Timer update (#856)

Faster macos timer
Faster linux timer (hopefully works on emscripten)

I confirmed these timers are fast

Updated M2 MacOS benchmarks
Updated N100 Linux benchmarks

benchmark/m2air_neon/joint_grid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,515.583
-2,968.592
-3,1393.95
-4,1652.07
+1,214.889
+2,344.308
+3,499.192
+4,589.811

benchmark/m2air_neon/large_pyramid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,327.475
-2,606.037
-3,843.779
-4,1057.89
+1,326.445
+2,594.514
+3,826.976
+4,1043.61

benchmark/m2air_neon/many_pyramids.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,85.4238
-2,159.788
-3,219.475
-4,261.439
+1,85.882
+2,160.755
+3,220.359
+4,266.172

benchmark/m2air_neon/rain.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,152.46
+2,253.474
+3,326.158
+4,396.199

benchmark/m2air_neon/smash.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,166.289
-2,262.228
-3,329.267
-4,393.513
+1,193.98
+2,307.236
+3,386.559
+4,459.442

benchmark/m2air_neon/spinner.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,312.26
+2,493.611
+3,639.152
+4,780.522

benchmark/m2air_neon/tumbler.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,351.845
-2,549.496
-3,674.889
-4,801.015
+1,431.639
+2,671.981
+3,818.123
+4,993.461

benchmark/n100_avx2/joint_grid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,115.85
-2,209.251
-3,266.537
-4,319.147
+1,75.178
+2,123.214
+3,158.983
+4,182.112

benchmark/n100_avx2/large_pyramid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,133.69
-2,239.395
-3,312.458
-4,360.458
+1,132.297
+2,234.621
+3,308.883
+4,355.701

benchmark/n100_avx2/many_pyramids.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,32.0345
-2,56.8865
-3,71.4046
-4,76.2903
+1,32.0528
+2,56.6901
+3,71.2562
+4,77.5141

benchmark/n100_avx2/rain.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,71.7432
+2,117.863
+3,141.721
+4,161.414

benchmark/n100_avx2/smash.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,73.9639
-2,115.759
-3,138.925
-4,157.394
+1,88.4488
+2,135.265
+3,162.135
+4,184.08

benchmark/n100_avx2/spinner.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,159.879
+2,261.516
+3,308.75
+4,360.254

benchmark/n100_avx2/tumbler.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,163.663
-2,261.691
-3,317.057
-4,362.4
+1,205.268
+2,320.216
+3,389.513
+4,449.153

benchmark/n100_sse2/joint_grid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,119.293
-2,213.357
-3,268.455
-4,321.339
+1,75.5947
+2,123.228
+3,160.379
+4,181.545

benchmark/n100_sse2/large_pyramid.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,128.152
-2,229.067
-3,302.297
-4,348.367
+1,127.236
+2,226.291
+3,297.628
+4,345.526

benchmark/n100_sse2/many_pyramids.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,30.9601
-2,55.0378
-3,69.5563
-4,77.3663
+1,30.8828
+2,55.0462
+3,69.5406
+4,77.7339

benchmark/n100_sse2/rain.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,72.2901
+2,118.753
+3,142.61
+4,162.35

benchmark/n100_sse2/smash.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,73.2282
-2,114.239
-3,137.388
-4,156.341
+1,86.2381
+2,132.306
+3,160.725
+4,181.842

benchmark/n100_sse2/spinner.csv

@@ -0,0 +1,5 @@
+threads,fps
+1,156.855
+2,258.638
+3,303.717
+4,358.492

benchmark/n100_sse2/tumbler.csv

@@ -1,5 +1,5 @@
 threads,fps
-1,161.228
-2,257.781
-3,311.219
-4,356.481
+1,199.492
+2,313.012
+3,381.983
+4,441.825

include/box2d/base.h

@@ -115,9 +115,11 @@ typedef struct b2Timer
 {
 #if defined( _WIN32 )
 	int64_t start;
-#elif defined( __linux__ ) || defined( __APPLE__ ) || defined( __EMSCRIPTEN__ )
-	unsigned long long start_sec;
-	unsigned long long start_usec;
+#elif defined( __linux__ ) || defined( __EMSCRIPTEN__ )
+	int64_t tv_sec;
+	int64_t tv_nsec;
+#elif defined( __APPLE__ )
+	uint64_t start;
 #else
 	int32_t dummy;
 #endif

include/box2d/collision.h

@@ -21,7 +21,7 @@ typedef struct b2Hull b2Hull;
 
 /// The maximum number of vertices on a convex polygon. Changing this affects performance even if you
 /// don't use more vertices.
-#define b2_maxPolygonVertices 8
+#define B2_MAX_POLYGON_VERTICES 8
 
