Benchmarking integer clamping methods using BenchMarkDotNet

Clamp.md

BenchmarkDotNet=v0.10.1, OS=Microsoft Windows NT 6.2.9200.0
Processor=Intel(R) Core(TM) i7-6600U CPU 2.60GHz, ProcessorCount=4
Frequency=2742193 Hz, Resolution=364.6716 ns, Timer=TSC
  [Host]     : Clr 4.0.30319.42000, 64bit RyuJIT-v4.6.1586.0
  DefaultJob : Clr 4.0.30319.42000, 64bit RyuJIT-v4.6.1586.0

Allocated=0 B

Method	Value	Mean	StdErr	StdDev	Median	Scaled	Scaled-StdDev
'Maths Clamp'	-1	1.2373 ns	0.0154 ns	0.0595 ns	1.2511 ns	1.00	0.00
'No Maths Clamp'	-1	1.2183 ns	0.0147 ns	0.0570 ns	1.2248 ns	0.99	0.07
'No Maths No Equals Clamp'	-1	2.0156 ns	0.0107 ns	0.0399 ns	2.0256 ns	1.63	0.08
'No Maths Clamp No Ternary'	-1	1.1519 ns	0.0069 ns	0.0247 ns	1.1617 ns	0.93	0.05
'No Maths No Equals Clamp No Ternary'	-1	1.1801 ns	0.0145 ns	0.0562 ns	1.1920 ns	0.96	0.06
'Clamp using Bitwise Abs'	-1	0.9129 ns	0.0090 ns	0.0326 ns	0.9092 ns	0.74	0.04
'Maths Clamp'	0	1.2245 ns	0.0103 ns	0.0385 ns	1.2348 ns	1.00	0.00
'No Maths Clamp'	0	1.2442 ns	0.0116 ns	0.0450 ns	1.2560 ns	1.02	0.05
'No Maths No Equals Clamp'	0	1.9620 ns	0.0140 ns	0.0541 ns	1.9488 ns	1.60	0.07
'No Maths Clamp No Ternary'	0	1.1913 ns	0.0280 ns	0.1047 ns	1.1699 ns	0.97	0.09
'No Maths No Equals Clamp No Ternary'	0	1.9778 ns	0.0112 ns	0.0432 ns	1.9723 ns	1.62	0.06
'Clamp using Bitwise Abs'	0	0.9206 ns	0.0074 ns	0.0287 ns	0.9145 ns	0.75	0.03
'Maths Clamp'	255	1.2562 ns	0.0187 ns	0.0725 ns	1.2320 ns	1.00	0.00
'No Maths Clamp'	255	0.9064 ns	0.0299 ns	0.1157 ns	0.8482 ns	0.72	0.10
'No Maths No Equals Clamp'	255	1.9107 ns	0.0031 ns	0.0118 ns	1.9119 ns	1.53	0.08
'No Maths Clamp No Ternary'	255	0.9426 ns	0.0081 ns	0.0314 ns	0.9389 ns	0.75	0.05
'No Maths No Equals Clamp No Ternary'	255	1.9511 ns	0.0047 ns	0.0182 ns	1.9515 ns	1.56	0.08
'Clamp using Bitwise Abs'	255	0.9079 ns	0.0052 ns	0.0195 ns	0.9053 ns	0.72	0.04
'Maths Clamp'	256	1.1499 ns	0.0038 ns	0.0137 ns	1.1484 ns	1.00	0.00
'No Maths Clamp'	256	1.2231 ns	0.0114 ns	0.0410 ns	1.2134 ns	1.06	0.04
'No Maths No Equals Clamp'	256	2.0982 ns	0.0285 ns	0.1067 ns	2.0862 ns	1.82	0.09
'No Maths Clamp No Ternary'	256	0.9000 ns	0.0036 ns	0.0139 ns	0.8975 ns	0.78	0.01
'No Maths No Equals Clamp No Ternary'	256	0.7765 ns	0.0433 ns	0.1785 ns	0.6871 ns	0.68	0.15
'Clamp using Bitwise Abs'	256	1.1191 ns	0.0479 ns	0.2582 ns	0.9897 ns	0.97	0.22

@JimBobSquarePants as a quick solution PackFromBytes() is also OK. My main point is to perform the float->byte conversion in the end.

In long term we have to review the .PackFrom...() mechanism. It's a major, library-wide performance bottleneck in it's current form.