EGTB Crowd

[jshriver]

Just popping in to say hi and hope everyone is well. Seems the scene has kinda died down a good bit from the original days when this started.
Hope work is still being done for 6 and beyond, but still here to read and discuss.

Take it easy
-Josh

[koistinen]

Yes, work is still going on, but only very little.

[redpawn]

Hi koistinen:

Exactly what kind of work you're talking about

[koistinen]

I am talking about improving/writing programs for computing and using 6 annd 7-man endgame tablebases. I think I have seen mention of some improvements in last 6 months and as for what I have been doing is changing my planned algorithm to make it easier to implement. (caring less about theoretical performance, which is kind of hard for me).

Having an index like 10x64x64^(N-2) the computation can be done in one sweep per side and iteration if you have 8G of ram. Doing it in one sweep per side and and iteration and piece would make it about 3 to 7 times slower for 7-man but be easier to code and would still be faster than what we have now and would require very little ram. I am trying to make myself code it that way, telling myself performance can be improved later.

[Kirill Kryukov]

I am trying to make myself code it that way, telling myself performance can be improved later.

I normally code a very dumb first version to get the initial statistics for the simplest endgames. Then I gradually implement faster versions, always comparing the statistics with previous runs. This way if I can't complete the next rewrite, I will still have something functional, even if not very fast. My coding priorities: 1. Clarity. 2. Correctness. 3. Speed+Features. Note that clarity comes before correctness, since you can debug (and then improve) clear code, but correct spaghetti code - you can't do anything with it. (Unfortunately looking at my code you'd think I have them in opposite order. How this happens is a mystery).