There were 84 errors for GCP, but the breakdown says 74 409s and 5 timeouts. Maybe it was 79 409s? Or 10 timeouts? I suspect the 409 conflicts are probably from the instance name not being unique in the test. It looks like the instance name used was: instance_name = f"gpu-test-{int(time())}" which has a 1-second precision. The test harness appears to do a `sleep(1)` between test creations, but this sort of thing can…
I've naively used millisecond precision things for a long time - not in anything critical I don't think - but I've only recently come to more of an awareness that a millisecond is a pretty long time. Recent example is that I used a timestamp to version a record in a database, but it's feasible that in a Go application, a record could feasilby be mutated multiple times a millisecond by different users / processes / re…
Re: AWS vs. GCP reliability is wildly different
#241To put this into perspective, a game (or anything) running at 60 FPS only as a bit over 16 milliseconds to render each frame. These days higher frame rates are common enough, often putting you down into single-digit milliseconds per frame. Not think about how many things there are simulated and rendered in each frame for common games. Way more than 16.