 /// Low level ray cast input data
 typedef struct b2RayCastInput
@@ -42,7 +42,7 @@ typedef struct b2RayCastInput
 typedef struct b2ShapeCastInput
 {
 	/// A point cloud to cast
-	b2Vec2 points[b2_maxPolygonVertices];
+	b2Vec2 points[B2_MAX_POLYGON_VERTICES];
 
 	/// The number of points
 	int32_t count;
@@ -115,17 +115,17 @@ typedef struct b2Capsule
 
 /// A solid convex polygon. It is assumed that the interior of the polygon is to
 /// the left of each edge.
-/// Polygons have a maximum number of vertices equal to b2_maxPolygonVertices.
+/// Polygons have a maximum number of vertices equal to B2_MAX_POLYGON_VERTICES.
 /// In most cases you should not need many vertices for a convex polygon.
 /// @warning DO NOT fill this out manually, instead use a helper function like
 /// b2MakePolygon or b2MakeBox.
 typedef struct b2Polygon
 {
 	/// The polygon vertices
-	b2Vec2 vertices[b2_maxPolygonVertices];
+	b2Vec2 vertices[B2_MAX_POLYGON_VERTICES];
 
 	/// The outward normal vectors of the polygon sides
-	b2Vec2 normals[b2_maxPolygonVertices];
+	b2Vec2 normals[B2_MAX_POLYGON_VERTICES];
 
 	/// The centroid of the polygon
 	b2Vec2 centroid;
@@ -273,7 +273,7 @@ B2_API b2CastOutput b2ShapeCastPolygon( const b2ShapeCastInput* input, const b2P
 typedef struct b2Hull
 {
 	/// The final points of the hull
-	b2Vec2 points[b2_maxPolygonVertices];
+	b2Vec2 points[B2_MAX_POLYGON_VERTICES];
 
 	/// The number of points
 	int32_t count;
@@ -284,7 +284,7 @@ typedef struct b2Hull
 /// - all points very close together
 /// - all points on a line
 /// - less than 3 points
-/// - more than b2_maxPolygonVertices points
+/// - more than B2_MAX_POLYGON_VERTICES points
 /// This welds close points and removes collinear points.
 /// @warning Do not modify a hull once it has been computed
 B2_API b2Hull b2ComputeHull( const b2Vec2* points, int32_t count );
@@ -333,7 +333,7 @@ B2_API b2SegmentDistanceResult b2SegmentDistance( b2Vec2 p1, b2Vec2 q1, b2Vec2 p
 typedef struct b2DistanceProxy
 {
 	/// The point cloud
-	b2Vec2 points[b2_maxPolygonVertices];
+	b2Vec2 points[B2_MAX_POLYGON_VERTICES];
 
 	/// The number of points
 	int32_t count;
@@ -601,21 +601,13 @@ B2_API b2Manifold b2CollideChainSegmentAndPolygon( const b2ChainSegment* segment
  * A dynamic AABB tree broad-phase, inspired by Nathanael Presson's btDbvt.
  * A dynamic tree arranges data in a binary tree to accelerate
  * queries such as AABB queries and ray casts. Leaf nodes are proxies
- * with an AABB. These are used to hold a user collision object, such as a reference to a b2Shape.
+ * with an AABB. These are used to hold a user collision object.
  * Nodes are pooled and relocatable, so I use node indices rather than pointers.
  * The dynamic tree is made available for advanced users that would like to use it to organize
  * spatial game data besides rigid bodies.
- *
- * @note This is an advanced feature and normally not used by applications directly.
  * @{
  */
 
-/// The default category bit for a tree proxy. Used for collision filtering.
-#define b2_defaultCategoryBits ( 1 )
-
-/// Convenience mask bits to use when you don't need collision filtering and just want all results.
-#define b2_defaultMaskBits ( UINT64_MAX )
-
 /// A node in the dynamic tree. This is private data placed here for performance reasons.
 typedef struct b2TreeNode
 {
@@ -653,7 +645,7 @@ typedef struct b2TreeNode
 /// The dynamic tree structure. This should be considered private data.
 /// It is placed here for performance reasons.
 typedef struct b2DynamicTree
-{
+{  
 	/// The tree nodes
 	b2TreeNode* nodes;
 

samples/sample_benchmark.cpp

@@ -1297,7 +1297,7 @@ class BenchmarkCast : public Sample
 		int nodeVisits = 0;
 		int leafVisits = 0;
 		float ms = 0.0f;
-		int sampleCount = m_origins.size();
+		int sampleCount = (int)m_origins.size();
 
 		if ( m_queryType == e_rayCast )
 		{

samples/sample_collision.cpp

@@ -554,7 +554,7 @@ class DynamicTree : public Sample
 					p->fatBox.lowerBound = b2Sub( p->box.lowerBound, aabbMargin );
 					p->fatBox.upperBound = b2Add( p->box.upperBound, aabbMargin );
 
