buying a new computer

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Daniel Shawul
Posts: 4185
Joined: Tue Mar 14, 2006 11:34 am
Location: Ethiopia

Re: buying a new computer

Post by Daniel Shawul »

Daniel Shawul wrote: Fri Jul 19, 2019 3:26 pm
crem wrote: Fri Jul 19, 2019 9:36 am
Daniel Shawul wrote: Fri Jul 19, 2019 4:49 am
I don't think lc0 consumes that much memory and from my calculations it could go upto 17 hours analysis with just 12 gb ram.
Thats one of the reason i decided to stick with 16 gb ram.

Daniel
1 node takes 250 bytes, or in other words 1GB is needed for 4M nodes.
(and also one NN cache entry takes 350 bytes. It doesn't grow with time, but it's a thing to consider when setting huge cache sizes).

So at 33333 nps it takes 1GB per 2 minutes, per hour it takes 30GB, and after 17 hours it needs 512GB of RAM.

If 12GB is enough for 17 hours (and it overflows into swap partition), it seems that average nps is around 780.
I thought you are using "10 bytes per leaf, and 24 bytes per internal node" according to your post
https://github.com/LeelaChessZero/lc0/issues/13

So if i divide that by a factor of (250 bytes / 24 bytes = 10), it will take 51 GB for 17 hour analysis. Also i used an nps of 20knps.
I did 2-hours analysis of 10 positions for Dann Corbit a while ago

Code: Select all

info depth 30 seldepth 73 time 3616851 nodes 76130139 score cp 18 hashfull 1000 nps 21048 tbhits 0 pv c1b2 d7d5 e4d5 e6d5 f1b5 b8c6 d1e2 g8e7 g1f3 c8e6 f3g5 e6f5 b5c6 b7c6 b1c3 f7f6 g5e6 f5e6 e2e6 d8d7 e6d7 e8d7 c3a4 e7g6 e1g1 g6f4 f1e1 f8d6 e1e3 f4e6 g1f1 d6f4 e3d3 f4d6 a1e1 h7h5 d3e3 a8e8 h2h4 e6d4 c2c4 d4c2 e3e8 h8e8 e1e8 d7e8 f1e2 e8f7 d2d3 f7e6 e2d2 c2b4
So per hour it was taking 76 million nodes with an nps of 21 knps. So if i take 24 bytes / per node, in 17 hours it uses up 28 Gb.
I see my mistake now the nodes per second (nps) doesn't actually tell the number of nodes actually generated.
Unviisted nodes will increase memory consumtpion a factor of the averaging branching factor ...
crem
Posts: 177
Joined: Wed May 23, 2018 9:29 pm

Re: buying a new computer

Post by crem »

Daniel Shawul wrote: Fri Jul 19, 2019 3:26 pm
I thought you are using "10 bytes per leaf, and 24 bytes per internal node" according to your post
https://github.com/LeelaChessZero/lc0/issues/13

So if i divide that by a factor of (250 bytes / 24 bytes = 10), it will take 51 GB for 17 hour analysis. Also i used an nps of 20knps.
I did 2-hours analysis of 10 positions for Dann Corbit a while ago

Code: Select all

info depth 30 seldepth 73 time 3616851 nodes 76130139 score cp 18 hashfull 1000 nps 21048 tbhits 0 pv c1b2 d7d5 e4d5 e6d5 f1b5 b8c6 d1e2 g8e7 g1f3 c8e6 f3g5 e6f5 b5c6 b7c6 b1c3 f7f6 g5e6 f5e6 e2e6 d8d7 e6d7 e8d7 c3a4 e7g6 e1g1 g6f4 f1e1 f8d6 e1e3 f4e6 g1f1 d6f4 e3d3 f4d6 a1e1 h7h5 d3e3 a8e8 h2h4 e6d4 c2c4 d4c2 e3e8 h8e8 e1e8 d7e8 f1e2 e8f7 d2d3 f7e6 e2d2 c2b4
So per hour it was taking 76 million nodes with an nps of 21 knps. So if i take 24 bytes / per node, in 17 hours it uses up 28 Gb.
That was written prior to edge-node separation. (Memory usage estimation is still roughly true though, "internal node" included data from edge pointing to it, and "leaf node" only contained edge information but not node itself). Prior to edge-node separation, for every visited nodes all its children Nodes were also created even they were not visited (so after visiting X nodes, 20*X of node objects were in memory). We stored priors there.

Now it's the roughly the same, but edges are stored together with node.
Edge size is 4 bytes.
Node size is `80 bytes + number_possible_moves*edge_size`. With 30 possible moves in average given position, this gives 200.
Because of memory allocation overhead, memory fragmentation, blah blah blah, in reality it's more towards 250 bytes per node.
Daniel Shawul
Posts: 4185
Joined: Tue Mar 14, 2006 11:34 am
Location: Ethiopia

Re: buying a new computer

Post by Daniel Shawul »

Daniel Shawul wrote: Fri Jul 19, 2019 4:00 am Well it turns out lc0 performs well on the rtx 2070-super cards getting nps of

Code: Select all

FP16 28k nps
FP32 10k nps
So this must have been a problem with TensorRT library i am using in scorpio. I used to get comparable nps with FP16 and FP32
as lc0 on a volta chip so I am not sure what is going on here. However scorpio still get 35knps with INT8 so it is ok for now..
I have also tried to ssh into the desktop while the screen is dim since the GPU is dual purpose right now. The effect is minimal to the
nps, but I may want to stick in a cheap gpu, maybe gtx 1650 for 150$ on the second pci slot, to handle the display, and the rtx can do the compute.
So this hardware combo ryzen 9 3900x + RTX 2070 super seems to be pretty good for stockfish+lc0 for those interested.

