Re: Unboxed structures (yes, again, only this time for real)

On 16 Jan 2026, at 18:24, Martin Simmons <[email protected]> wrote:
> 
> I wouldn't trust any timings that are less than a few seconds.  This is
> especially true on hardware where the clock speed varies according to CPU
> usage (hence the same code runs more slowly until the speed has ramped up).

FWIW (Martin will know this), the opposite is also true.  If you write something that makes the machine work hard enough then it can end up being thermally throttled on a system like a laptop which can't deal with dissipating the heat.  This is certainly true for an Apple M1 laptop: I have some code which manages to pretty much saturate the four high-performance cores in the thing and you can see it being throttled after some tens of seconds or perhaps less.  So there's a window of long-enough-to-be-useful vs short-enough-to-not-get-throttled you have to aim at.

I don't know if a single thread can cause thermal throttling on an M1.

Obviously won't apply to systems which can deal with the heat.

--tim