Highest depth ever achieved

[wihdv]

Hello!

I just saw a guy that went in depth 70 with no searchmoves and 41 192 897 168 184 nodes from the start position in stockfish few years ago.

I would like to know if this is still a record or if someone has achieved an higher depth/nodes with stockfish or equivalent in the start position?

Thanks!

[Milos]

Hello!

I just saw a guy that went in depth 70 with no searchmoves and 41 192 897 168 184 nodes from the start position in stockfish few years ago.

I would like to know if this is still a record or if someone has achieved an higher depth/nodes with stockfish or equivalent in the start position?

Thanks!

That's like 2 weeks analysis on a moderately powerful machine and using NNUE. Nothing special for today's standards.

[wihdv]

Thanks for answering! I'm pretty new to this domain.

Oh yes that's not so long but have you seen results with more depth/nodes or not?

[Dann Corbit]

I guess it would not take very long for someone with a really high end machine like Bojung Guo to beat it.
But the fascination for deepest search from the root has sort of worn off.

[wihdv]

Yeah it doesn't have much sense I know but I'm still really interested by that, could you explain me what is Bojung Guo?

Thanks

[Jouni]

There seems to be great progress after 2018 test. It took 162.433.347.175 nodes for 50 plies. But in my test SF dev took only 18.214.961.106 nodes! I am sure that with CCC hardware (2x EPYC=256 threads) it reach 80 plies in one week. May be 90-100? Enough to solve chess .

[Jouni]

In second try 50 plies after 7 giganodes:

Analysis by Stockfish 110121:

...
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.0-0 Nxe4 5.Re1 Nd6 6.Nxe5 Be7 7.Bf1 Nxe5 8.Rxe5 0-0 9.Nc3 Ne8 10.Nd5 Bd6 11.Re1 c6 12.Ne3 Be7 13.c4 Nc7 14.d4 d5 15.cxd5 Nxd5 16.Nxd5 cxd5 17.Bf4 Be6 18.Be5 Bd6 19.g3 a5 20.h4 Bxe5 21.Rxe5 Qb6 22.Qd2 h6 23.Re3 a4 24.Rae1 Rac8 25.Re5 Rc7 26.h5 Rfc8 27.Kg2
= (0.30) Depth: 49/62 00:16:19 7216mN, tb=6
1.e4 e5
= (0.22 --) Depth: 50/59 00:17:09 7588mN, tb=6

Komodo11 needed 4.606.843.996.444 nodes.

[xr_a_y]

Stockfish really has an impressive EBF !

Look how Minic behaves in terms of time and nodes. Already 8.569.919.342 at depth 35 (using 1 thread) !

Code: Select all

depth	cp	time	nodes	nps	seldepth
1	102	1	21	20999	1
2	45	1	62	61999	2
3	91	1	90	89999	3
4	58	1	171	170999	4
5	53	1	475	474999	6
6	72	2	806	402999	6
7	74	7	2384	340571	9
8	63	13	5012	385538	12
9	50	28	14996	535571	15
10	51	56	34560	617142	16
11	52	102	66121	648245	21
12	51	154	104550	678896	24
13	52	265	180393	680728	25
14	48	431	297078	689276	28
15	64	948	649614	685246	28
16	59	1622	1117842	689175	29
17	50	2642	1830673	692911	29
18	60	3664	2542736	693978	33
19	57	6268	4327811	690461	33
20	59	12521	8604702	687221	34
21	61	18760	12899503	687606	34
22	64	45110	30865052	684217	39
23	69	54193	37079002	684202	39
24	63	120101	82216957	684565	42
25	49	175372	120087800	684760	42
26	61	196574	134506434	684253	42
27	61	249760	170427743	682366	43
28	63	545162	370802370	680169	46
29	57	688936	467841235	679078	46
30	62	806498	546633624	677786	46
31	62	1104246	746908691	676397	46
32	53	2324883	1566373300	673742	49
33	51	4349627	2928430158	673260	53
34	64	6416503	4314360409	672385	53
35	53	12759884	8569919342	671629	55

[Uri Blass]

Stockfish really has an impressive EBF !

Look how Minic behaves in terms of time and nodes. Already 8.569.919.342 at depth 35 (using 1 thread) !

Code: Select all

depth	cp	time	nodes	nps	seldepth
1	102	1	21	20999	1
2	45	1	62	61999	2
3	91	1	90	89999	3
4	58	1	171	170999	4
5	53	1	475	474999	6
6	72	2	806	402999	6
7	74	7	2384	340571	9
8	63	13	5012	385538	12
9	50	28	14996	535571	15
10	51	56	34560	617142	16
11	52	102	66121	648245	21
12	51	154	104550	678896	24
13	52	265	180393	680728	25
14	48	431	297078	689276	28
15	64	948	649614	685246	28
16	59	1622	1117842	689175	29
17	50	2642	1830673	692911	29
18	60	3664	2542736	693978	33
19	57	6268	4327811	690461	33
20	59	12521	8604702	687221	34
21	61	18760	12899503	687606	34
22	64	45110	30865052	684217	39
23	69	54193	37079002	684202	39
24	63	120101	82216957	684565	42
25	49	175372	120087800	684760	42
26	61	196574	134506434	684253	42
27	61	249760	170427743	682366	43
28	63	545162	370802370	680169	46
29	57	688936	467841235	679078	46
30	62	806498	546633624	677786	46
31	62	1104246	746908691	676397	46
32	53	2324883	1566373300	673742	49
33	51	4349627	2928430158	673260	53
34	64	6416503	4314360409	672385	53
35	53	12759884	8569919342	671629	55

I am not impressed by depth number when there is a lot of pruning.
It is easy to make it bigger without making the engine better.

[xr_a_y]

Stockfish really has an impressive EBF !

Look how Minic behaves in terms of time and nodes. Already 8.569.919.342 at depth 35 (using 1 thread) !

Code: Select all

depth	cp	time	nodes	nps	seldepth
1	102	1	21	20999	1
2	45	1	62	61999	2
3	91	1	90	89999	3
4	58	1	171	170999	4
5	53	1	475	474999	6
6	72	2	806	402999	6
7	74	7	2384	340571	9
8	63	13	5012	385538	12
9	50	28	14996	535571	15
10	51	56	34560	617142	16
11	52	102	66121	648245	21
12	51	154	104550	678896	24
13	52	265	180393	680728	25
14	48	431	297078	689276	28
15	64	948	649614	685246	28
16	59	1622	1117842	689175	29
17	50	2642	1830673	692911	29
18	60	3664	2542736	693978	33
19	57	6268	4327811	690461	33
20	59	12521	8604702	687221	34
21	61	18760	12899503	687606	34
22	64	45110	30865052	684217	39
23	69	54193	37079002	684202	39
24	63	120101	82216957	684565	42
25	49	175372	120087800	684760	42
26	61	196574	134506434	684253	42
27	61	249760	170427743	682366	43
28	63	545162	370802370	680169	46
29	57	688936	467841235	679078	46
30	62	806498	546633624	677786	46
31	62	1104246	746908691	676397	46
32	53	2324883	1566373300	673742	49
33	51	4349627	2928430158	673260	53
34	64	6416503	4314360409	672385	53
35	53	12759884	8569919342	671629	55

I am not impressed by depth number when there is a lot of pruning.
It is easy to make it bigger without making the engine better.

Indeed, but i this case, Stockfish is of course way deeper AND way stronger than Minic.