The full log

Code: Select all

daniel@danidesti-desktop:~/engines/lc0/build/release$ ./lc0 -t 2 --backend=cudnn-fp16 --verbose-move-stats --weights=../../../ID-32742.gz
       _
|   _ | |
|_ |_ |_| v0.19.1 built Jul 18 2019
go movetime 50000
Loading weights file from: ../../../ID-32742.gz
Creating backend [cudnn-fp16]...
info depth 1 seldepth 2 time 15 nodes 4 score cp 63 hashfull 0 nps 266 tbhits 0 pv e2e4 e7e5
info depth 2 seldepth 3 time 18 nodes 8 score cp 55 hashfull 0 nps 444 tbhits 0 pv e2e4 e7e5 g1f3
info depth 3 seldepth 4 time 21 nodes 16 score cp 61 hashfull 0 nps 761 tbhits 0 pv e2e4 c7c6 b1c3 d7d5
info depth 3 seldepth 5 time 24 nodes 23 score cp 60 hashfull 0 nps 958 tbhits 0 pv e2e4 c7c6 c2c4 d7d5 e4d5
info depth 4 seldepth 6 time 27 nodes 31 score cp 60 hashfull 0 nps 1148 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4
info depth 4 seldepth 7 time 35 nodes 41 score cp 60 hashfull 0 nps 1171 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 e4d5
info depth 5 seldepth 8 time 42 nodes 50 score cp 60 hashfull 1 nps 1190 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 9 time 50 nodes 62 score cp 60 hashfull 1 nps 1240 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 10 time 56 nodes 88 score cp 58 hashfull 1 nps 1571 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 11 time 62 nodes 121 score cp 58 hashfull 1 nps 1951 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3
info depth 6 seldepth 12 time 68 nodes 159 score cp 58 hashfull 2 nps 2338 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3 e4d5
info depth 6 seldepth 13 time 73 nodes 208 score cp 57 hashfull 2 nps 2849 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3 d1f3
info depth 6 seldepth 14 time 81 nodes 261 score cp 58 hashfull 2 nps 3222 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3 d1f3 e7e6
info depth 6 seldepth 15 time 87 nodes 343 score cp 56 hashfull 2 nps 3942 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 f1b5 c6d4 f3d4 e5d4 e4e5
info depth 6 seldepth 16 time 93 nodes 405 score cp 56 hashfull 3 nps 4354 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 f1b5 f8b4 e1g1 e8g8 d2d3
info depth 6 seldepth 17 time 99 nodes 478 score cp 56 hashfull 3 nps 4828 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 f1b5 c6d4 f3d4 e5d4 e4e5 d4c3
info depth 6 seldepth 18 time 112 nodes 585 score cp 58 hashfull 3 nps 5223 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7
info depth 7 seldepth 18 time 115 nodes 625 score cp 60 hashfull 4 nps 5434 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 h3h4
info depth 7 seldepth 19 time 118 nodes 641 score cp 59 hashfull 4 nps 5432 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4
info depth 7 seldepth 20 time 124 nodes 697 score cp 60 hashfull 4 nps 5620 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8
info depth 7 seldepth 21 time 130 nodes 790 score cp 63 hashfull 4 nps 6076 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2
info depth 7 seldepth 22 time 136 nodes 840 score cp 63 hashfull 5 nps 6176 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8
info depth 7 seldepth 23 time 144 nodes 936 score cp 63 hashfull 5 nps 6500 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5
info depth 7 seldepth 24 time 149 nodes 997 score cp 63 hashfull 5 nps 6691 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 a2a3 e7f8 b1d2 h7h6
info depth 7 seldepth 25 time 156 nodes 1108 score cp 62 hashfull 5 nps 7102 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 e4e5
info depth 8 seldepth 26 time 164 nodes 1226 score cp 61 hashfull 6 nps 7475 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2
info depth 8 seldepth 27 time 173 nodes 1366 score cp 61 hashfull 6 nps 7895 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2 c6b8 b2b3
info depth 8 seldepth 27 time 247 nodes 2874 score cp 63 hashfull 11 nps 11635 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8
info depth 9 seldepth 28 time 290 nodes 3815 score cp 61 hashfull 14 nps 13155 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5
info depth 9 seldepth 29 time 322 nodes 4571 score cp 61 hashfull 16 nps 14195 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5 a1c1 h7h6
info depth 9 seldepth 30 time 342 nodes 4996 score cp 60 hashfull 18 nps 14608 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5 a1c1 h7h6
info depth 9 seldepth 31 time 390 nodes 6084 score cp 59 hashfull 21 nps 15600 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5 d1e2 c8d7 a1d1
