Teaching and learning resources • Re: Advent of Code 2023

The code posted above is slightly different than what's on Github

Yes, the Github version is multithreaded which doesn't work on the 32 bit OS running the Pi 0 (doesn't support 64 bit atomics). This is an earlier single threaded version, using the same logic.

Surprising that the Ryzen 4650 only achieves the same performance as the Pi5.

On Ryzen 7 3700X, with 4 threads I get

Code:

--- Day 12: Hot Springs ---Part 1  - 7694Part 2  - 5071883216318Elapsed - 10.043 ms.

I wonder if differences with the puzzle input could account for the longer times.

Also, I optimized with -O3 but didn't specify any architecture specific flags.

$ ./lurk12 # Pi Zero 1000 MHz--- Day 12: Hot Springs ---Part 1  - 7694Part 2  - 5071883216318Elapsed - 1221.5 ms.$ ./lurk12--- Day 12: Hot Springs ---Part 1  - 7694Part 2  - 5071883216318Elapsed - 1102.53 ms.$ ./lurk12--- Day 12: Hot Springs ---Part 1  - 7694Part 2  - 5071883216318Elapsed - 1064.17 ms.$ suPassword:# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_governorperformance# vcgencmd measure_clock armfrequency(48)=1000000000

Statistics: Posted by ejolson — Thu Jan 11, 2024 3:38 am