'Best of four' on AMD Ryzen1800 TC 3m2s 12threads

[beram]

First tournament played on my AMd Ryzen 1800X with the top three engines Stockfish 030417, Komodo 10.4, Houdini 5 and Fizbo 1.9 because of its attractive agressive approach of the game
TC 3m2s, each engine 12 threads, 4 Gb hash, privat testsuite 25 lines, round robin
Sensational good result for Komodo 10.4 coming second only two points behind Stockfish beating Houdini 5 58% and ‘just’ loosing with 54% (23-27) against latest SF dev 030417(ultimaq compile)
Will post some games later

[beram]

Two nice wins
First the shortest win of the tournament by Fizbo1.9 in a Kings Gambit just 17 moves to reach +9 and a win over the strong Komodo
Second game a convincingly strong black win for Komodo over Stockfish

[pgn]
[Event "four best april 2017"]
[Site "?"]
[Date "2017.04.08"]
[Round "25.3"]
[White "Komodo 10.4 64-bit"]
[Black "Fizbo 1.9"]
[Result "0-1"]
[ECO "C37"]
[Annotator "-0.43;-0.73"]
[PlyCount "34"]
[EventDate "2017.04.07"]
[EventType "tourn"]
[Source "Bram"]

{AMD Ryzen 7 1800X Eight-Core Processor 3593 MHz W=23.1 plies; 14.
858kN/s B=18.9 plies; 7.412kN/s} 1. e4 e5 2. f4 exf4 3. Nf3 g5 {[%eval -73,17]
[%emt 0:00:05]} 4. Bc4 {[%eval -43,23] [%emt 0:00:17] (h2-h3)} g4 {[%eval -121,
18] [%emt 0:00:06]} 5. Ne5 {[%eval -41,23] [%emt 0:00:14] (Pf3-g1)} Qh4+ {
[%eval -160,17] [%emt 0:00:05]} 6. Kf1 {[%eval -49,23] [%emt 0:00:06]} Nc6 {
[%eval -156,19] [%emt 0:00:09] (Pg8-f6)} 7. Nxf7 {[%eval -30,22] [%emt 0:00:04]
} Bc5 {[%eval -194,20] [%emt 0:00:06] (f4-f3)} 8. Qe1 {[%eval 84,22] [%emt 0:
00:05] (d2-d4)} g3 {[%eval -272,19] [%emt 0:00:11] (Dh4-f6)} 9. Nxh8 {[%eval 0,
21] [%emt 0:00:06]} Nf6 {[%eval -168,19] [%emt 0:00:14] (Pg8-h6)} 10. Bf7+ {
[%eval -128,23] [%emt 0:00:24] (d2-d4)} Kd8 {[%eval -384,18] [%emt 0:00:13]
(Ke8-f8)} 11. c3 {[%eval -161,22] [%emt 0:00:09] (d2-d3)} Bf2 {[%eval -319,18]
[%emt 0:00:10]} 12. Qd1 {[%eval -351,24] [%emt 0:00:06]} d6 {[%eval -435,18]
[%emt 0:00:13]} 13. h3 {[%eval -396,24] [%emt 0:00:04] (d2-d4)} Bg4 {[%eval
-509,17] [%emt 0:00:04]} 14. Qc2 {[%eval -451,24] [%emt 0:00:04] (Dd1-b3)} Ne5
{[%eval -690,19] [%emt 0:00:03] (Kd8-e7)} 15. d4 {[%eval -748,22] [%emt 0:00:
22]} Nf3 {[%eval -771,19] [%emt 0:00:04]} 16. Bxf4 {[%eval -827,24] [%emt 0:00:
03]} Ne1 {[%eval -878,20] [%emt 0:00:07]} 17. Qd2 {[%eval -882,27] [%emt 0:00:
07]} Bf3 {[%eval -1031,22] [%emt 0:00:03]} 0-1

[Event "four best april 2017"]
[Site "?"]
[Date "2017.04.07"]
[Round "1.2"]
[White "Stockfish 030417 64 POPCNT"]
[Black "Komodo 10.4 64-bit"]
[Result "0-1"]
[ECO "B22"]
[Annotator "0.07;-0.12"]
[PlyCount "90"]
[EventDate "2017.04.07"]
[EventType "tourn"]
[Source "Bram"]

