C# .NET versus C# pgo fastest performance

Always look at the source code.

If the fastest programs are hand-written vector instructions, does the host language matter? You might be more interested in the less optimised programs — more seconds, less gz source code.

fannkuch-redux
source	secs	mem	gz	cpu secs
C# .NET #8	2.23	31,164	1485	8.60
C# .NET #6	8.65	30,684	1075	34.32
C# .NET #9	9.07	32,108	1232	35.94
C# .NET #7		Bad Output
n-body
source	secs	mem	gz	cpu secs
C# .NET #9	3.16	30,044	2480	3.15
C# .NET #4	3.79	30,088	2137	3.77
C# .NET #7	4.79	30,216	1715	4.77
C# .NET #6	4.82	30,196	1754	4.80
C# .NET #3	6.64	31,212	1311	6.62
C# .NET #2	6.82	31,168	1350	6.81
C# .NET #8	6.82	31,208	1339	6.81
spectral-norm
source	secs	mem	gz	cpu secs
C# .NET #5	0.93	36,020	776	3.34
C# .NET #3	1.56	33,348	890	5.85
mandelbrot
source	secs	mem	gz	cpu secs
C# .NET	3.23	66,812	1980	12.56
C# .NET #9	3.23	66,560	822	12.59
C# .NET #5	4.02	64,276	851	15.75
pidigits
source	secs	mem	gz	cpu secs
C# .NET #5	0.92	34,692	990	0.94
C# .NET #6	0.96	32,836	1182	0.94
C# .NET #4	1.36	33,556	1135	1.35
C# .NET #7	6.49	53,064	579	7.05
regex-redux
source	secs	mem	gz	cpu secs
C# .NET #8	1.36	388,756	1881	2.60
C# .NET #5	2.07	861,292	679	6.28
fasta
source	secs	mem	gz	cpu secs
C# .NET #5	1.48	49,500	1702	5.38
C# .NET	1.83	52,060	2113	6.61
k-nucleotide
source	secs	mem	gz	cpu secs
C# .NET #6	4.89	187,780	1615	12.63
C# .NET		Make Error
reverse-complement
source	secs	mem	gz	cpu secs
C# .NET #8	0.55	718,428	2641	1.30
C# .NET #5	2.60	1,026,320	1620	3.58
C# .NET #6	1.58	1,027,952	1627	3.80
C# .NET #7	1.53	942,900	1312	4.16
binary-trees
source	secs	mem	gz	cpu secs
C# .NET #2	4.48	1,624,564	717	14.25
C# .NET #7	4.54	1,608,796	706	14.53
C# .NET #3	4.84	2,089,712	688	15.89
C# .NET	5.08	2,046,692	668	16.32
C# .NET	.NET SDK 8.0.200 Host Version: 8.0.2 Commit: 1381d5ebd2 <ServerGarbageCollection>true
C# pgo	.NET SDK 7.0.200 Host Version: 7.0.3 Commit: 0a2bda10e8 <ServerGarbageCollection>true <TieredPGO>true