HNHacker News
TopNewBestAskShowJobs

wtallis

22,096 karma · joined September 20, 2009

submissionscomments
wtallis··on Coreutils for Windows
The filesystem may choose to store new data at different logical block addresses than older versions. The SSD will definitely choose to store those newly written blocks at different physical addresses, both for the sake of wear leveling and for performance, because a read-erase-rearite cycle on an entire NAND flash erase block (several MB at minimum) is a very slow operation.
wtallis··on Nvidia RTX Spark
I think DGX Spark has poor memory bandwidth because these laptops were the plan all along. NVIDIA didn't want to commit to the extra costs of a 512-bit memory bus for their first laptop SoC, so they went with the more modest 256-bit bus, same as AMD did for Strix Halo.
wtallis··on Nvidia RTX Spark
There are two new things being announced here: the GB10 chip being put into laptops, and GB10 running Windows. GB10 running Linux is not news, it's a product that's been shipping since last fall.
wtallis··on Nvidia RTX Spark
GB300 is nominally "available" in desktop form factor workstations priced around $100k. That's a few orders of magnitude away from the ordinary desktop PC market that consumers participate in.
wtallis··on Nvidia RTX Spark
NVIDIA hasn't done custom CPU cores for anything they've yet branded "Grace". The original Grace data center CPU (paired with the Hopper data center GPU) used ARM Neoverse V2 cores. The "GB10" chip shipped in DGX Spark and announced here for RTX Spark uses Cortex X925 and Cortex A725 CPU cores.

Physically, NVIDIA did the GPU chiplet and Mediatek did the other chiplet that has the CPU, DRAM controller, and IO.

wtallis··on What Is a Dickover?
It's my computer, not yours. The browser is my user agent, not your server agent. If you don't want me viewing your page according to my preferences, then the world wide web is not the right medium for you to be working in. Go write a native app and try to convince people to install it. Once you serve up the contents of a web page to my user agent, you have consented to letting me digest that content however I want.

This is not a sense of "entitlement", it's just the fundamental reality of what the web is.

wtallis··on Why Gentoo?
> One day you'll be happy with your pam config, then bam, package updates and you're forced into passwdqc and faillock.

Are you saying you've had your PAM config broken by a package update when you didn't update anything in /etc, or that you had your PAM config broken by a package update where you blindly accepted changes to stuff in /etc?

wtallis··on Claude Opus 4.8
Raw training data is raw. A really big model trained on it has already done a first-pass of finding patterns and squeezing out redundancy. Re-ingesting the full training set to train a smaller model is probably more expensive, for marginal quality improvement over distilling from the large model.
wtallis··on I'm Getting into Mesh Networks (Meshtastic, MeshCore, and Reticulum)
The request was for "a proper, working emergency meshing standard". Ham radio absolutely is not that. Yes, ham radios are used in emergency situations, but usually not as a mesh and certainly not as a standard.
wtallis··on Warm up your MacBook (2019)
For a very long time now, it has been the case that in the short term most processors will boost high enough for the die to reach around 100°C. When you see a reading substantially lower than that like your example of 75°C, either the system has throttled for a reason other than processor die temperature (eg. throttling to limit the temperature of the outer case of the machine) or the temperature reading you're seeing does not correspond to the hottest part of the chip and the throttling is based on the presumption that a different part of the chip that is not directly monitored will be much hotter.

In the specific case of the i7-7700T, the "T" suffix for Intel CPUs usually means you have mainstream desktop silicon with arbitrarily reduced long-term turbo limits, intended to be used in small form factor PCs with limited cooling capacity. Its limitation to 35W sustained and official Tj max of 80°C are artificial and essentially fiction, and the same silicon will readily do 91W sustained with a Tj max of 100°C as seen on the i7-7700K.

Processor temperature under load tells you almost nothing about the power draw or efficiency of the chip, because the temperature can be controlled to almost any value desired through a combination of varying cooling effort and varying clock speeds.

wtallis··on Motorola phones have started hijacking the Amazon app to insert affiliate codes
Asus has been doing it since before that article was written. It's amazing that a "put rootkit here" Windows feature has been around for so long.
wtallis··on California moves to exempt Linux from its age-verification law after backlash
> every computer, even the tiny embedded CPU in your microwave oven, might just need to ask your age before it allows you to do anything.

Why do I keep seeing this bullshit exaggeration? It doesn't help anything to make such ridiculous statements. Nobody's microwave oven has an app store.

wtallis··on Bug 1950764: Work Around Crash on Intel Raptor Lake CPU
It's a dumb question, because it's in reply to a comment that already implies the answer, and it's trivial to find an answer online in less time than it takes to post that question and wait for someone to supply an answer.

The subject of CPU microcode update mechanisms is an interesting and relevant topic, but such a shallow, low-effort question is not a good way to promote interesting discussion on that topic.

wtallis··on Don't Roll Your Own
> You should always write your own code whenever possible, that's how we get better things.

