[英]Must I keep a virtual address range reservation if it has active mappings?
CUDA's low-level virtual memory management mechanism involves: CUDA的低级虚拟memory管理机制涉及:
Conveniently, if you map a physical allocation to some address range - you can "free" the physical allocation and keep using it until it is unmapped.方便的是,如果您 map 对某个地址范围进行物理分配 - 您可以“释放”物理分配并继续使用它,直到它被取消映射。
Can we also do this for virtual address range reservations?我们也可以为虚拟地址范围保留执行此操作吗? ie will they be kept alive until the mappings are gone?
即在映射消失之前它们会一直保持活力吗? The documentation doesn't say.
文档没有说。
Virtual memory address range reservations apparently don't have the reference-counting mechanism which physical allocations have;虚拟 memory 地址范围保留显然没有物理分配所具有的引用计数机制; the driver expects them to be released when they are no longer in use, period.
司机希望它们在不再使用时被释放,期间。 An attempt to free the reservation will fail with
CUDA_ERROR_INVALID_VALUE
.释放预留的尝试将失败并显示
CUDA_ERROR_INVALID_VALUE
。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.