Page Index - unodb-dev/unodb GitHub Wiki
111 page(s) in this GitHub Wiki:
- Home
- Concurrent ART flavor overhead in single threaded workloads
- Please reload this page
- Effect of annotating pmr_allocate with attribute((malloc,returns_nonnull))
- Please reload this page
- Effect of AVX2 for Node48 insert position search
- Please reload this page
- Effect of branching exactly once in remove code path
- Please reload this page
- Effect of fully inlining try_read_unlock
- Please reload this page
- Effect of initializing Node48 child_indexes at declaration
- Please reload this page
- Effect of jemalloc sdallocx sized free
- Please reload this page
- Effect of LD_PRELOAD'ing jemalloc
- Please reload this page
- Effect of lock free QSBR step 1
- Please reload this page
- Effect of lock free QSBR step 10
- Please reload this page
- Effect of lock free QSBR step 11
- Please reload this page
- Effect of lock free QSBR step 12
- Please reload this page
- Effect of lock free QSBR step 2
- Please reload this page
- Effect of lock free QSBR step 3
- Please reload this page
- Effect of lock free QSBR step 4
- Please reload this page
- Effect of lock free QSBR step 5
- Please reload this page
- Effect of lock free QSBR step 6
- Please reload this page
- Effect of lock free QSBR step 7
- Please reload this page
- Effect of lock free QSBR step 8
- Please reload this page
- Effect of lock free QSBR step 9
- Please reload this page
- Effect of merging find_child and add node type dispatch
- Please reload this page
- Effect of Node16 NEON search
- Please reload this page
- Effect of Node16 NEON search replacing uint128_t with a ternary
- Please reload this page
- Effect of Node4 NEON search
- Please reload this page
- Effect of Node48 AVX2 insert position search unrolling x2
- Please reload this page
- Effect of Node48 AVX2 insert position search unrolling x4
- Please reload this page
- Effect of Node48 NEON search
- Please reload this page
- Effect of Node48 NEON search unrolled x2
- Please reload this page
- Effect of optimistic_lock::try_read_lock testing once in the straight path
- Please reload this page
- Effect of padding OLC ART counters
- Please reload this page
- Effect of removing branching on children_count in N4 insert_position_search
- Please reload this page
- Effect of removing is_full dispatch on add node path
- Please reload this page
- Effect of removing node type check before growing inode
- Please reload this page
- Effect of removing Node256 fullness check on add
- Please reload this page
- Effect of removing polymorphism on inode entry removal
- Please reload this page
- Effect of removing pool PMR allocators
- Please reload this page
- Effect of removing redundant nullptr checks in deleters
- Please reload this page
- Effect of removing std::uncaught_exceptions calls on OL obsoleting critical section enter and exit
- Please reload this page
- Effect of replacing RMW atomics with load non atomic modify store in OL critical sections
- Please reload this page
- Effect of replacing RMW with W in optimistic lock obsoletion
- Please reload this page
- Effect of replacing RMW with W in optimistic lock write unlock
- Please reload this page
- Effect of SIMD'ing Node4 insert position search
- Please reload this page
- Effect of simpler Node4 add_two_to_empty key2_i
- Please reload this page
- Effect of tagged node pointers
- Please reload this page
- Performance comparison with @justinasvd art_map
- Please reload this page
- Performance of GCC 11 vs GCC 10
- Please reload this page
- Performance of GCC 12 vs GCC 11
- Please reload this page
- Performance of LLVM 12 vs LLVM 11
- Please reload this page
- Performance of LLVM 12 vs LLVM 11, take 2
- Please reload this page
- Performance of LLVM 13 vs LLVM 12
- Please reload this page
- Performance of LLVM 14 vs LLVM 13
- Please reload this page
- Performance of LLVM 15 vs LLVM 14
- Please reload this page
- Performance of LLVM 16 vs LLVM 15
- Please reload this page
- Performance of LLVM 17 vs LLVM 16
- Please reload this page
- Performance Tests
- Please reload this page