Optimized Partitioning Using SW-buffers

Classes
union	cacheline_t

Macros
#define	TUPLESPERCACHELINE   (CACHE_LINE_SIZE/sizeof(tuple_t))

Functions
void	parallel_radix_partition_optimized (part_t *const part)

Variables
typedef	__attribute__

Detailed Description

Macro Definition Documentation

TUPLESPERCACHELINE

#define TUPLESPERCACHELINE   (CACHE_LINE_SIZE/sizeof(tuple_t))

Definition at line 798 of file parallel_radix_join.c.

Function Documentation

parallel_radix_partition_optimized()

void parallel_radix_partition_optimized

(

part_t *const

part

)

This function implements the parallel radix partitioning of a given input relation. Parallel partitioning is done by histogram-based relation re-ordering as described by Kim et al. Parallel partitioning method is commonly used by all parallel radix join algorithms. However this implementation is further optimized to benefit from write-combining and non-temporal writes.

Parameters

part	description of the relation to be partitioned

Definition at line 857 of file parallel_radix_join.c.