info depth 9 seldepth 31 time 410 nodes 6495 score cp 53 hashfull 23 nps 15841 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5
info depth 10 seldepth 32 time 442 nodes 7249 score cp 54 hashfull 25 nps 16400 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7
info depth 10 seldepth 33 time 493 nodes 8139 score cp 56 hashfull 28 nps 16509 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7
info depth 10 seldepth 34 time 505 nodes 8314 score cp 56 hashfull 29 nps 16463 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7
info depth 10 seldepth 35 time 527 nodes 8758 score cp 57 hashfull 31 nps 16618 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7
info depth 11 seldepth 36 time 724 nodes 13952 score cp 53 hashfull 47 nps 19270 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5
info depth 11 seldepth 37 time 810 nodes 16153 score cp 51 hashfull 53 nps 19941 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 11 seldepth 38 time 936 nodes 19518 score cp 51 hashfull 64 nps 20852 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2
info depth 11 seldepth 39 time 999 nodes 21084 score cp 51 hashfull 68 nps 21105 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2
info depth 12 seldepth 39 time 1169 nodes 25472 score cp 50 hashfull 83 nps 21789 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2
info depth 12 seldepth 40 time 1424 nodes 32173 score cp 52 hashfull 104 nps 22593 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2 h7h6
info depth 12 seldepth 41 time 1483 nodes 33380 score cp 52 hashfull 109 nps 22508 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2 h7h6
info depth 13 seldepth 42 time 1901 nodes 44498 score cp 53 hashfull 145 nps 23407 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6
info depth 13 seldepth 43 time 2789 nodes 68783 score cp 53 hashfull 223 nps 24662 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 43 time 3142 nodes 78005 score cp 54 hashfull 254 nps 24826 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 13 seldepth 43 time 3148 nodes 78134 score cp 54 hashfull 254 nps 24820 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 43 time 3192 nodes 79099 score cp 53 hashfull 257 nps 24780 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 13 seldepth 43 time 3207 nodes 79518 score cp 53 hashfull 259 nps 24795 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 43 time 3235 nodes 80266 score cp 53 hashfull 261 nps 24811 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 13 seldepth 43 time 3245 nodes 80473 score cp 53 hashfull 262 nps 24799 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 13 seldepth 44 time 3680 nodes 94512 score cp 52 hashfull 301 nps 25682 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 44 time 4005 nodes 103067 score cp 52 hashfull 329 nps 25734 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 45 time 4316 nodes 111440 score cp 52 hashfull 356 nps 25820 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 46 time 4493 nodes 115830 score cp 52 hashfull 371 nps 25780 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6
info depth 14 seldepth 47 time 6373 nodes 170980 score cp 52 hashfull 539 nps 26828 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 g8h7 h3h4 d5b4
info depth 14 seldepth 48 time 9203 nodes 260573 score cp 51 hashfull 792 nps 28313 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4
info depth 15 seldepth 48 time 9307 nodes 263254 score cp 51 hashfull 800 nps 28285 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4
info depth 15 seldepth 49 time 10582 nodes 301814 score cp 51 hashfull 912 nps 28521 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4
info depth 15 seldepth 49 time 15596 nodes 445421 score cp 50 hashfull 1000 nps 28559 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 15 seldepth 49 time 20597 nodes 581291 score cp 49 hashfull 1000 nps 28222 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 16 seldepth 49 time 20661 nodes 582663 score cp 49 hashfull 1000 nps 28201 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 15 seldepth 49 time 20675 nodes 582978 score cp 49 hashfull 1000 nps 28197 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 16 seldepth 49 time 20692 nodes 583326 score cp 49 hashfull 1000 nps 28190 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 16 seldepth 49 time 25702 nodes 718377 score cp 48 hashfull 1000 nps 27950 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info depth 16 seldepth 49 time 25788 nodes 720361 score cp 48 hashfull 1000 nps 27933 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4 a1a4 b7b2
