Sesse__

20 ms·

301.
by Sesse__ 1y ago
What in my comments makes you think I haven't read it already? :-)
302.
by Sesse__ 1y ago
The first audio CD demonstrations were also like that. People drilling holes in the discs, smearing them with ketchup, etc., to show how reliable the system was. At least the next 15 seconds…
303.
by Sesse__ 1y ago
Why? Ubershaders were a common technique in graphics for many years before Dolphin adopted them.
304.
by Sesse__ 1y ago
Yes. (I'm not going into a pissing contest.)
305.
by Sesse__ 1y ago
https://distantprovince.by/posts/its-rude-to-show-ai-output-... In any case, if you really believe mmap is great for an assembler, then sure, go ahead. But it's not.
306.
by Sesse__ 1y ago
Why would anyone be interested in arguing against a confused AI?
307.
by Sesse__ 1y ago
Linux also uses mmap for running executables. But explicit I/O does not mean you have to start off by a gigabyte-long read().
308.
by Sesse__ 1y ago
You're probably thinking of gperf, not flex and bison.
309.
by Sesse__ 1y ago
mmap-based I/O still needs to go through the kernel, including memory allocation (in the page cache) and all. If you've got 10x speedup from mmap, it is usually because your explicit I/O was very inefficient; there are situat
310.
by Sesse__ 1y ago
FWIW, this is basically an implementation of perfect hashing, and there's a myriad of different strategies. Sometimes “switch on length + well-chosen characters” are good, sometimes you can do better (e.g. just looking up in a table in
311.
by Sesse__ 1y ago
Every problem in computer science can be solved with another layer of indirection. (And in particular, every design failure related to backwards compatibility can be solved with a virtual environment, evidently)
312.
by Sesse__ 1y ago
That “with compatible routing” thing pulls a lot of weight… I mean, if you have literal magic, then sure. Apart from that, IPv6 _is_ IPv4 with a bigger address space. It's so similar it's remarkable.
313.
by Sesse__ 1y ago
It is difficult to call Intel's E-core server chips a “clone” of Zen 5c. Sierra Forest was announced before Zen 4c, and they're actually designed-for-area cores instead of just resynthesizing the same chip with a different library
314.
by Sesse__ 1y ago
Is there _any_ confirmed information about Diamond Rapids, beyond one leaked (and seemingly slightly dubious) slide?
315.
by Sesse__ 1y ago
Are you talking about the fact that NVMe works by MMIO and DMA? So is pretty much any SATA controller, so there's no inherent difference there (there are _many_ years since the dominant way of talking to devices was through programmed
316.
by Sesse__ 1y ago
What about... single-core machines?
317.
by Sesse__ 1y ago
> And why does the SSD allow this to happen? A SSD has its own onboard computer, it's not just allowing the OS to do whatever it wants. If the device is DRAM-less, much of its central information (large parts of the FTL, in particul
318.
by Sesse__ 1y ago
`:not(.what) .ever` should be fairly fast, unless you have lots of `class="ever"` elements. Not ideal, but not as bad as e.g. `.ever :not(.what)` would be. `:has()` is just inherently slow if there's a significant amount of e
319.
by Sesse__ 1y ago
> You certainly can build slow apps with React, it doesn't make building slow things that hard. By all means. It sometimes feels like React is more the symptom than the actual issue, though. Personally I generally just like having l
320.
by Sesse__ 1y ago
Interesting how _everyone_ here blames JS and React, yet the fixes you linked are about CSS performance.
321.
by Sesse__ 1y ago
> and with the internal usage of the program (I only started in 2008, but spoke to Jeff extensively about the history of MR as part of Google's early infra) where the map function can be fed with recordio (list containers) or sstabl
322.
by Sesse__ 1y ago
> MapReduce is a great example of that- Google certainly didn't invent the concepts of Map or Reduce, or even the idea of using those for doing high throughput computing (and the shuffle phase of MapReduce is more "interesting&
323.
by Sesse__ 1y ago
There are _lots_ of objects in a large git repository. E.g., I happen to have a fork of VLC lying around. VLC has 70k+ commits (on that version). Each commit has about 10k files. The typical AI crawler wants, for every commit, to download
324.
by Sesse__ 1y ago
Well, like everything else: Is bandwidth your primary problem, or is CPU? Whenever I run apt now, download time is nearly nil but installation time is forever. Increasing installation time (and complexity, and disk space and cacheability on
325.
by Sesse__ 1y ago
C? Java? Perl? PHP? I have significant amounts of 20+ year old code in production, you just need to steer clear of the desire to use a huge framework/tons of dependencies and a language that breaks compatibility every few years.
326.
by Sesse__ 1y ago
It also does not add anything interesting to the discussion. Like, why add a bland summary of the article?
327.
by Sesse__ 1y ago
It seems to be some sort of read-only reimplementation of MySQL/Postgres that can ingest their replication streams and materialize views (for caching). Complete with a really primitive optimizer, if the article is to be believed.
328.
by Sesse__ 1y ago
This isn't really the same as MySQL's ICP; it seems more like what MySQL would call a “ref” or “eq_ref” lookup, i.e. a simple lookup on an indexed value on the right side of a nested-loop join. It's bread and butter for basic
329.
by Sesse__ 1y ago
You can do that just fine with the occasional video inside the article, though. I mean, significant parts of the video is just panning by some device while he's talking.
330.
by Sesse__ 1y ago
There's a perfectly good format for long-form dives: An article. But no, everything needs to be a video because otherwise, how would anyone bother to consume it.

More ›