m0th87

hace 12 mesesdiscuss

https://www.intel.com/content/www/us/en/docs/intrinsics-guid...

Intel's docs are unfortunately spartan, but the guarantees around program order is a hint that this is what it does.

m0th87

hace 12 mesesdiscuss

I see, thanks. I had assumed incorrectly that NT writes operated the same as NT accesses, where there is no dedicated cache.

m0th87

hace 12 mesesdiscuss

There are no guarantees even if everything operates on the same core. Rust docs have some details: https://doc.rust-lang.org/stable/core/arch/x86_64/fn._mm_sfe...

m0th87

hace 12 mesesdiscuss

I had interpreted GP to mean that you don’t slap on NTs for correctness reasons, rather you do it for performance reasons.

m0th87

hace 12 mesesdiscuss

I work on optimizations like this at work, and yes this is largely correct. But do you have a source on this?

> or (more likely) go into just some special small subsection of it reserved for non-temporal writes only.

I hadn’t heard of this before. It looks like older x86 CPUs may have had a dedicated cache.

m0th87

el año pasadodiscuss

That’s what I hope for, but everything that isn’t bananas expensive with unified memory has very low memory bandwidth. DGX (Digits), Framework Desktop, and non-Ultra Macs are all around 128 gb/s, and will produce single digits tokens per second for larger models: https://github.com/XiongjieDai/GPU-Benchmarks-on-LLM-Inferen...

So there’s a fundamental tradeoff between cost, inference speed, and hostable model size for the foreseeable future.

m0th87

hace 3 añosdiscuss

This is almost perfect for my needs (communication and location sharing in areas without phone signal.) But it's missing GPS/GNSS. I'm not sure how niche that use case is, so more broadly I wish Lilygo offered something akin to Raspberry Pi hats -- i.e. some way to add extensions.

m0th87

hace 3 añosdiscuss

I’ve heard Beaglebone Blacks are great (never used them), but have to say the experience with Blue was awful. I had a hardware defect that’s apparently commonplace. It gave me the feeling (maybe unjustified) that there’s a low bar for slapping the Beaglebone label on a product. In contrast, none of my several RPi’s have had any hardware issues, and I’m pretty brutal with them.

m0th87

hace 3 añosdiscuss

I’m definitely interested! Was just looking over this code and wishing it were in rust or C.

m0th87

hace 4 añosdiscuss

That's interesting, I didn't know they did that. But it could be better:

1) This alert doesn't show up at all on mobile.

2) IMO this should either affect the overall score or be displayed in search results as well, which is not currently the case: https://www.glassdoor.com/Search/results.htm?keyword=ZURU

m0th87

hace 4 añosdiscuss

For what it’s worth, the person you’re replying to is one of the heavyweights of the rust community.

m0th87

hace 5 añosdiscuss

Just whipped up a (naive) solve in rust: https://gist.github.com/ysimonson/01a1dee41b1b5990c30568fd25...

It usually solves this in under 6 guesses. The guessing could be improved; at the moment it's random, but it could select for words with non-repeating letters to narrow down the search space faster.

m0th87

hace 5 añosdiscuss

R2 is marketed as "one less than S3!"

I used the exact same tagline for a project open sourced in 2019: https://github.com/pachyderm/s2/

So there you go VCs, I can market as well as cloudflare. Money please!

m0th87

hace 5 añosdiscuss

I messed around with and contributed to this project a bit. The biggest challenge I saw in day to day usage was that line numbers aren't included in error messages: https://github.com/murarth/ketos/issues/11

m0th87

hace 5 añosdiscuss

I wrote up some rust low-level bindings for guile [1]. The problem is that Guile liberally uses setjmp/longjmp, which breaks rust destructors. It might be possible to fix this by wrapping every call the way rlua does [2], but I'm not familiar enough with Guile to know.

1: https://github.com/ysimonson/guile-sys

2: https://github.com/amethyst/rlua/issues/21

m0th87

hace 5 añosdiscuss

For rust, I manage a graph database that can be embedded as a library: https://github.com/indradb/indradb/

m0th87

hace 6 añosdiscuss

Understanding Power by Chomsky had a big impact on me. I read it after this compelling blog post by Aaron Swartz (RIP): http://www.aaronsw.com/weblog/epiphany

m0th87

hace 6 añosdiscuss

You're not alone. I've been writing go since before 1.0, and can handle modules/packages in _any other language_ fine, yet go's regularly leaves me stumped.

A little while ago we were hitting yet another issue with dependencies related to version pinning at work. I tried to divine what the disparate versions mean, and came up with this flow chart from a legalistic reading of stack overflow issues [1] and the go modules doc [2]:

  has the package opted into go modules (i.e. is there a go.mod?)
    yes
      does the repository use semver?
        yes
          is the major version <= 1?
            yes
              => normal versioning, e.g. github.com/stretchr/testify v1.3.0
            no
              => when importing, we need add /vN at the end, e.g., import "github.com/my/mod/v4", and in go.mod it will behave like github.com/my/mod/v4 v4.1.0
        no
          => pseudo-versioning will be used, e.g. github.com/kolo/xmlrpc v0.0.0-20190717152603-07c4ee3fd181
    no
      does the repository use semver?
        yes
          is the major version <= 1?
            yes
              => pseudo-versioning will be used, e.g. github.com/kolo/xmlrpc v0.0.0-20190717152603-07c4ee3fd181
            no
              => version will be marked as incompatible, e.g. github.com/zeromq/goczmq v4.1.0+incompatible
        no
          => ???

I still don't know if this is right. But that's not really the point. I think the go authors painted themselves into a corner with such an anemic initial release. Now that they're trying to address shortcomings, the complexity is blowing up to handle the very large/diverse ecosystem. It reminds me a lot of java around 1.4 when they added in generics. In the long-run it was the right move, but man was it painful.

1: https://stackoverflow.com/a/57372286 2: https://github.com/golang/go/wiki/Modules

m0th87

hace 6 añosdiscuss

It's not well known, because it's not true, at least not absolutely. Speed could either mean latency or throughput. Reference counting is almost always worse than GC at throughput, but is usually better for latency.

m0th87

hace 6 añosdiscuss

Yes, though this is arguably a big reason why SCTP didn't take off, which is similar to QUIC in a lot of ways (there is SCTP over UDP, but it was a second class citizen until WebRTC.)

Submissions

Markets, Discrimination, and “Lowering the Bar”

WaveFunctionCollapse

Taking ZUFS Upstream

Genera concepts

The QNX Neutrino Microkernel (2007)

Tokamak: A Rust IDE for Atom

My gamedever wishlist for Rust

My gamedever wishlist for Rust

Hostage saves herself via Pizza Hut app

CartoDB

Python3 features you can't use in python2 (2014) [pdf]

Git-fit

Cosmos

Beating A/B Tests

Prerender

Swift object storage

Objc_msgSend() Tour

Tornado gives you wings

Pilbox

Libc considered harmful

comments