From 0c7ad4e85944b38ef6e4d2becfd58bf58385812e Mon Sep 17 00:00:00 2001
From: Adam Sawicki <adam.sawicki@amd.com>
Date: Tue, 29 Mar 2022 18:07:09 +0200
Subject: [PATCH] Optimized VmaCountBitsSets to use std::popcount when C++20 is
 enabled

Closes #251
---
 include/vk_mem_alloc.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/include/vk_mem_alloc.h b/include/vk_mem_alloc.h
index 256b654..98aa40f 100644
--- a/include/vk_mem_alloc.h
+++ b/include/vk_mem_alloc.h
@@ -2574,6 +2574,9 @@ VMA_CALL_PRE void VMA_CALL_POST vmaFreeStatsString(
 #ifdef _MSC_VER
     #include <intrin.h> // For functions like __popcnt, _BitScanForward etc.
 #endif
+#if __cplusplus >= 202002L || _MSVC_LANG >= 202002L // C++20
+    #include <bit> // For std::popcount
+#endif
 
 /*******************************************************************************
 CONFIGURATION SECTION
@@ -3176,12 +3179,16 @@ But you need to check in runtime whether user's CPU supports these, as some old
 */
 static inline uint32_t VmaCountBitsSet(uint32_t v)
 {
+#if __cplusplus >= 202002L || _MSVC_LANG >= 202002L // C++20
+    return std::popcount(v);
+#else
     uint32_t c = v - ((v >> 1) & 0x55555555);
     c = ((c >> 2) & 0x33333333) + (c & 0x33333333);
     c = ((c >> 4) + c) & 0x0F0F0F0F;
     c = ((c >> 8) + c) & 0x00FF00FF;
     c = ((c >> 16) + c) & 0x0000FFFF;
     return c;
+#endif
 }
 
 static inline uint8_t VmaBitScanLSB(uint64_t mask)