From 9a013b7c68f94e6be0fe8748c9012a441fe0273f Mon Sep 17 00:00:00 2001
From: Sebastiano Tronto <sebastiano@tronto.net>
Date: Mon, 12 Jan 2026 18:09:43 +0100
Subject: Improve performance of H48 solver with prefetching

With this commit we re-structure how the node expansion in the
H48 solution search works to allow prefetching of pruning values,
showing performance improvements in the range of 30-45% on x86,
depending on table size and solution length.

A small bug fix related to appending solutions is included in this commit.
---
 src/utils/prefetch.h | 15 +++++++++++++++
 src/utils/utils.h    |  1 +
 2 files changed, 16 insertions(+)
 create mode 100644 src/utils/prefetch.h

(limited to 'src/utils')

diff --git a/src/utils/prefetch.h b/src/utils/prefetch.h
new file mode 100644
index 0000000..89b14e8
--- /dev/null
+++ b/src/utils/prefetch.h
@@ -0,0 +1,15 @@
+#if defined(AVX2)
+
+#define prefetch(a, i) _mm_prefetch(a+i, _MM_HINT_T0)
+
+#else
+#if defined(__GNUC__) || defined(__clang__)
+
+#define prefetch(a, i) __builtin_prefetch(a+i, 0, 0)
+
+#else
+
+#define prefetch(a, i) (void)i
+
+#endif
+#endif
diff --git a/src/utils/utils.h b/src/utils/utils.h
index 383781a..e322f20 100644
--- a/src/utils/utils.h
+++ b/src/utils/utils.h
@@ -4,3 +4,4 @@
 #include "math.h"
 #include "sleep.h"
 #include "wrapthread.h"
+#include "prefetch.h"
-- 
cgit v1.3