cuda.bindings._v2.driver.SharedCarveout#

class cuda.bindings._v2.driver.SharedCarveout(value: int)#

Shared memory carveout configurations. These may be passed to cuFuncSetAttribute or cuKernelSetAttribute

See CUshared_carveout.

DEFAULT: -1

No preference for shared memory or L1 (default)

MAX_SHARED: 100

Prefer maximum available shared memory, minimum L1 cache

MAX_L1: 0

Prefer maximum available L1 cache, minimum shared memory