{AMD Ryzen 7 1800X Eight-Core Processor 3593 MHz W=26.5 plies; 16.
767kN/s B=26.2 plies; 14.630kN/s} 1. e4 c5 2. c3 b6 3. d4 Bb7 4. Bd3 Nf6 5.
Qe2 cxd4 6. cxd4 Nc6 7. Nf3 Nb4 8. Nc3 Nxd3+ 9. Qxd3 e6 {[%eval -12,21] [%emt
0:00:03]} 10. O-O {[%eval 7,23] [%emt 0:00:06] (Lc1-g5)} Bb4 {[%eval -13,25]
[%emt 0:00:09]} 11. e5 {[%eval -4,29] [%emt 0:00:08] (Tf1-e1)} Nd5 {[%eval -14,
26] [%emt 0:00:06]} 12. Bg5 {[%eval -16,28] [%emt 0:00:03]} f6 {[%eval -13,25]
[%emt 0:00:06] (Lb4-e7)} 13. Nxd5 {[%eval 12,28] [%emt 0:00:04]} Bxd5 {[%eval
-17,24] [%emt 0:00:03]} 14. exf6 {[%eval 0,26] [%emt 0:00:02]} gxf6 {[%eval
-17,24] [%emt 0:00:04]} 15. Bf4 {[%eval 11,27] [%emt 0:00:07]} Qe7 {[%eval 0,
26] [%emt 0:00:10]} 16. Rac1 {[%eval 9,27] [%emt 0:00:07] (a2-a3)} Rg8 {[%eval
-21,25] [%emt 0:00:08] (0-0)} 17. Bg3 {[%eval 0,28] [%emt 0:00:07]} Qg7 {
[%eval -30,24] [%emt 0:00:07]} 18. Kh1 {[%eval 0,27] [%emt 0:00:07] (a2-a3)}
Be7 {[%eval -33,24] [%emt 0:00:14]} 19. b3 {[%eval -10,28] [%emt 0:00:27]
(Tf1-e1)} Qg6 {[%eval -40,23] [%emt 0:00:08]} 20. Qb5 {[%eval -15,24] [%emt 0:
00:08] (Dd3-e2)} Rd8 {[%eval -47,24] [%emt 0:00:05]} 21. Rfe1 {[%eval 0,26]
[%emt 0:00:13]} Qh5 {[%eval -46,25] [%emt 0:00:12] (h7-h5)} 22. Qd3 {[%eval
-21,25] [%emt 0:00:10]} Kf7 {[%eval -44,25] [%emt 0:00:11] (Le7-b4)} 23. Rc7 {
[%eval -10,24] [%emt 0:00:05]} a5 {[%eval -58,24] [%emt 0:00:04]} 24. Ra7 {
[%eval -33,26] [%emt 0:00:20] (Te1-g1)} Bb4 {[%eval -46,23] [%emt 0:00:06]} 25.
Rg1 {[%eval -24,25] [%emt 0:00:09]} Ke7 {[%eval -79,26] [%emt 0:00:12]} 26. Bc7
{[%eval -63,28] [%emt 0:00:24] (a2-a3)} Rc8 {[%eval -114,23] [%emt 0:00:03]}
27. h3 {[%eval -135,29] [%emt 0:00:07]} Bd6 {[%eval -153,26] [%emt 0:00:04]}
28. Bxd6+ {[%eval -155,25] [%emt 0:00:01]} Kxd6 {[%eval -160,27] [%emt 0:00:02]
} 29. Ne1 {[%eval -154,29] [%emt 0:00:08] (Ta7-a6)} Qf5 {[%eval -207,26] [%emt
0:00:09]} 30. Qe3 {[%eval -166,28] [%emt 0:00:01] (Dd3xf5)} Qg5 {[%eval -169,
27] [%emt 0:00:07]} 31. Qe2 {[%eval -160,30] [%emt 0:00:11]} Ke7 {[%eval -202,
26] [%emt 0:00:03] (Tc8-c3)} 32. f3 {[%eval -129,25] [%emt 0:00:03]} Qg3 {
[%eval -204,27] [%emt 0:00:05] (Dg5-f4)} 33. Ra6 {[%eval -118,28] [%emt 0:00:
04]} Rg5 {[%eval -342,27] [%emt 0:00:08] (Dg3-c7)} 34. Rxb6 {[%eval -214,29]
[%emt 0:00:08] (Tg1-f1)} Rh5 {[%eval -412,31] [%emt 0:00:17] (Tc8-c1)} 35. Rb5
{[%eval -258,26] [%emt 0:00:01]} Rg8 {[%eval -375,31] [%emt 0:00:03] (Tc8-c1)}
36. Qf1 {[%eval -139,23] [%emt 0:00:02]} Rg6 {[%eval -407,30] [%emt 0:00:04]
(a5-a4)} 37. Rxd5 {[%eval -349,26] [%emt 0:00:07]} Rxd5 {[%eval -425,32] [%emt
0:00:16]} 38. f4 {[%eval -372,26] [%emt 0:00:07] (Pe1-d3)} Rxd4 {[%eval -508,
27] [%emt 0:00:02]} 39. Nf3 {[%eval -412,29] [%emt 0:00:15]} Rxf4 {[%eval -766,
28] [%emt 0:00:06] (Td4-d5)} 40. Qe2 {[%eval -418,24] [%emt 0:00:05] (a2-a3)}
Rxf3 {[%eval -983,27] [%emt 0:00:03] (Tf4-f5)} 41. Qxf3 {[%eval -318,23] [%emt
0:00:02]} Qxf3 {[%eval -1046,29] [%emt 0:00:02]} 42. gxf3 {[%eval -326,21]
[%emt 0:00:01]} Rxg1+ {[%eval -1180,27] [%emt 0:00:04] (d7-d5)} 43. Kxg1 {
[%eval -1540,29] [%emt 0:00:08]} Kd6 {[%eval -1281,28] [%emt 0:00:02]} 44. a3 {
[%eval -2078,23] [%emt 0:00:05] (Kg1-f2)} Kd5 {[%eval -1533,27] [%emt 0:00:04]
(Kd6-c5)} 45. Kf2 {[%eval -5271,31] [%emt 0:00:05] (Kg1-g2)} Kd4 {[%eval -1539,
30] [%emt 0:00:14]} 0-1
[/pgn]

[Dann Corbit]

Were SF and Komodo specifically compiled for AMD?
If not, the giant leap above the other two does not make sense to me.

[beram]

Were SF and Komodo specifically compiled for AMD?
If not, the giant leap above the other two does not make sense to me.

Not specifically compiled versions
Btw, there is non such as a giant leap above Fizbo, because 26% overall score against top three of Fizbo is in line with the scores at CCRL 40/4 there Fizbo scores mixed 23-24%
Only special score is that of Komodo 10.4 58% against Houdini 4. But as we have to say just 50 games

[Waschbaer]

Did you check whether Houdini got the license data?

I ask because on my system after upgrading to new mainboard and ryzen CPU Houdini played sometimes very strange moves and lost.

Reason: Did not get the license information,
a new install of Houdini solved that problem.

Just start Houdini from command line and read the output.