Optimized Encoder.Encode()

[greymistcube]

Just squeezing out more performance. 🙄
Depending on the case, the speed gain seems to be about 20% ~ 30%.
On the memory side, the memory usage seems go down more drastically as the collection size gets bigger.

This is achieved through the following:

• Removed value argument passing and returning values for tracking the encoder state.
• Other than declaring temporary local variables, the global offset is tracked using a single variable using ref.

As a bonus, the code became more readable as well. Also a bug was fixed where an unused overloaded Encode() method was wrongly implemented (not sure if we still need to keep the method tho). 😗

Encode() original

Method	SetSize	WordSize	Mean	Error	StdDev	Gen0	Allocated
EncodeList	8	8	3.838 ms	0.6533 ms	0.3888 ms	265.6250	1.6 MB
EncodeDict	8	8	13.948 ms	1.2701 ms	0.8401 ms	453.1250	2.73 MB
EncodeNestedList	8	8	35.165 ms	6.0021 ms	3.9700 ms	1923.0769	11.65 MB
EncodeNestedDict	8	8	160.819 ms	25.9170 ms	17.1425 ms	3500.0000	21.6 MB
EncodeList	8	16	5.465 ms	0.1529 ms	0.1012 ms	789.0625	4.73 MB
EncodeDict	8	16	17.626 ms	0.6054 ms	0.3603 ms	1500.0000	8.99 MB
EncodeNestedList	8	16	45.677 ms	1.5729 ms	1.0404 ms	6090.9091	36.67 MB
EncodeNestedDict	8	16	168.689 ms	2.9600 ms	1.9579 ms	12333.3333	74.78 MB
EncodeList	16	8	7.854 ms	0.1353 ms	0.0895 ms	390.6250	2.42 MB
EncodeDict	16	8	31.121 ms	1.2161 ms	0.8044 ms	718.7500	4.33 MB
EncodeNestedList	16	8	130.883 ms	1.7192 ms	1.0231 ms	5750.0000	35.3 MB
EncodeNestedDict	16	8	521.700 ms	10.7511 ms	7.1112 ms	11000.0000	67.73 MB
EncodeList	16	16	10.420 ms	0.4796 ms	0.2854 ms	1437.5000	8.68 MB
EncodeDict	16	16	33.950 ms	1.6944 ms	1.1208 ms	2785.7143	16.84 MB
EncodeNestedList	16	16	160.629 ms	3.9966 ms	2.6435 ms	22500.0000	135.4 MB
EncodeNestedDict	16	16	581.700 ms	5.6855 ms	3.3833 ms	45000.0000	274.18 MB

Encode() optimized

Method	SetSize	WordSize	Mean	Error	StdDev	Gen0	Allocated
EncodeList	8	8	3.322 ms	0.5204 ms	0.3442 ms	265.6250	1.61 MB
EncodeDict	8	8	13.067 ms	1.4576 ms	0.9641 ms	453.1250	2.73 MB
EncodeNestedList	8	8	30.082 ms	2.4822 ms	1.4771 ms	1937.5000	11.65 MB
EncodeNestedDict	8	8	124.848 ms	11.0279 ms	7.2943 ms	3500.0000	21.61 MB
EncodeList	8	16	3.674 ms	0.4520 ms	0.2989 ms	382.8125	2.29 MB
EncodeDict	8	16	13.914 ms	1.5466 ms	1.0230 ms	671.8750	4.1 MB
EncodeNestedList	8	16	34.074 ms	4.8878 ms	3.2330 ms	2800.0000	17.14 MB
EncodeNestedDict	8	16	138.955 ms	11.0044 ms	7.2787 ms	5500.0000	33.28 MB
EncodeList	16	8	7.396 ms	0.7947 ms	0.5257 ms	398.4375	2.42 MB
EncodeDict	16	8	27.286 ms	1.7213 ms	1.1385 ms	718.7500	4.33 MB
EncodeNestedList	16	8	108.680 ms	4.5182 ms	2.9885 ms	5800.0000	35.3 MB
EncodeNestedDict	16	8	471.722 ms	9.6321 ms	6.3710 ms	11000.0000	67.73 MB
EncodeList	16	16	7.570 ms	0.1171 ms	0.0775 ms	632.8125	3.8 MB
EncodeDict	16	16	27.453 ms	0.9822 ms	0.6496 ms	1156.2500	7.07 MB
EncodeNestedList	16	16	114.874 ms	4.2063 ms	2.7822 ms	9400.0000	57.27 MB
EncodeNestedDict	16	16	462.164 ms	56.9407 ms	37.6628 ms	18000.0000	113.04 MB

[CLAassistant]

All committers have signed the CLA.