> 35x less system calls = others wait less for the kernel to handle their system calls
That isn't how it works. There isn't a fixed syscall budget distributed among running programs. Internally, the kernel is taking many of the same locks and resources to satisfy io_uring requests as ordinary syscall requests.
More system calls mean more overall OS overhead eg. more context switches, or as you say more contention on internal locks etc.
Also, more fs-related system calls mean less available kernel threads to process these system calls. eg. XFS can paralellize mutations only up to its number of allocation groups (agcount)
> More system calls mean more overall OS overhead [than the equivalent operations performed with io_uring]
Again, this just isn't true. The same "stat" operations are being performed one way or another.
> Also, more fs-related system calls mean less available kernel threads to process these system calls.
Generally speaking sync system calls are processed in the context of the calling (user) thread. They don't consume kernel threads generally. In fact the opposite is true here -- io_uring requests are serviced by an internal kernel thread pool, so to the extent this matters, io_uring requests consume more kernel threads.
syscalls are expensive and their relative latency compared with the rest of code only grow especially in view of mitigations against cache-related and other other hardware bugs.
Under the "tools" part of the README it shows the following observation tools:
- browser_snapshot_dom
- browser_query
- browser_accessible_tree
- browser_read_text
- browser_screenshot
So most likely the LLM can chose how to "see" the page?
I was being a bit sarcastic when I asked "which winlator?" because there are sooo many variations of it. At this point none is more official than the other. Maybe people picked up on that and didn't value the url as much.
> The only true random sources in the universe are quantum. But it's not practical or performant to hook your workload up to a raw quantum measurement like the spin of a fundamental particle, or radioactive decay.
I was under the impression that's how entropy generators for servers work?
That is good example why one timezone doesn't work. The locals in Xinjiang use a local time zone +6, instead of China time +8, because the latter is too far off the daylight hours.
My understanding is that use of Xinjiang time has dropped recently because of the crack down on Uygurs and government forcing China time.
There are various countries that optimize the number of time zones for administrative purposes, but this is much easier and sensical to implement within one country than globally.
UTC is used globally when it makes sense, e.g. for the schedules of international radio broadcasts.
