work_queue_persistent_inl.h File Reference

#include <nvbio/basic/types.h>
#include <nvbio/basic/numbers.h>
#include <nvbio/basic/cuda/arch.h>
#include <cub/cub.cuh>
#include <thrust/copy.h>

Go to the source code of this file.

Namespaces
	nvbio
	Define a vector_view POD type and plain_view() for std::vector.
	nvbio::cuda
	nvbio::cuda::wq

Functions
template<uint32 BLOCKDIM, typename WorkUnitT , typename WorkStreamT >
__global__ void	nvbio::cuda::wq::persistent_warps_work_queue_kernel (uint32 *pool, const WorkStreamT stream, WorkQueueStats::View stats)
template<uint32 BLOCKDIM, typename WorkUnitT , typename WorkStreamT >
__global__ void	nvbio::cuda::wq::__launch_bounds__ (BLOCKDIM, 6) persistent_threads_work_queue_kernel(uint32 *pool
	nvbio::cuda::wq::while (__any(active))

Variables
__global__ void const uint32	nvbio::cuda::wq::max_inactive_lanes
__global__ void const uint32 const WorkStreamT	nvbio::cuda::wq::stream
__global__ void const uint32 const WorkStreamT WorkQueueStats::View	nvbio::cuda::wq::stats
const uint32	nvbio::cuda::wq::thread_id = threadIdx.x + blockIdx.x*BLOCKDIM
WorkUnit	nvbio::cuda::wq::unit
const uint32	nvbio::cuda::wq::NUM_WARPS = BLOCKDIM >> cuda::Arch::LOG_WARP_SIZE
__shared__ volatile uint32	nvbio::cuda::wq::sm_broadcast [NUM_WARPS]
const uint32	nvbio::cuda::wq::invalid_unit = uint32(-1)
uint32	nvbio::cuda::wq::work_id = invalid_unit
uint32	nvbio::cuda::wq::work_iter = 0u
bool	nvbio::cuda::wq::active = true
const uint32	nvbio::cuda::wq::stream_size = stream.size()