Xeon E7-8890 v4 4x24cores=96c/192t

	A	B	C	D	E	F	G	H	I	J	K	L
1	Xeon E7-8890 v4 4x24cores = 96cores/192threads
2	Operating system : Windows Server 2012 R2
3	HyperThreading(HT) = OFF
4	Tests done can be follow under this link:
5	http://www.ipmanchess.yolasite.com/testing-xeon-e7-8890-v4-96cores-192threads.php
6
7	Second run with Stockfish version without CounterMovesHistory(CMH)
8	Cores	8	16	20	22	24	32	44	48	64	80	96
9	%Faster		122,51	27,93	8,70	8,07	30,73	34,52	8,89	-0,28	2,63	-0,63
10	Nodes/sec.	14984877	33342846	42654404	46367168	50110196	65509679	88121814	95957082	95691348	98209077	97592638
11
12	Stockfish without CMH - dubbeling total cores									Stockfish without CMH
13	Cores	8	16	32	64					Cores	24	48
14	%Faster		122,51	96,47	46,07					%Faster		91,49
15	Nodes/sec.	14984877	33342846	65509679	95691348					Nodes/sec.	50110196	95957082
16
17	Stockfish 110616 last source bmi2 compile - First run with same Hash=1024					Stockfish without CMH for comparing TCEC9 final system				Stockfish with CMH as normal abrok version
18	Cores	24	64	96		Cores	22	44		Cores	24	48
19	%Faster		61,22	0,42		%Faster		90,05		%Faster		62,60
20	Nodes/sec.	44789208	72208319	72508544		Nodes/sec.	46367168	88121814		Nodes/sec.	40404128	65698485
21
22	Compare first & second run with same Hash=1024 & Cores
23	Cores	24	64	96
24	%Faster	11,88	32,52	34,59
25	Nodes/sec. R1	44789208	72208319	72508544
26	Nodes/sec. R2	50110196	95691348	97592638
27
28	Stockfish 110616 last source bmi2 compile - First run with same Hash=8192
29	Cores	8	16	24	32	48	64	80	96
30	%Faster		128,56	39,27	24,14	30,98	6,04	0,86	1,04
31	Nodes/sec.	12692959	29010783	40404128	50157804	65698485	69668345	70268202	70998622
32
33
34	Intel Xeon E7-8870 v3 @2.1Ghz 4x18cores=72cores/144threads system
35	Operating system : Windows Server 2012 R2
36	HyperThreading(HT) = OFF
37
38	3th run with Stockfish version without CounterMovesHistory(CMH) all test used same Hash=1024
39	Cores	8	16	18	22	24	32	36	44	54	72
40	%Faster		84,27	24,77	9,94	7,87	37,85	15,80	-0,06	0,87	1,06
41	Nodes/sec.	16732330	30832367	38470097	42293556	45620949	62890172	72825522	72778913	73414009	74190398
42
43
44	Stockfish without CMH using bench 1024 36 24 default depth				Stockfish without CMH for comparing TCEC9 final system
45	Cores	18	36		Cores	18	36
46	%Faster		99,67		%Faster		89,30
47	Nodes/sec.	24144143	48207412		Nodes/sec.	38470097	72825522
48	24 default depth bench takes much longer time
49	And shows almost perfect scaling from 18cores to 36cores
50	result from 1cpu to use 2cpu's!
51
52	Texel 1.06a48 512cores Numa - setoption name threads value x - go depth 22
53	Cores	8	16	18	22	32	36	44	54	64	72
54	%Faster		88,50	14,98	12,82	24,35	2,32	-32,05	-17,73	13,54	-0,08
55	Nodes/sec.	7208965	13589203	15624894	17628099	21920702	22428969	15240947	12538295	14235759	14224983
56
57	Texel 1.06a48 512cores Numa - setoption name threads value x - go depth 22
58	Cores	18	36
59	%Faster		43,55
60	Nodes/sec.	15624894	22428969
61
62	Andscacs 0.86189 - setoption name threads value x - go depth 25
63	Cores	8	16	18	22	36	44	54	64	72
64	%Faster		108,59	6,27	22,28	41,20	1,52	-2,43	3,09	-2,08
65	Nodes/sec.	6482131	13520811	14368402	17569682	24808340	25184920	24573952	25332783	24806192
66
67	Andscacs 0.86189 - setoption name threads value x - go depth 25				Andscacs 0.86189 - setoption name threads value x - go depth 26				Andscacs 0.86189 - setoption name threads value x - go depth 27
68	Cores	18	36		Cores	18	36		Cores	18	36
69	%Faster		72,66		%Faster		68,05		%Faster		72,42
70	Nodes/sec.	14368402	24808340		Nodes/sec.	14963031	25144641		Nodes/sec.	14579126	25137979
71
72	Houdini 4 Pro B Numa 32cores				Komodo 1656 200cores				Komodo 1656 200cores
73	Cores	16	32		Cores	24	48		Cores	8	16
74	%Faster		67,48		%Faster		17,89		%Faster		112,77
75	Nodes/sec.	18732000	31373000		Nodes/sec.	18059567	21289807		Nodes/sec.	7942805	16899827
76
77	Scaling from 1cpu to 2cpu's		Procent		Crafty 25.0.1				asmFish 2016.06.13 bmi2
78	Stockfish without CMH		89% - 99,67%		Cores	18	36		Cores	24	48
79	Andscacs 0.86189		68% - 72,66%		%Faster		5,75		%Faster		70,19
80	asmFish 2016.06.13 bmi2		70,19%		Nodes/sec.	38004860	40191380		Nodes/sec.	24635035	41925656
81	Houdini 4 Pro B Numa		67,48%%
82	Stockfish with CMH normal ver.		62,60%						DON 100616 256cores
83	DON 100616		51,10%						Cores	24	48
84	Texel 1.06a48		43,55%						%Faster		51,10
85	Komodo 1656		17,89%						Nodes/sec.	30890862	46676864
86	Crafty 25.0.1		5,75%
87
88	Texel 1.06a48 512cores Numa - setoption name threads value x - go depth 22 - Hash=1024
89	Cores	8	16	18	22	32	36	44	54	64	72
90	%Faster		89,67	10,23	13,33	49,22	13,45	20,03	15,37	14,96	10,73
91	Nodes/sec.	8781890	16656990	18361567	20808858	31049985	35225499	42282183	48781273	56077700	62092922
92
93	Texel 1.06a48 512cores Numa - compare from 1cpu-2cpu's-3cpu's-4cpu's								Texel 1.06a48 512cores Numa - go depth 24
94	Cores	18	36	54	72				Cores	64	72
95	%Faster		91,84	38,48	27,29				%Faster		10,26
96	Nodes/sec.	18361567	35225499	48781273	62092922				Nodes/sec.	63032176	69499644
97
98	Texel 1.06a48 512cores Numa - dubbling cores
99	Cores	36	72
100	%Faster		76,27