-					p->proxyId = b2DynamicTree_CreateProxy( &m_tree, p->fatBox, b2_defaultCategoryBits, m_proxyCount );
+					p->proxyId = b2DynamicTree_CreateProxy( &m_tree, p->fatBox, B2_DEFAULT_CATEGORY_BITS, m_proxyCount );
 					p->rayStamp = -1;
 					p->queryStamp = -1;
 					p->moved = false;
@@ -682,7 +682,7 @@ class DynamicTree : public Sample
 		if ( m_queryDrag )
 		{
 			b2AABB box = { b2Min( m_startPoint, m_endPoint ), b2Max( m_startPoint, m_endPoint ) };
-			b2DynamicTree_Query( &m_tree, box, b2_defaultMaskBits, QueryCallback, this );
+			b2DynamicTree_Query( &m_tree, box, B2_DEFAULT_MASK_BITS, QueryCallback, this );
 
 			g_draw.DrawAABB( box, b2_colorWhite );
 		}
@@ -693,7 +693,7 @@ class DynamicTree : public Sample
 		if ( m_rayDrag )
 		{
 			b2RayCastInput input = { m_startPoint, b2Sub( m_endPoint, m_startPoint ), 1.0f };
-			b2TreeStats result = b2DynamicTree_RayCast( &m_tree, &input, b2_defaultMaskBits, RayCallback, this );
+			b2TreeStats result = b2DynamicTree_RayCast( &m_tree, &input, B2_DEFAULT_MASK_BITS, RayCallback, this );
 
 			g_draw.DrawSegment( m_startPoint, m_endPoint, b2_colorWhite );
 			g_draw.DrawPoint( m_startPoint, 5.0f, b2_colorGreen );
@@ -2162,7 +2162,7 @@ class OverlapWorld : public Sample
 		{
 			b2World_OverlapPolygon( m_worldId, &m_queryBox, transform, b2DefaultQueryFilter(), OverlapWorld::OverlapResultFcn,
 									this );
-			b2Vec2 points[b2_maxPolygonVertices] = { 0 };
+			b2Vec2 points[B2_MAX_POLYGON_VERTICES] = { 0 };
 			for ( int i = 0; i < m_queryBox.count; ++i )
 			{
 				points[i] = b2TransformPoint( transform, m_queryBox.vertices[i] );
@@ -3316,7 +3316,7 @@ class ShapeCast : public Sample
 						   output.iterations, output.fraction, distanceOutput.distance );
 		m_textLine += m_textIncrement;
 
-		b2Vec2 vertices[b2_maxPolygonVertices];
+		b2Vec2 vertices[B2_MAX_POLYGON_VERTICES];
 
 		for ( int i = 0; i < m_countA; ++i )
 		{
@@ -3392,11 +3392,11 @@ class ShapeCast : public Sample
 		g_draw.DrawSegment( m_transformB.p, b2Add( m_transformB.p, m_translationB ), b2_colorGray );
 	}
 
-	b2Vec2 m_vAs[b2_maxPolygonVertices];
+	b2Vec2 m_vAs[B2_MAX_POLYGON_VERTICES];
 	int m_countA;
 	float m_radiusA;
 
-	b2Vec2 m_vBs[b2_maxPolygonVertices];
+	b2Vec2 m_vBs[B2_MAX_POLYGON_VERTICES];
 	int m_countB;
 	float m_radiusB;
 
@@ -3497,7 +3497,7 @@ class TimeOfImpact : public Sample
 		//                        b2_toiMaxRootIters);
 		m_textLine += m_textIncrement;
 
-		b2Vec2 vertices[b2_maxPolygonVertices];
+		b2Vec2 vertices[B2_MAX_POLYGON_VERTICES];
 
 		// Draw A
 		b2Transform transformA = b2GetSweepTransform( &sweepA, 0.0f );

samples/sample_geometry.cpp

@@ -15,7 +15,7 @@ class ConvexHull : public Sample
 public:
 	enum
 	{
-		e_count = b2_maxPolygonVertices
+		e_count = B2_MAX_POLYGON_VERTICES
 	};
 
 	explicit ConvexHull( Settings& settings )
@@ -211,7 +211,7 @@ class ConvexHull : public Sample
 		return new ConvexHull( settings );
 	}
 
-	b2Vec2 m_points[b2_maxPolygonVertices];
+	b2Vec2 m_points[B2_MAX_POLYGON_VERTICES];
 	int32_t m_count;
 	int32_t m_generation;
 	bool m_auto;