info string g2g4  (378 ) N:     225 (+ 0) (P:  1.38%) (Q: -0.38179) (U: 0.53325) (Q+U:  0.15146) (V:  -.----) 
info string f2f3  (346 ) N:     444 (+ 0) (P:  1.75%) (Q: -0.20453) (U: 0.34239) (Q+U:  0.13785) (V:  -.----) 
info string g1h3  (161 ) N:     579 (+ 0) (P:  1.95%) (Q: -0.15908) (U: 0.29315) (Q+U:  0.13408) (V:  -.----) 
info string b1a3  (34  ) N:     695 (+ 0) (P:  2.00%) (Q: -0.11876) (U: 0.24964) (Q+U:  0.13088) (V:  -.----) 
info string h2h4  (403 ) N:     711 (+ 0) (P:  2.10%) (Q: -0.12508) (U: 0.25644) (Q+U:  0.13136) (V:  -.----) 
info string f2f4  (351 ) N:     989 (+ 0) (P:  2.28%) (Q: -0.07299) (U: 0.20013) (Q+U:  0.12713) (V:  -.----) 
info string b2b4  (234 ) N:    1158 (+ 0) (P:  2.60%) (Q: -0.06846) (U: 0.19564) (Q+U:  0.12718) (V:  -.----) 
info string a2a4  (207 ) N:    1239 (+ 0) (P:  2.47%) (Q: -0.04817) (U: 0.17338) (Q+U:  0.12521) (V:  -.----) 
info string a2a3  (204 ) N:    1794 (+ 0) (P:  2.63%) (Q: -0.00585) (U: 0.12766) (Q+U:  0.12180) (V:  -.----) 
info string d2d3  (288 ) N:    1916 (+ 0) (P:  2.68%) (Q: -0.00047) (U: 0.12168) (Q+U:  0.12121) (V:  -.----) 
info string b2b3  (230 ) N:    1972 (+ 0) (P:  2.87%) (Q: -0.00488) (U: 0.12647) (Q+U:  0.12159) (V:  -.----) 
info string h2h3  (400 ) N:    2614 (+ 0) (P:  3.01%) (Q:  0.01929) (U: 0.10030) (Q+U:  0.11959) (V:  -.----) 
info string c2c3  (259 ) N:    2899 (+ 0) (P:  2.98%) (Q:  0.02917) (U: 0.08948) (Q+U:  0.11865) (V:  -.----) 
info string b1c3  (36  ) N:    3086 (+ 0) (P:  2.69%) (Q:  0.04187) (U: 0.07582) (Q+U:  0.11769) (V:  -.----) 
info string e2e3  (317 ) N:    5121 (+ 0) (P:  3.47%) (Q:  0.05743) (U: 0.05903) (Q+U:  0.11645) (V:  -.----) 
info string g2g3  (374 ) N:    7912 (+ 0) (P:  4.50%) (Q:  0.06614) (U: 0.04950) (Q+U:  0.11565) (V:  -.----) 
info string c2c4  (264 ) N:   25819 (+ 0) (P:  7.94%) (Q:  0.08684) (U: 0.02677) (Q+U:  0.11361) (V:  -.----) 
info string g1f3  (159 ) N:   75596 (+ 0) (P: 13.63%) (Q:  0.09638) (U: 0.01569) (Q+U:  0.11208) (V:  -.----) 
info string d2d4  (293 ) N:  116631 (+ 0) (P: 14.02%) (Q:  0.10070) (U: 0.01046) (Q+U:  0.11116) (V:  -.----) 
info string e2e4  (322 ) N:  468960 (+284) (P: 23.05%) (Q:  0.10686) (U: 0.00428) (Q+U:  0.11114) (V:  -.----) 
bestmove e2e4 ponder c7c5
quit
daniel@danidesti-desktop:~/engines/lc0/build/release$ ./lc0 -t 2 --backend=cudnn --verbose-move-stats --weights=../../../ID-32742.gz
       _
|   _ | |
|_ |_ |_| v0.19.1 built Jul 18 2019
go movetime 50000
Loading weights file from: ../../../ID-32742.gz
Creating backend [cudnn]...
info depth 1 seldepth 2 time 22 nodes 4 score cp 63 hashfull 0 nps 181 tbhits 0 pv e2e4 e7e5
info depth 2 seldepth 3 time 30 nodes 8 score cp 55 hashfull 0 nps 266 tbhits 0 pv e2e4 e7e5 g1f3
info depth 3 seldepth 4 time 39 nodes 16 score cp 61 hashfull 0 nps 410 tbhits 0 pv e2e4 c7c6 b1c3 d7d5
info depth 3 seldepth 5 time 48 nodes 24 score cp 60 hashfull 0 nps 500 tbhits 0 pv e2e4 c7c6 c2c4 d7d5 e4d5
info depth 4 seldepth 6 time 66 nodes 33 score cp 60 hashfull 0 nps 500 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4
info depth 4 seldepth 7 time 82 nodes 43 score cp 60 hashfull 0 nps 524 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 4 seldepth 8 time 100 nodes 51 score cp 60 hashfull 1 nps 510 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 8 time 109 nodes 55 score cp 60 hashfull 1 nps 504 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 9 time 117 nodes 64 score cp 60 hashfull 1 nps 547 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3
info depth 5 seldepth 10 time 134 nodes 97 score cp 59 hashfull 1 nps 723 tbhits 0 pv e2e4 c7c6 c2c4 d7d5 e4d5 g8f6 b1c3 c6d5
info depth 5 seldepth 11 time 150 nodes 131 score cp 58 hashfull 1 nps 873 tbhits 0 pv e2e4 c7c6 c2c4 d7d5 e4d5 g8f6 b1c3 c6d5 d2d4
info depth 6 seldepth 12 time 168 nodes 170 score cp 58 hashfull 1 nps 1011 tbhits 0 pv e2e4 c7c6 c2c4 d7d5 e4d5 g8f6 b1c3 c6d5 d2d4 b8c6
info depth 6 seldepth 13 time 185 nodes 225 score cp 57 hashfull 2 nps 1216 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3 d1f3
info depth 6 seldepth 14 time 202 nodes 285 score cp 58 hashfull 2 nps 1410 tbhits 0 pv e2e4 c7c6 b1c3 d7d5 g1f3 c8g4 h2h3 g4f3 d1f3 e7e6
info depth 6 seldepth 15 time 221 nodes 372 score cp 56 hashfull 2 nps 1683 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 d2d4 f8b4 f3e5 f6e4 d1f3
