Hacker News: crest

New comment by crest in "Resurrecting iChat audio and video conferencing"

crest — Tue, 06 Oct 2026 10:42:04 +0000

Having to optimise for three very different microarchitectures with the G3 (short narrow pipeline, no AltiVec, small caches), G4 (short wide pipeline, focus on in-order AltiVec, useful OoO otherwise, large L2 yet memory bandwidth starved), and G5 (long and wide pipeline, focus on super-scalar OoO, AltiVec just glued on, more memory bandwidth) must have been "fun".

How many variants of the codecs did you have to implement to make it work?

New comment by crest in "Treachery in the Rodin Museum 3D scan verdict"

crest — Mon, 05 Oct 2026 02:20:53 +0000

Sometimes a bit of wood, rope, and a sturdy blade will help.

New comment by crest in "GPT-6 Astra has gained the ability to drive a car"

crest — Wed, 23 Sep 2026 17:20:21 +0000

"You've reached your quota. Please hold your credit card or mobile phone on the card reader within the next 10 seconds or you'll be liable for the resulting crash."

New comment by crest in "Apple Mac mini review"

crest — Mon, 21 Sep 2026 15:51:27 +0000

Best of luck getting a good deal for new hardware right now from anyone right now. All hardware pricing negotiations these days start with give us your company credit card, drop your pants and bend over the table.

New comment by crest in "The scourge of x86 emulation"

crest — Fri, 18 Sep 2026 19:23:53 +0000

Does any other architecture allow split cacheline atomic operations across sockets?

New comment by crest in "OpenAI expands ChatGPT ads with Sponsored Agents"

crest — Wed, 16 Sep 2026 14:13:17 +0000

"Nuke the entire site from orbit ... it's the only way to be sure."

New comment by crest in "GEFS on OpenBSD: A Early Preview"

crest — Wed, 16 Sep 2026 00:55:45 +0000

They also fucked up their parity RAID repair so that the next disk failure on a different disk will destroy the data, argued that the pathological B-tree performance reported by a user couldn't be real, because they didn't understood the tree depth would only be logarithmic to the base of the fanout if the entries are the same size. insert shocked pikachu face

New comment by crest in "List of references on Sony websites to players "owning" their digital games"

crest — Fri, 11 Sep 2026 10:47:30 +0000

No reasonable person has ever argued that a consumer buying a (digital) copy of a game obtains the exclusive rights to the game. This is such a cheap embarrassingly stupid attempt to come up with a straw man.

I'll listen to Sony management and lawyers whining about copyright after they paid the damages and fines and get out of jail for every rootkit they deployed. Should be any century now.

New comment by crest in "FreeBSD 14.5-Release"

crest — Wed, 09 Sep 2026 11:05:33 +0000

A CDN *cache* server is a special case. Netflix cache appliances can accept the risk of running FreeBSD -CURRENT and upgrading half their fleet to the latest snapshot every ~2 weeks and we thank them for battle testing FreeBSD's active development branch in production. It really helps to reveal regressions (both correctness and performance) early in the subsystems and drivers that matter to their usecase.

Most users will neither be willing nor able to accept the trade-offs that Netflix chose.

New comment by crest in "FreeBSD 14.5-Release"

crest — Wed, 09 Sep 2026 08:48:33 +0000

If you run the latest (laptop) hardware you may still have to track stable or backport some patches from -CURRENT.

New comment by crest in "FreeBSD 14.5-Release"

crest — Wed, 09 Sep 2026 08:47:07 +0000

You're just wrong. FreeBSD stable branches are the development branches from which new minor releases (with a stable ABI) are forked. If you want to be pedantic you want to follow the releng branches for most production deployments (release + security and non-security errata patches). Unless you build from source yourself you the tools (pkgbase, freebsd-update) don't make that distinction visible to the user.

New comment by crest in "Async Rust vs RTOS showdown (2022)"

crest — Thu, 03 Sep 2026 11:18:02 +0000

The ARM v7M architecture helps a lot here because the NVIC supports priority inheritance without disabling interrupts. Yes you have to assign your priorities correctly and yield CPU time for anything compute bound, but again the architecture helps you in the form of the SVCall+PendSV exception pair. The PendSV handler can easily be multiplexed on the Cortex-M3/M4 using atomic operations on bit-banded memory. I've needed only a few dozen lines of Forth to set this up on bare metal without an RTOS to split peripheral drivers in a short interrupt handler saving the required state and delegating protocol handling to a lower priority callback.

New comment by crest in "Async Rust vs RTOS showdown (2022)"

crest — Thu, 03 Sep 2026 11:11:55 +0000

A 1GHz superscalar Cortex-M7 can still hit a 12 cycle interrupt latency if the system stack is kept in the 2x32 bit wide zero cycle dTCM, and the VTOR and handler are in the iTCM thanks to lots of bandwidth and a reasonably short pipeline. The NVIC is truly the underappreciated superpower of the Cortex-M family. You can implement a full RTOS or complex bare metal application without ever having to disable interrupts. On other MCU architectures (most RISC-V, PIC32, even older crap refusing to die) you'll easily end up with closer to 100 then 10 cycles of jitter from critical sections having to run with disabled interrupts e.g. context switching code.

New comment by crest in "Arbitrary code execution in QubesOS via copy-to-VM error reporting backchannel"

crest — Sun, 30 Aug 2026 17:53:19 +0000

Really an attacker controlled system() call?!?

New comment by crest in "Memory Ordering in CPUs"

crest — Thu, 27 Aug 2026 01:34:07 +0000

One advantage they have for building larger systems is that they allow you to crack stores into two parts. The part that has to retire in order for the local CPU and the write back to memory. Now assume you have multiple banks of memory and one of them stalls. A system with TSO can't allow the other banks to make progress because that would make the stores visible to other CPUs (or peripherals) out of program order. So you have to stall stores to otherwise ready memory banks and hold on to valuable ROB and store queue resources for longer.

New comment by crest in "GitHub Publishes IPv6 Addresses for Git SSH remotes"

crest — Thu, 27 Aug 2026 01:24:47 +0000

It must be a snow day in hell.

New comment by crest in "US sanctions Italian hosting provider Autistici Inventati"

crest — Thu, 27 Aug 2026 01:04:20 +0000

Should we also ask the Saudis what drawings to punish by beheading or the Taliban which writings to punish by stoning?

New comment by crest in "Run OpenBSD on DigitalOcean for $4/month"

crest — Tue, 25 Aug 2026 19:30:42 +0000

They're also too stupid to understand that they've been dropping UDP datagrams with src port == dst port and < 1024. This obviously breaks IPsec IKE. I showed them dummy traffic captured with netcat and tcpdump, but they refused to admit they caused the problem. They repeatedly claimed to see nothing dropped between the pcap files, that I was doing it to myself with a firewall that wasn't even enabled, or that this is just how a network is supposed to work.

New comment by crest in "Apple introduces M6 and M5 Ultra"

crest — Tue, 25 Aug 2026 18:22:31 +0000

A lot of the important applications took a while to get ported. Also the last two generations of G4 hardware just overclocked basically unchanged chips. The 12" G4 laptop was infamous for slowly cooking your junk with their titanium shell.

New comment by crest in "IPFS Maintainers Winding Down"

crest — Mon, 24 Aug 2026 18:52:40 +0000

What people did or planned to do with IPFS is pretty silly, but the idea of content addressed storage with cryptographic (not in a crypo coin way) names makes a lot of sense. It would make for a great software distribution method e.g. your game downloads get faster the more people want to play the latest version. It's the combination of Git and BitTorrent and has uses far beyond just hosting warez and NFTs.