Help with Scorpio 3.0.8

[Daniel Shawul]

Not a problem, I changed the default to 30, so now it should report from 30 to 45 max most of the time.
In the endgame ( <10 pieces ) it will offset this by the depth the AB search managed to reach to avoid duplicate depths.

[Dann Corbit]

Not a problem, I changed the default to 30, so now it should report from 30 to 45 max most of the time.
In the endgame ( <10 pieces ) it will offset this by the depth the AB search managed to reach to avoid duplicate depths.

Thank you kindly sir.

Now, to get started, where is the software to download (I want to be sure to get the right version) and where shall I find instructions to run the calibrations?

[Daniel Shawul]

Information is on my website http://scorpiozero.ddns.net

There is a link to the client github page on the left. Git clone (or download) that and run client.bat for windows or client.sh for linux.
The only requirement is a java runtime which most people have anyway.
It will install Scorpio for selfplay game generation and then connect to the server. Once you create a username and password,
it will keep how many games you generated for specific training runs.

That is it

[Terje]

Why does scorpio not report depths in the same way leela or allie does?

[Modern Times]

At first I thought that it was a mistake it was launching only 10 threads, but it isn't.
The GTX 1050 has only 5 SMs, so it allocates 2x the number of threads.

I am curious though how much pps it will get with mt set to 20 or 40.
Could you set mt to these values in scoprio.ini and run scorpio.bat go quit again, and post the results here?
My estimate of using 2x number of SMs of GPU could be wrong on low end GPUs.

Thanks.
Daniel

This is mt 20:

Code: Select all

C:\Engines\ScorpioNN\Scorpio\bin\Windows>scorpio.bat go quit
feature done=0
Number of cores 8 of 8
treeht 83886080 X 320 = 25600.0 MB
processors [20]
ht 67108864 X 16 = 1024.0 MB
eht 262144 X 8 X 20 = 80.0 MB
pht 32768 X 24 X 20 = 15.0 MB
EgbbProbe 4.3 by Daniel Shawul
egbb_cache 4084 X 8216 = 32.0 MB
180 egbbs loaded !
nn_cache 131072 X 1568 = 196.0 MB
Loading neural network : C:\Engines\ScorpioNN\Scorpio\nets-scorpio/ens-net-20x256.uff
With "INT8" precision
Loading graph on /gpu:0
0. main_input 2048 = 32 8 810
1. policy/Reshape 1024 = 1024 1 1
2. value/BiasAdd 3 = 3 1 1
Neural network loaded !
nn_cache 131072 X 1568 = 196.0 MB
Loading neural network : C:\Engines\ScorpioNN\Scorpio\nets-maddex/ME.uff
With "INT8" precision
Loading graph on /gpu:0
0. main_input 7168 = 112 8 80
1. value_head 3 = 3 1 1
2. policy_head 1858 = 1858 1 1
Neural network loaded !
loading_time = 1400s
# rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
# [st = 12502ms, mt = 29250ms , hply = 0 , moves_left 10]
63 32 87 2095 63 2394 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 Ng8-f6 Ke1-g1 Nf6xe4 Rf1-e1 Ne4-d6 Nf3xe5 Bf8-e7 Bb5-f1 Nc6xe5 Re1xe5 Ke8-g8 d2-d4 Be7-f6
64 32 175 4342 64 2481 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 Ng8-f6 Ke1-g1 Nf6xe4 Rf1-e1 Ne4-d6 Nf3xe5 Bf8-e7 Bb5-f1 Nc6xe5 Re1xe5 Ke8-g8 d2-d4 Be7-f6 Re5-e1 Rf8-e8 c2-c3 Re8xe1 Qd1xe1 Nd6-e8 Bc1-f4 d7-d5 Nb1-d2 c7-c6 Bf1-d3 g7-g6 Nd2-f3
65 31 264 6663 65 2516 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 Ng8-f6 Ke1-g1 Nf6xe4 Rf1-e1 Ne4-d6 Nf3xe5 Bf8-e7 Bb5-f1 Nc6xe5 Re1xe5 Ke8-g8 d2-d4 Be7-f6 Re5-e1 Rf8-e8 c2-c3 Re8xe1 Qd1xe1 Nd6-e8 Bc1-f4 d7-d5 Nb1-d2 c7-c6 Bf1-d3 g7-g6 Nd2-f3
66 31 352 9056 66 2570 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Bf8-e7 Rf1-e1 b7-b5 Ba4-b3 d7-d6 c2-c3 Ke8-g8 h2-h3 Bc8-b7 d2-d4 Rf8-e8 Nb1-d2 Be7-f8 a2-a4 h7-h6 Bb3-c2 Ra8-b8
67 30 439 11525 67 2620 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Bf8-e7 Rf1-e1 b7-b5 Ba4-b3 d7-d6 c2-c3 Ke8-g8 h2-h3 Bc8-b7 d2-d4 Rf8-e8 Nb1-d2 Be7-f8 a2-a4 h7-h6 Bb3-c2 Ra8-b8
68 30 528 14276 68 2699 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Bf8-e7 Rf1-e1 b7-b5 Ba4-b3 d7-d6 c2-c3 Ke8-g8 h2-h3 Bc8-b7 d2-d4 Rf8-e8 Nb1-d2 Be7-f8 a2-a4 h7-h6 Bb3-c2 Ra8-b8
69 30 616 16729 69 2714 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Bf8-e7 Rf1-e1 b7-b5 Ba4-b3 d7-d6 c2-c3 Ke8-g8 h2-h3 Bc8-b7 d2-d4 Rf8-e8 Nb1-d2 Be7-f8 a2-a4 h7-h6 Bb3-c2 Ra8-b8 Bc2-b1
70 30 703 19160 70 2722 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Nf6xe4 d2-d4 b7-b5 Ba4-b3 d7-d5 d4xe5 Bc8-e6 c2-c3 Bf8-e7 Nb1-d2 Ne4-c5 Bb3-c2 d5-d4 Nd2-b3 d4xc3 Nb3xc5 Be7xc5 Bc2-e4 Qd8-d7 b2xc3 Ra8-d8
71 31 791 21572 71 2726 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Nf6xe4 d2-d4 b7-b5 Ba4-b3 d7-d5 d4xe5 Bc8-e6 c2-c3 Bf8-e7 Nb1-d2 Ne4-c5 Bb3-c2 d5-d4 Nd2-b3 d4xc3 Nb3xc5 Be7xc5 Bc2-e4 Qd8-d7 b2xc3 Ra8-d8 Qd1xd7 Be6xd7
72 31 878 24040 72 2735 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Nf6xe4 d2-d4 b7-b5 Ba4-b3 d7-d5 d4xe5 Bc8-e6 c2-c3 Bf8-e7 Nb1-d2 Ne4-c5 Bb3-c2 d5-d4 Nd2-b3 d4xc3 Nb3xc5 Be7xc5 Bc2-e4 Qd8-d7 b2xc3 Ra8-d8 Qd1xd7 Be6xd7
73 31 946 25919 73 2738 0 e2-e4 e7-e5 Ng1-f3 Nb8-c6 Bf1-b5 a7-a6 Bb5-a4 Ng8-f6 Ke1-g1 Nf6xe4 d2-d4 b7-b5 Ba4-b3 d7-d5 d4xe5 Bc8-e6 c2-c3 Bf8-e7 Nb1-d2 Ne4-c5 Bb3-c2 d5-d4 Nd2-b3 d4xc3 Nb3xc5 Be7xc5 Bc2-e4 Qd8-d7 b2xc3 Ra8-d8 Qd1xd7 Be6xd7
# nodes 267354 <0% qnodes> time 9550ms nps 27995 eps 0 nneps 2733 tbhits 0
# Tree: nodes 33428 depth 33 pps 2732 visits 25927 qcalls 0 scalls 0
move e2e4
Bye Bye