info depth 6 seldepth 16 time 240 nodes 433 score cp 56 hashfull 2 nps 1804 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 f1b5 f8b4 e1g1 e8g8 d2d3
info depth 6 seldepth 17 time 260 nodes 514 score cp 56 hashfull 3 nps 1976 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 b1c3 g8f6 f1b5 c6d4 f3d4 e5d4 e4e5 d4c3
info depth 7 seldepth 17 time 280 nodes 608 score cp 58 hashfull 3 nps 2171 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 b5b4
info depth 7 seldepth 18 time 287 nodes 615 score cp 58 hashfull 3 nps 2142 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7
info depth 6 seldepth 18 time 295 nodes 649 score cp 60 hashfull 3 nps 2200 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 h3h4
info depth 7 seldepth 19 time 301 nodes 661 score cp 60 hashfull 4 nps 2196 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4
info depth 7 seldepth 20 time 320 nodes 722 score cp 60 hashfull 4 nps 2256 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8
info depth 7 seldepth 21 time 340 nodes 823 score cp 62 hashfull 4 nps 2420 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2
info depth 7 seldepth 22 time 353 nodes 882 score cp 63 hashfull 4 nps 2498 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8
info depth 7 seldepth 23 time 370 nodes 955 score cp 63 hashfull 5 nps 2581 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5
info depth 7 seldepth 24 time 384 nodes 1026 score cp 62 hashfull 5 nps 2671 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 a2a3 e7f8 b1d2 h7h6
info depth 7 seldepth 25 time 399 nodes 1115 score cp 62 hashfull 5 nps 2794 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 e4e5
info depth 7 seldepth 26 time 423 nodes 1254 score cp 61 hashfull 6 nps 2964 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2
info depth 8 seldepth 26 time 436 nodes 1334 score cp 60 hashfull 6 nps 3059 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2 c6b8
info depth 8 seldepth 27 time 449 nodes 1402 score cp 60 hashfull 6 nps 3122 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2 c6b8
info depth 8 seldepth 28 time 478 nodes 1568 score cp 59 hashfull 7 nps 3280 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a3 h7h6 b3c2 c6b8
info depth 8 seldepth 28 time 682 nodes 3230 score cp 61 hashfull 12 nps 4736 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8
info depth 9 seldepth 28 time 760 nodes 3906 score cp 61 hashfull 14 nps 5139 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5
info depth 9 seldepth 29 time 878 nodes 4785 score cp 60 hashfull 17 nps 5449 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5
info depth 9 seldepth 30 time 944 nodes 5320 score cp 59 hashfull 19 nps 5635 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5
info depth 9 seldepth 31 time 1146 nodes 6992 score cp 59 hashfull 24 nps 6101 tbhits 0 pv d2d4 g8f6 c2c4 e7e6 g1f3 d7d5 b1c3 d5c4 e2e3 a7a6 a2a4 f8e7 f1c4 c7c5 e1g1 c5d4 e3d4 b8c6 f1e1 e8g8 h2h3 d8b6 c4b3 f8d8 c1e3 b6a5 a1c1 h7h6
info depth 9 seldepth 32 time 1186 nodes 7329 score cp 53 hashfull 25 nps 6179 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5
info depth 9 seldepth 33 time 1249 nodes 7730 score cp 53 hashfull 26 nps 6188 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5
info depth 10 seldepth 33 time 1271 nodes 7913 score cp 54 hashfull 27 nps 6225 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5
info depth 10 seldepth 34 time 1342 nodes 8402 score cp 55 hashfull 29 nps 6260 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b5b4
info depth 10 seldepth 35 time 1438 nodes 9037 score cp 57 hashfull 31 nps 6284 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7
info depth 10 seldepth 36 time 1932 nodes 13727 score cp 52 hashfull 45 nps 7105 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 g5c1
info depth 10 seldepth 37 time 2001 nodes 14258 score cp 52 hashfull 47 nps 7125 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 11 seldepth 37 time 2027 nodes 14596 score cp 51 hashfull 48 nps 7200 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 11 seldepth 38 time 2535 nodes 19489 score cp 50 hashfull 63 nps 7687 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 11 seldepth 39 time 2633 nodes 20358 score cp 50 hashfull 66 nps 7731 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 12 seldepth 39 time 3382 nodes 27532 score cp 50 hashfull 87 nps 8140 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b7c6 c1g5 b8d7 f3h2
