Benchmarking.

diff --git a/benchmarks.md b/benchmarks.md
index e6e93548698003215cabfe85a8a8aab1941a90b6..bd678546deed225de28d6947a2955626daa46c50 100644 (file)
--- a/benchmarks.md
+++ b/benchmarks.md
@@ -225,7 +225,7 @@ NanoPow (WebGPU) 3070 (0x400)
        "geometric": 168.54025739379932
 }
 
-NanoPow (WebGPU) 3070 (0x400)
+NanoPow (WebGPU) 3070 (dispatch 0x400 workgroup 256)
 (after converting `v` scalars to vectors)
 {
        "count": 512,
@@ -239,6 +239,175 @@ NanoPow (WebGPU) 3070 (0x400)
        "geometric": 162.9622640163259
 }
 
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 127778.10000002384,
+       "rate": 4.006946417264809,
+       "min": 15,
+       "max": 1386.9000000059605,
+       "arithmetic": 249.56660156254657,
+       "truncated": 176.96152343729045,
+       "harmonic": 88.48986919127314,
+       "geometric": 152.85173378506695
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 126437.00000026822,
+       "rate": 4.04944755094564,
+       "min": 8.299999982118607,
+       "max": 1592.5999999940395,
+       "arithmetic": 246.94726562552387,
+       "truncated": 197.15312500030268,
+       "harmonic": 83.44136709616109,
+       "geometric": 151.69548923418716
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x1000 workgroup 32)
+{
+       "count": 512,
+       "total": 130610.90000012517,
+       "rate": 3.920040364161868,
+       "min": 16.899999976158142,
+       "max": 1475.5,
+       "arithmetic": 255.09941406274447,
+       "truncated": 201.2101562502212,
+       "harmonic": 108.50998153048026,
+       "geometric": 169.9536026991119
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 133471.2999998033,
+       "rate": 3.836030667272699,
+       "min": 9.400000005960464,
+       "max": 1703.5,
+       "arithmetic": 260.68613281211583,
+       "truncated": 205.00488281232538,
+       "harmonic": 88.65601312209833,
+       "geometric": 160.81454554342605
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64) debug mode
+{
+       "count": 512,
+       "total": 138230.30000001192,
+       "rate": 3.70396360276984,
+       "min": 11.199999988079071,
+       "max": 1709.9000000059605,
+       "arithmetic": 269.9810546875233,
+       "truncated": 223.11933593766298,
+       "harmonic": 86.31741549131624,
+       "geometric": 164.2553862861393
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64) debug mode
+{
+       "count": 512,
+       "total": 137174.0999996364,
+       "rate": 3.732483027053628,
+       "min": 9.300000011920929,
+       "max": 1598.0999999940395,
+       "arithmetic": 267.91816406178987,
+       "truncated": 189.36367187427823,
+       "harmonic": 88.46031367791892,
+       "geometric": 157.678025720888
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64) debug mode
+{
+       "count": 512,
+       "total": 140824.9999999702,
+       "rate": 3.63571808982857,
+       "min": 12.100000023841858,
+       "max": 1640.300000011921,
+       "arithmetic": 275.0488281249418,
+       "truncated": 209.06249999988358,
+       "harmonic": 97.42050559825007,
+       "geometric": 174.1405646344008
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 256) debug mode
+{
+       "count": 512,
+       "total": 134499.49999976158,
+       "rate": 3.806705601142812,
+       "min": 24.200000017881393,
+       "max": 1324.5999999940395,
+       "arithmetic": 262.69433593703434,
+       "truncated": 206.15488281205762,
+       "harmonic": 117.14110375247057,
+       "geometric": 178.66910197648477
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 135615.80000019073,
+       "rate": 3.7753713062879095,
+       "min": 11.300000011920929,
+       "max": 1967.5,
+       "arithmetic": 264.87460937537253,
+       "truncated": 214.91972656268626,
+       "harmonic": 95.29029745129836,
+       "geometric": 166.91867186739316
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 129097.29999995232,
+       "rate": 3.9660008381289855,
+       "min": 9.099999994039536,
+       "max": 1335.9000000059605,
+       "arithmetic": 252.14316406240687,
+       "truncated": 200.86796874983702,
+       "harmonic": 87.70052168819247,
+       "geometric": 159.12504898549835
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 129204.69999992847,
+       "rate": 3.9627041431177306,
+       "min": 14.900000005960464,
+       "max": 1968.5,
+       "arithmetic": 252.3529296873603,
+       "truncated": 205.91660156229045,
+       "harmonic": 97.46470054285273,
+       "geometric": 162.08582278945588
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 64)
+{
+       "count": 512,
+       "total": 140579.30000019073,
+       "rate": 3.6420724815054943,
+       "min": 11.200000017881393,
+       "max": 1399.0999999940395,
+       "arithmetic": 274.56894531287253,
+       "truncated": 213.73515625024447,
+       "harmonic": 104.79799509895179,
+       "geometric": 179.31860807870936
+}
+
+NanoPow (WebGPU) 3070 (dispatch 0x800 workgroup 256)
+{
+       "count": 512,
+       "total": 138712.99999949336,
+       "rate": 3.691074376603996,
+       "min": 24.799999982118607,
+       "max": 1613.0999999940395,
+       "arithmetic": 270.92382812401047,
+       "truncated": 207.68144531175494,
+       "harmonic": 117.38847021261739,
+       "geometric": 179.3126486381509
+}
+
 
 
 NanoPowGpu: Time to calculate proof-of-work for a send block 32 times