[Dann Corbit]

Why does scorpio not report depths in the same way leela or allie does?

I guess because depths are relatively meaningless with MCTS programs. It's already iffy with alpha-beta (Does a single reply extension count? (for instance))

With alpha-beta, you at least have a depth measure that is passed into the search algorithm. The MCTS searches kind of have an approximation of that, but it is not the same as an alpha-beta depth.

A 20 ply MCTS search is not commensurate with a 20 ply alpha-beta search.

So I guess he was being funny and making a joke about depth in the PGN standard for EPD and FEN.

When he answers, I'll see if my guess is right.

[Dann Corbit]

Information is on my website http://scorpiozero.ddns.net

There is a link to the client github page on the left. Git clone (or download) that and run client.bat for windows or client.sh for linux.
The only requirement is a java runtime which most people have anyway.
It will install Scorpio for selfplay game generation and then connect to the server. Once you create a username and password,
it will keep how many games you generated for specific training runs.

That is it

Suggestion:
Have the client download the EGBB files if they are missing.
I don't have them on the machine I am running the client on, so I got a "0 egbbs loaded !" message
Since we already downloaded a pile of stuff, what is one more pile?

I need to load Java on it too
Oops, that's my bad, you told me about that and I forgot I did not do that yet

[Dann Corbit]

I am not sure if it is working. It looks like it is not.
After I gave my username and password (pw had a single quote in it, is that OK)?
it just sits there, apparently doing nothing.
My 64 threads are idle.
One GPU is loafing at 1% and the other one is doing nothing.
Do I need to take another step?

[Dann Corbit]

It was the single quote.
I made a new user/pass and it started doing something.

[Dann Corbit]

It is only using one of the two GPUs. Is that intentional?
Seems to be running with very low resources.
It is using 2% of CPU bandwidth and 1% of one GPU.
I think you should go ahead and open the throttle!
Give it both GPUs at 100%.
Should I run multiple sessions of the client?