info depth 12 seldepth 40 time 3940 nodes 32731 score cp 51 hashfull 104 nps 8307 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2
info depth 12 seldepth 41 time 4200 nodes 34930 score cp 51 hashfull 111 nps 8316 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 d4d5 c6b8 d2f1 c7c6 d5c6 b8c6 c1g5 c6a5 b3c2
info depth 12 seldepth 42 time 4861 nodes 41857 score cp 51 hashfull 132 nps 8610 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7
info depth 13 seldepth 42 time 5392 nodes 46267 score cp 52 hashfull 147 nps 8580 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6
info depth 13 seldepth 43 time 7257 nodes 63316 score cp 53 hashfull 204 nps 8724 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1
info depth 14 seldepth 43 time 8961 nodes 78786 score cp 54 hashfull 257 nps 8792 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 a4b5 a6b5
info depth 14 seldepth 44 time 9681 nodes 85954 score cp 54 hashfull 279 nps 8878 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6 d5b4
info depth 14 seldepth 45 time 11438 nodes 104583 score cp 53 hashfull 334 nps 9143 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6 d5b4
info depth 14 seldepth 45 time 16464 nodes 156742 score cp 52 hashfull 492 nps 9520 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6 d5b4 c2b1
info depth 14 seldepth 46 time 17926 nodes 171702 score cp 52 hashfull 539 nps 9578 tbhits 0 pv e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c8b7 d2d4 f8e8 b1d2 e7f8 a2a4 e5d4 c3d4 d8d7 a4b5 a6b5 a1a8 b7a8 f3g5 c6d8 e4e5 f6d5 d1c2 g7g6 d2e4 f8g7 e1d1 h7h6 g5f3 d8e6 c1h6 d5b4 c2b1
info depth 14 seldepth 46 time 22964 nodes 228407 score cp 51 hashfull 696 nps 9946 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4
info depth 14 seldepth 47 time 23736 nodes 236879 score cp 51 hashfull 720 nps 9979 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4
info depth 14 seldepth 47 time 25185 nodes 253065 score cp 51 hashfull 765 nps 10048 tbhits 0 pv e2e4 c7c5 g1f3 e7e6 a2a3 b8c6 d2d4 c5d4 f3d4 g8f6 b1c3 d8c7 d4b5 c7b8 f2f4 d7d6 c1e3 a7a6 b5d4 f8e7 f1d3 e8g8 e1g1 c6d4 e3d4 e6e5 d4b6 c8d7 g1h1 e7d8 b6e3 b7b5 d1f3 b8b7 f4e5 d6e5 a3a4 b5a4 c3a4 d7a4
info string g2g4  (378 ) N:     100 (+ 0) (P:  1.39%) (Q: -0.40060) (U: 0.57051) (Q+U:  0.16991) (V: -0.4643) 
info string f2f3  (346 ) N:     196 (+ 0) (P:  1.75%) (Q: -0.21720) (U: 0.36934) (Q+U:  0.15214) (V: -0.2128) 
info string g1h3  (161 ) N:     269 (+ 0) (P:  1.96%) (Q: -0.15453) (U: 0.30095) (Q+U:  0.14642) (V: -0.1817) 
info string h2h4  (403 ) N:     333 (+ 0) (P:  2.09%) (Q: -0.11766) (U: 0.26056) (Q+U:  0.14290) (V: -0.1436) 
info string b1a3  (34  ) N:     337 (+ 0) (P:  2.02%) (Q: -0.10664) (U: 0.24866) (Q+U:  0.14202) (V: -0.0977) 
info string f2f4  (351 ) N:     449 (+ 0) (P:  2.28%) (Q: -0.07175) (U: 0.21023) (Q+U:  0.13849) (V: -0.0892) 
info string a2a4  (207 ) N:     570 (+ 0) (P:  2.49%) (Q: -0.04548) (U: 0.18151) (Q+U:  0.13602) (V: -0.0194) 
info string b2b4  (234 ) N:     584 (+ 0) (P:  2.60%) (Q: -0.04868) (U: 0.18464) (Q+U:  0.13596) (V: -0.0746) 
info string a2a3  (204 ) N:     845 (+ 0) (P:  2.66%) (Q:  0.00089) (U: 0.13068) (Q+U:  0.13157) (V:  0.0110) 
info string d2d3  (288 ) N:     859 (+ 0) (P:  2.69%) (Q:  0.00144) (U: 0.13010) (Q+U:  0.13154) (V:  0.0156) 
info string b2b3  (230 ) N:     880 (+ 0) (P:  2.87%) (Q: -0.00328) (U: 0.13542) (Q+U:  0.13214) (V:  0.0124) 
info string b1c3  (36  ) N:    1093 (+ 0) (P:  2.71%) (Q:  0.02592) (U: 0.10294) (Q+U:  0.12885) (V:  0.0458) 
info string h2h3  (400 ) N:    1112 (+ 0) (P:  3.02%) (Q:  0.01740) (U: 0.11257) (Q+U:  0.12997) (V:  0.0164) 
info string c2c3  (259 ) N:    1274 (+ 0) (P:  2.99%) (Q:  0.03115) (U: 0.09757) (Q+U:  0.12872) (V:  0.0471) 
info string e2e3  (317 ) N:    2130 (+ 0) (P:  3.49%) (Q:  0.05792) (U: 0.06808) (Q+U:  0.12600) (V:  0.0668) 
info string g2g3  (374 ) N:    3694 (+ 0) (P:  4.53%) (Q:  0.07320) (U: 0.05093) (Q+U:  0.12413) (V:  0.0885) 
info string c2c4  (264 ) N:   11288 (+ 0) (P:  8.02%) (Q:  0.09253) (U: 0.02953) (Q+U:  0.12207) (V:  0.1122) 
info string g1f3  (159 ) N:   26864 (+ 0) (P: 13.70%) (Q:  0.09873) (U: 0.02119) (Q+U:  0.11991) (V:  0.1171) 
info string d2d4  (293 ) N:   38442 (+ 0) (P: 13.90%) (Q:  0.10484) (U: 0.01502) (Q+U:  0.11987) (V:  0.1258) 
info string e2e4  (322 ) N:  161745 (+273) (P: 22.84%) (Q:  0.11386) (U: 0.00586) (Q+U:  0.11972) (V:  0.1398) 
bestmove e2e4 ponder c7c5
quit
Ok I am able to get pretty good number with scorpio now if i don't use the maddex net.

