Skip to content

Conversation

@protonspring
Copy link

This is a functional simplification. The resulting values are substantially close to master, but no lookup is required, so this is presumably faster.

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 17785 W: 4020 L: 3890 D: 9875
http://tests.stockfishchess.org/tests/view/5d39c2160ebc5925cf0f0e03

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 28867 W: 4939 L: 4831 D: 19097
http://tests.stockfishchess.org/tests/view/5d3a04430ebc5925cf0f12e9

bench 3533582

bench 3533582
@snicolet
Copy link
Member

This is almost like a [0..4] test

@snicolet snicolet mentioned this pull request Jul 28, 2019
@protonspring
Copy link
Author

Combo Test passed:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 25465 W: 5684 L: 5570 D: 14211
http://tests.stockfishchess.org/tests/view/5d3e17070ebc5925cf0f6a00

@mrobaczyk
Copy link

mrobaczyk commented Jul 30, 2019

Is calculating r^4 better and faster than looking up in const array?
IMO const array are better - they are more flexible (you can tune it better).

Const array values: 0, 7, 8, 12, 29, 48, 86
R^4 function values: 7, 7, 8, 12, 23, 46, 88

@protonspring
Copy link
Author

After lots of benchmarks and testing. It does appear as though r^3 is faster than the array lookup, but r^4 is not. Closing.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants