perf

[bitfaster]

Inheriting from IBoundedPolicy/ITimePolicy comes with some overhead, so use proxy. Caching an instance of proxy inside lru core also results in the same penalty.

Before:

Method	Runtime	Mean	Error	Ratio	Code Size	Allocated
ConcurrentDictionary	.NET 6.0	7.368 ns	0.0568 ns	1.00	1,523 B	-
FastConcurrentLru	.NET 6.0	10.248 ns	0.2043 ns	1.39	2,352 B	-
ConcurrentLru	.NET 6.0	13.693 ns	0.1725 ns	1.86	2,378 B	-
AtomicFastLru	.NET 6.0	19.988 ns	0.1546 ns	2.71	846 B	-
FastConcurrentTLru	.NET 6.0	26.544 ns	0.3709 ns	3.61	2,544 B	-
ConcurrentTLru	.NET 6.0	32.986 ns	0.3864 ns	4.48	2,619 B	-
ClassicLru	.NET 6.0	48.563 ns	0.3721 ns	6.60	3,041 B	-
RuntimeMemoryCacheGet	.NET 6.0	121.734 ns	2.3760 ns	16.60	49 B	32 B
ExtensionsMemoryCacheGet	.NET 6.0	53.987 ns	0.4352 ns	7.33	78 B	24 B
ConcurrentDictionary	.NET Framework 4.8	14.333 ns	0.2964 ns	1.00	4,207 B	-
FastConcurrentLru	.NET Framework 4.8	15.140 ns	0.2779 ns	1.06	15,182 B	-
ConcurrentLru	.NET Framework 4.8	17.627 ns	0.3523 ns	1.23	15,212 B	-
AtomicFastLru	.NET Framework 4.8	39.460 ns	0.7806 ns	2.76	358 B	-
FastConcurrentTLru	.NET Framework 4.8	45.415 ns	0.5684 ns	3.17	15,358 B	-
ConcurrentTLru	.NET Framework 4.8	47.494 ns	0.7391 ns	3.31	15,403 B	-
ClassicLru	.NET Framework 4.8	61.447 ns	0.8935 ns	4.28	6,945 B	-
RuntimeMemoryCacheGet	.NET Framework 4.8	282.429 ns	3.1541 ns	19.72	33 B	32 B
ExtensionsMemoryCacheGet	.NET Framework 4.8	114.503 ns	1.4632 ns	7.99	82 B	24 B

After:

Method	Runtime	Mean	Error	Ratio	Code Size	Allocated
ConcurrentDictionary	.NET 6.0	8.132 ns	0.1803 ns	1.00	1,523 B	-
FastConcurrentLru	.NET 6.0	10.059 ns	0.0674 ns	1.24	2,352 B	-
ConcurrentLru	.NET 6.0	13.665 ns	0.0694 ns	1.68	2,378 B	-
AtomicFastLru	.NET 6.0	19.910 ns	0.3562 ns	2.45	846 B	-
FastConcurrentTLru	.NET 6.0	25.743 ns	0.2754 ns	3.17	2,544 B	-
ConcurrentTLru	.NET 6.0	29.895 ns	0.0757 ns	3.68	2,619 B	-
ClassicLru	.NET 6.0	48.791 ns	0.8167 ns	6.01	3,041 B	-
RuntimeMemoryCacheGet	.NET 6.0	117.031 ns	1.9028 ns	14.44	49 B	32 B
ExtensionsMemoryCacheGet	.NET 6.0	54.898 ns	1.0983 ns	6.77	78 B	24 B
ConcurrentDictionary	.NET Framework 4.8	13.810 ns	0.1752 ns	1.00	4,207 B	-
FastConcurrentLru	.NET Framework 4.8	15.032 ns	0.3198 ns	1.09	15,182 B	-
ConcurrentLru	.NET Framework 4.8	17.699 ns	0.3127 ns	1.28	15,212 B	-
AtomicFastLru	.NET Framework 4.8	40.034 ns	0.3762 ns	2.90	358 B	-
FastConcurrentTLru	.NET Framework 4.8	45.270 ns	0.1457 ns	3.28	15,358 B	-
ConcurrentTLru	.NET Framework 4.8	46.651 ns	0.7994 ns	3.38	15,403 B	-
ClassicLru	.NET Framework 4.8	61.811 ns	0.8297 ns	4.47	6,945 B	-
RuntimeMemoryCacheGet	.NET Framework 4.8	277.511 ns	1.1883 ns	20.08	33 B	32 B
ExtensionsMemoryCacheGet	.NET Framework 4.8	113.441 ns	1.1372 ns	8.21	82 B	24 B