From d008a2678397bd2c45c53e997f7d14f1cfadef4c Mon Sep 17 00:00:00 2001
From: ash9146 <52308288+ash9146@users.noreply.github.com>
Date: Sat, 1 Aug 2026 05:24:32 +0900
Subject: [PATCH] Cross-link GPU-accelerated vector indexing from
 approximate-knn-search.md

This page covers ways to speed up index construction and merging
(increasing max_merged_segment, disabling refresh during bulk uploads),
but didn't mention GPU-accelerated indexing via NVIDIA cuVS, which
addresses the same problem (faster HNSW construction, faster
force-merge) and already has a dedicated reference page.

Noted as GA since 9.4, per the official Elastic 9.4 release blog
(https://www.elastic.co/blog/whats-new-elastic-9-4-0), corroborated by
an independent release-notes aggregator. One Elasticsearch Labs post
describes self-managed Enterprise GA as still targeted for 9.5, but
that appears to be an outdated statement given both other sources are
more recent and agree on 9.4.
---
 .../optimize-performance/approximate-knn-search.md           | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/deploy-manage/production-guidance/optimize-performance/approximate-knn-search.md b/deploy-manage/production-guidance/optimize-performance/approximate-knn-search.md
index 3c34413133..b69f7c35db 100644
--- a/deploy-manage/production-guidance/optimize-performance/approximate-knn-search.md
+++ b/deploy-manage/production-guidance/optimize-performance/approximate-knn-search.md
@@ -190,6 +190,11 @@ A common pattern is to first perform an initial bulk upload, then make an index
 * Give {{es}} a large indexing buffer so it can accept more documents before flushing. By default, the [`indices.memory.index_buffer_size`](elasticsearch://reference/elasticsearch/configuration-reference/indexing-buffer-settings.md) is set to 10% of the heap size. With a substantial heap size like 32GB, this is often enough. To allow the full indexing buffer to be used, you should also increase the limit [`index.translog.flush_threshold_size`](elasticsearch://reference/elasticsearch/index-settings/translog.md).
 
 
+### Use GPU-accelerated indexing [_use_gpu_accelerated_indexing]
+
+For indexing-heavy workloads on large vector datasets, GPU acceleration (generally available since {{es}} 9.4) can significantly speed up HNSW index construction and reduce the cost of merging segments into larger ones. See [GPU accelerated vector indexing](elasticsearch://reference/elasticsearch/mapping-reference/gpu-vector-indexing.md) for supported configurations and setup.
+
+
 ## Avoid heavy indexing during searches [_avoid_heavy_indexing_during_searches]
 
 Actively indexing documents can have a negative impact on approximate kNN search performance, since indexing threads steal compute resources from search. When indexing and searching at the same time, {{es}} also refreshes frequently, which creates several small segments. This also hurts search performance, since approximate kNN search is slower when there are more segments.