Skip to content

MDEV-40827: prefetch MHNSW neighbours during search #5571

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
chanztuying wants to merge 2 commits into
base: main
Choose a base branch
from
+28 −0
Open

MDEV-40827: prefetch MHNSW neighbours during search #5571

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8c0ba52
MDEV-38721: prefetch MHNSW neighbours during search
chanztuying Aug 18, 2026
edf9a46
Wrap prefetch in MHNSW_PREFETCH_READ() and implements MSVC prefetching
chanztuying Aug 19, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 28 additions & 0 deletions sql/vector_mhnsw.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,22 @@ static constexpr float subdist_margin= 1.05f;
static constexpr double subdist_stddev_threshold= 0.05; // 3σ, p>99.9%
static constexpr ulonglong subdist_stddev_valid= 10000; // sufficient

/* Portable read prefetch, modeled after UNIV_PREFETCH_R. */
#if defined(__GNUC__) || defined(__clang__)
# define MHNSW_PREFETCH_READ(addr) __builtin_prefetch((addr), 0, 3)
#elif defined(_MSC_VER)
# include <intrin.h>
# if defined(_M_IX86) || defined(_M_X64)
# define MHNSW_PREFETCH_READ(addr) \
_mm_prefetch(reinterpret_cast<const char*>(addr), _MM_HINT_T0)
# elif defined(_M_ARM64)
# define MHNSW_PREFETCH_READ(addr) __prefetch(addr)
# endif
#endif
#ifndef MHNSW_PREFETCH_READ
# define MHNSW_PREFETCH_READ(addr) ((void) 0)
#endif

/*
The class below can assume normal distribution and only collect
M1 and M2, or go beyond that and collect M3 and M4 to account
Expand Down Expand Up @@ -1347,6 +1363,18 @@ static int search_layer(MHNSW_param *p, const FVector *target, float threshold,
if (res == 0xff)
continue;

// A node and its vector share one allocation. Prefetch unseen nodes before
// computing distances so later loads can overlap with work on earlier ones.
for (size_t i= 0; i < 8; i++)
{
FVectorNode *link= links[i];
if (!(res & (1 << i)) && link)
{
MHNSW_PREFETCH_READ(link);
MHNSW_PREFETCH_READ(reinterpret_cast<const char*>(link) + 64);
}
}

for (size_t i= 0; i < 8; i++)
{
if (res & (1 << i))
Expand Down