Chess Engines binaries for Raspberry Pi 4 Linux 64 bits

[cpeters]

Well, I thought it was a test/worth a try.

Can you post please, when you're doing this (the statically linked too please) in a terminal? It may give me a hint for trying out other compiling (which it did with no errors) options.

Greetings

Edit:

FIXED!!! with chmod +700 fire8

Do they both calculate/are stable? How is the speed (is there a difference)? With your advise I can try Dietrichs toga then.

[AlexChess]

UPDATE:
Houston, we have a problem

It starts thinking, but after a while it exits:

Code: Select all

alex@alex-Parallels-ARM-Virtual-Machine:~/Scrivania/cutechess$ ./fire8
Fire 8.1 x64 popcnt
info string 2 available cores
go infinite
info time 1 nodes 1 nps 1000 tbhits 0 depth 1 seldepth 2 multipv 1 score cp -2 pv a2a3
info time 1 nodes 2 nps 2000 tbhits 0 depth 1 seldepth 2 multipv 1 score cp 0 pv b2b3
info time 1 nodes 3 nps 3000 tbhits 0 depth 1 seldepth 2 multipv 1 score cp 10 pv c2c3
info time 1 nodes 4 nps 4000 tbhits 0 depth 1 seldepth 2 multipv 1 score cp 27 pv d2d3
info time 1 nodes 5 nps 5000 tbhits 0 depth 1 seldepth 2 multipv 1 score cp 42 pv e2e3
info time 2 nodes 27 nps 13000 tbhits 0 depth 2 seldepth 4 multipv 1 score cp 45 pv e2e3 a7a6
info time 2 nodes 56 nps 28000 tbhits 0 depth 2 seldepth 4 multipv 1 score cp 78 pv e2e4 a7a6
info time 2 nodes 84 nps 42000 tbhits 0 depth 3 seldepth 6 multipv 1 score cp 49 pv e2e4 c7c6 d2d4
info time 2 nodes 154 nps 77000 tbhits 0 depth 4 seldepth 8 multipv 1 score cp 40 upperbound pv e2e4 c7c6
info time 2 nodes 265 nps 132000 tbhits 0 depth 4 seldepth 9 multipv 1 score cp 25 upperbound pv e2e4 c7c6
info time 2 nodes 399 nps 199000 tbhits 0 depth 4 seldepth 9 multipv 1 score cp 35 pv e2e4 c7c6 d2d4 e7e6
info time 3 nodes 537 nps 179000 tbhits 0 depth 5 seldepth 9 multipv 1 score cp 28 upperbound pv e2e4 c7c6
info time 3 nodes 1058 nps 352000 tbhits 0 depth 5 seldepth 10 multipv 1 score cp 25 upperbound pv e2e4 c7c6
info time 4 nodes 1448 nps 362000 tbhits 0 depth 5 seldepth 11 multipv 1 score cp 11 upperbound pv e2e4 d7d5
fire8: movegen.cpp:249: s_move* movegen::moves_for_pawn(const position&, s_move*, uint64_t) [with side me = (side)1; move_gen mg = captures_promotions; uint64_t = long unsigned int]: Assertion `rank(pos.enpassant_square()) == relative_rank(me, rank_6)' failed.
Annullato (core dump creato)
alex@alex-Parallels-ARM-Virtual-Machine:~/Scrivania/cutechess$ 

Regards, Alex

[cpeters]

Well the shyte hits the fan at least in an early stage of calculation.


I have to get myself a proper Pi...

Gretings

[jshriver]

Well the shyte hits the fan at least in an early stage of calculation.

I have to get myself a proper Pi...

Gretings

I'll try to make a build later today, tried on first post but having to make a custom Makefile for armv8 w/ neone

[JohnWoe]

Looking at benchmarks it looks like Pi 4 8GB performs like some crappy old and slow i3 laptop.
I would love to have some super small credit card size computer.
But it would be so slow it's totally useless.
Maybe some shit with leds idk

It costs like 300€ after case + GPU. I could get a used 8th gen i5 laptop for that.

I'll check after Pi 7th gen. If there's enough horsepower. Plus credit card size

[AlexChess]

Looking at benchmarks it looks like Pi 4 8GB performs like some crappy old and slow i3 laptop.
I would love to have some super small credit card size computer.
But it would be so slow it's totally useless.
Maybe some shit with leds idk

It costs like 300€ after case + GPU. I could get a used 8th gen i5 laptop for that.

I'll check after Pi 7th gen. If there's enough horsepower. Plus credit card size

Igel 3.0.0 NNUE bench (starting position):

-PI4 compatible Ubuntu 20.04 ARM64 4CPUs: 4 mN /s (Under Parallels Desktop 16.3 M1) jsriver's compilation
-Windows 10 ARM64 NEON native 4CPUs 3 mN /s (Under Parallels Desktop 16.3 M1) official engine authors compilation
-Mac mini M1 Big Sur M1 native 4CPUs 7 mN /s acepoint's compilation

hsriver can give us the speed on Raspberry Pi4

[AlexChess]

Another strong open source challenger (and challenge to compile for Pi4 and Ubuntu ARM64) for my tutors cpeters and jshriver
LCfish NNUE
https://github.com/borg323/lc0/releases/tag/lcfish

[AlexChess]

Sorry, it seems that LcFish is only an April 1 joke.

[jshriver]

Here is a pi4 64bit armv8 build of Clover 2.1.2.1, sorry didnt make a pi 0 one but can if someone is interested.

http://chess.nyc3.digitaloceanspaces.co ... .1.2.1.zip

[amanjpro]

Zahak 2.0.0 for Pi 4 (64-bit):

https://github.com/amanjpro/zahak/relea ... rm64-2.0.0

and 32 bit for all Pi models:

https://github.com/amanjpro/zahak/relea ... rm32-2.0.0