Code: Select all

FP32   4753 nps
FP16   26994 nps
INT8   46208 nps
So it is now getting 46 knps with INT8, and 26knps with FP16 using my net-20x256.uff network.
So I suspect the problem is with TensorRT not being able to optimize leelas style nets as well as my own format.

One thing i noticed is that the RTX 2070-super when using FP16 has 2x more flops with FP16 accumulator than with FP32 accumulator.

Daniel
Zenmastur
Posts: 919
Joined: Sat May 31, 2014 8:28 am

Re: buying a new computer

Post by Zenmastur »

Daniel Shawul wrote: Thu Jul 18, 2019 7:46 am
Success at last!!


...

I did a quick benchmark on the CPU (ryzen 9 3900x) and GPU (RTX 2070 super). Stockfish seems to scale linearly across the 12 cores with its lazy smp implementation. So i got about 1.8 mnps on 1 core using latest source compiled with gcc 7.4, and get 21 mnps using all 12 cores. It goes up to 27 mnps if i use hyperthreading (24 threads). Similar scaling for scorpio as well. No overclocking for this test, so just the base clock of 3.8ghz.
The more I look at the numbers you posted for SF the more I think something isn't right. I'm not sure what, but the NPS seems unusually low and it really started to bug me. My first thought was JEDEC memory timings for slower memory are being used. But even if slow 2133 memory timings are used it would only slow the machine about 15% as compared to DDR4 3200 CL16 ram. Which makes me curios about which version of SF you used. Was it an Abrok Haswell compile of recent vintage? If so, which one? Or, was it a POPCNT compile?

Maybe I'm just paranoid but these numbers don't seem right and if it were my machine I would be scouring the earth to determine what the problem was especial since this is a new build. So you may want to look around a bit to make sure all the things that you normally can't see are actually set the way you believe they should be.

I'm going to a High Power Rocket launch tomorrow weather permitting so I won't be around until later in the day. But I am interested in determining if my intuition is correct.

Regards,

Zenmastur
Only 2 defining forces have ever offered to die for you.....Jesus Christ and the American Soldier. One died for your soul, the other for your freedom.
Daniel Shawul
Posts: 4185
Joined: Tue Mar 14, 2006 11:34 am
Location: Ethiopia

Re: buying a new computer

Post by Daniel Shawul »

Zenmastur wrote: Sat Jul 20, 2019 8:24 am
Daniel Shawul wrote: Thu Jul 18, 2019 7:46 am
Success at last!!


...

I did a quick benchmark on the CPU (ryzen 9 3900x) and GPU (RTX 2070 super). Stockfish seems to scale linearly across the 12 cores with its lazy smp implementation. So i got about 1.8 mnps on 1 core using latest source compiled with gcc 7.4, and get 21 mnps using all 12 cores. It goes up to 27 mnps if i use hyperthreading (24 threads). Similar scaling for scorpio as well. No overclocking for this test, so just the base clock of 3.8ghz.
The more I look at the numbers you posted for SF the more I think something isn't right. I'm not sure what, but the NPS seems unusually low and it really started to bug me. My first thought was JEDEC memory timings for slower memory are being used. But even if slow 2133 memory timings are used it would only slow the machine about 15% as compared to DDR4 3200 CL16 ram. Which makes me curios about which version of SF you used. Was it an Abrok Haswell compile of recent vintage? If so, which one? Or, was it a POPCNT compile?
I built it myself with gcc, but i did not do a profile build before. Now that I did that the single core nps had increased to 2.2 mnps from 1.8 mnps.
And with 24 threads, I now get 31 millons nps. I have also tried an abrok compile for modern-linux and it gives the same nps.

However, i did the nps measurement from the start position only which is probably the problem. If I do the "bench" command i get an average nps of 2.75 million nps. So using all 24 threads and bench it goes upto 38 million nps

Code: Select all

./stockfish_19071415_x64_modern bench 512 24 28 default depth

Total time (ms) : 165299
Nodes searched  : 6257478895
Nodes/second    : 37855515


I'm going to a High Power Rocket launch tomorrow weather permitting so I won't be around until later in the day. But I am interested in determining if my intuition is correct.
Have fun!

Daniel
Zenmastur
Posts: 919
Joined: Sat May 31, 2014 8:28 am

Re: buying a new computer

Post by Zenmastur »

Daniel Shawul wrote: Sat Jul 20, 2019 3:16 pm
Zenmastur wrote: Sat Jul 20, 2019 8:24 am
Daniel Shawul wrote: Thu Jul 18, 2019 7:46 am
Success at last!!


...

I did a quick benchmark on the CPU (ryzen 9 3900x) and GPU (RTX 2070 super). Stockfish seems to scale linearly across the 12 cores with its lazy smp implementation. So i got about 1.8 mnps on 1 core using latest source compiled with gcc 7.4, and get 21 mnps using all 12 cores. It goes up to 27 mnps if i use hyperthreading (24 threads). Similar scaling for scorpio as well. No overclocking for this test, so just the base clock of 3.8ghz.
The more I look at the numbers you posted for SF the more I think something isn't right. I'm not sure what, but the NPS seems unusually low and it really started to bug me. My first thought was JEDEC memory timings for slower memory are being used. But even if slow 2133 memory timings are used it would only slow the machine about 15% as compared to DDR4 3200 CL16 ram. Which makes me curios about which version of SF you used. Was it an Abrok Haswell compile of recent vintage? If so, which one? Or, was it a POPCNT compile?
I built it myself with gcc, but i did not do a profile build before. Now that I did that the single core nps had increased to 2.2 mnps from 1.8 mnps.
And with 24 threads, I now get 31 millons nps. I have also tried an abrok compile for modern-linux and it gives the same nps.

However, i did the nps measurement from the start position only which is probably the problem. If I do the "bench" command i get an average nps of 2.75 million nps. So using all 24 threads and bench it goes upto 38 million nps

Code: Select all

./stockfish_19071415_x64_modern bench 512 24 28 default depth

Total time (ms) : 165299
Nodes searched  : 6257478895
Nodes/second    : 37855515


I'm going to a High Power Rocket launch tomorrow weather permitting so I won't be around until later in the day. But I am interested in determining if my intuition is correct.
Have fun!

Daniel
That's a 40% improvement over the original numbers you gave. I was expecting numbers in the range of 35M nps to 43.5M nps. So, these are definitely in the “GOOD” range. It makes me want to go out and buy one just so I can tweak on it! :D :D :D

I'm trying mightily to resist the urge. :cry: :cry: :cry:

How much money do you think you saved by building your own computer?

And are you happy with your component decision?

Regards,

Zenmastur
Only 2 defining forces have ever offered to die for you.....Jesus Christ and the American Soldier. One died for your soul, the other for your freedom.
Daniel Shawul
Posts: 4185
Joined: Tue Mar 14, 2006 11:34 am
Location: Ethiopia

Re: buying a new computer

Post by Daniel Shawul »

I am quite happy with the system! The cpu is very powerfull ( maybe half the power of the tcec machine) and so is the RTX 2070-super.
I think i may have saved up 500$ which we will know for sure once pre-built pcs with ryzen 9 and rtx 2070-super show up.
This came at the right time that I needed to buy either a desktop or laptop -- glad i went for the desktop since it is way more powerful in
every aspect. Most of all, i enjoyed the experience a lot.

I made a miscalulation on RAM so i will upgrade it to 32 GB or more at some point, but it works for now.

Thanks
Daniel
corres
Posts: 3657
Joined: Wed Nov 18, 2015 11:41 am
Location: hungary

Re: buying a new computer

Post by corres »

crem wrote: Fri Jul 19, 2019 9:36 am
Daniel Shawul wrote: Fri Jul 19, 2019 4:49 am
I don't think lc0 consumes that much memory and from my calculations it could go upto 17 hours analysis with just 12 gb ram.
Thats one of the reason i decided to stick with 16 gb ram.

Daniel
1 node takes 250 bytes, or in other words 1GB is needed for 4M nodes.
(and also one NN cache entry takes 350 bytes. It doesn't grow with time, but it's a thing to consider when setting huge cache sizes).

So at 33333 nps it takes 1GB per 2 minutes, per hour it takes 30GB, and after 17 hours it needs 512GB of RAM.

If 12GB is enough for 17 hours (and it overflows into swap partition), it seems that average nps is around 780.
The huge memory usage of Lc0 is a returning theme on the sites of LC0.
But this is the first established answer for it.
Why, Dear "crem"?
Is it so great secret?
I can not believe the developers of Leela can not make modifiable the usage of RAM with an UCI parameter. Maybe it would decrease the Leela`s Elo in some measure but this effect would be decreased with a timed and structured Hash table as it is in AB engines.
It is not an elegant solving for this issue to let the Leela freezing down when it consumes the full of system RAM.
User avatar
Nordlandia
Posts: 2821
Joined: Fri Sep 25, 2015 9:38 pm
Location: Sortland, Norway

Re: buying a new computer

Post by Nordlandia »

Here is the performance difference of RTX 2080 Super card being released on 23rd (scheduled)

2080 Super is 7.5% or ~ 8% better overall.

Image
brianr
Posts: 536
Joined: Thu Mar 09, 2006 3:01 pm

Re: buying a new computer

Post by brianr »

corres wrote: Sun Jul 21, 2019 5:43 pm The huge memory usage of Lc0 is a returning theme on the sites of LC0.
But this is the first established answer for it.
Why, Dear "crem"?
Is it so great secret?
I can not believe the developers of Leela can not make modifiable the usage of RAM with an UCI parameter. Maybe it would decrease the Leela`s Elo in some measure but this effect would be decreased with a timed and structured Hash table as it is in AB engines.
It is not an elegant solving for this issue to let the Leela freezing down when it consumes the full of system RAM.
There already is a parameter, although there is some question about how accurate it is.

--ramlimit-mb=0..100000000
Maximum memory usage for the engine, in megabytes. The estimation is very rough,
and can be off by a lot. For example, multiple visits to a terminal node counted
several times, and the estimation assumes that all positions have 30 possible
moves. When set to 0, no RAM limit is enforced.
[UCI: RamLimitMb DEFAULT: 0 MIN: 0 MAX: 100000000]