixed-size heap in global memory.
trapâš Experimental
Generates the trap instruction TRAP
vprintfâš Experimental
Print formatted output from a kernel to a host-side output stream.