No, that's how you end up with a mountain of bad half-assed implementations. You should only roll your own when truly necessary, and only after thoroughly understanding the problem and the existing mature solutions and honestly comparing that against your own ideas to see if you're missing important aspects of the problem.

And even after doing all that, when it comes to implementing custom UI for something that already has a standard approach, you should still usually throw away your custom version because it won't be better by a wide enough margin to justify the effort users will need to invest in re-learning and breaking their existing habits.

Save your UI experimentation for your personal tools, and don't inflict it on innocent users.

wtallis··on Mounting git commits as folders with NFS (2023)
Files and folders can be read-only, a concept that has been around for about as long as the folders abstraction itself.
wtallis··on Nobody understands the point of hybrid cars [video]
The car wash in this video is a box with a solid floor and doors that close at the ends before the wash cycle starts. There's no reason for the car to be put in neutral for this car wash; it's not doing anything to move the car around. You park your car in this kind of car wash.
wtallis··on Tesla Wall Connector bootloader bypasses the firmware downgrade ratchet
> In fact, I believe I have never before heard someone combine the meaning of the word to be related to the ownership of the device being hacked.

Ownership implies authorization, and using the term "hacking" exclusively to refer to gaining unauthorized access to someone else's computer is a common usage that you must surely have encountered before.

wtallis··on Linux gaming is faster because Windows APIs are becoming Linux kernel features
A lot of that flexibility is what makes it hard to efficiently emulate (especially without kernel level support), but some of it seems too flexible to make sense as the default choice. How often does a video game really need a lock that can be shared between processes, and why should that lock type be the one that a game engine uses for almost all of its locks?
wtallis··on I’ve banned query strings
Postel's law is typically stated as "be conservative in what you do, be liberal in what you accept from others". It's unfortunately common for people to ignore the first half and hallucinate a third clause demanding that the recipient stay silent about the errors they receive.
wtallis··on Apple, Intel have reached preliminary chip-making deal
> It also was that Intel couldn’t execute reliably on their own roadmap,

Intel kept putting out delusional roadmaps that would assume their 10nm fab process was going to be ready for mass production in just another quarter or two. They spent years refusing to plan for 10nm to not be ready, so all their new architectures were unshippable and they had to resort to just using copy and paste on their 2015 CPU cores. Their fab fuck-up was hardly the only mistake they made in that era, but it was the biggest underlying cause of their problems.

wtallis··on A web page that shows you everything the browser told it without asking
That's much more reliably conveyed by looking at the viewport dimensions.
wtallis··on Apple, Intel have reached preliminary chip-making deal
Even if we make the not-particularly-reasonable assumption that Apple would want to own a fab, and the wildly unreasonable assumption that they would accept any outside customers for the fab: building a fab business from scratch to something competitive would take on the order of a decade or more even for a company with Apple's resources. And if Apple bought somebody else's fab business, it seems most likely that it would be Intel's and there would no longer be an Intel do design chips that would be in search of a foundry. (Intel has the only relevant logic fab business that could plausibly end up getting sold off to the likes of Apple.)
wtallis··on An Introduction to Meshtastic
The cookies used by the site appear to almost all be from Cloudflare, GitHub, YouTube: service providers that are not at all interested in enabling a cookie-free web site, and chose this route as malicious compliance. There's also one cookie to store a language preference that could be handled through HTTP headers instead.
wtallis··on DeepSeek 4 Flash local inference engine for Metal
Exactly. In the next few years, entry-level hardware will not be advancing beyond 16GB. And anything beyond 32GB will remain decidedly high-end.

And that's for laptops with unified memory. In the desktop space, 8GB discrete GPUs are going to be sticking around for a very long time.

wtallis··on I switched from Mac to a Lenovo Chromebook
The type-a ports that are 30 years old, have always been frustrating to orient correctly, and are significantly constrained in data rates and power delivery compared to type-c that has been common for a decade? Yes, calling that "old school" is reasonable.
wtallis··on 245TB Micron 6600 ION Data Center SSD Now Shipping
That's only going to change on similar timescales to the fab capacity, which is to say multiple years.
wtallis··on StarFighter 16-Inch
If an AI crash left the market with a glut of RDIMMs, we would absolutely see AliExpress sellers removing the individual DRAM packages from used RDIMMs and installing eight out of ten of them onto empty UDIMMs. The overhead costs would be a quick bit of hot air reflow, a 5V PMIC and associated components, and a new module PCB, and the sellers would get to build 10 UDIMMs for every eight RDIMMs they scavenge.
wtallis··on Google Chrome silently installs a 4 GB AI model on your device without consent
> and anyone who wants consistent rendering of sites

Yeah, that's where you lose me. The people who want to guarantee pixel-perfect rendering of web sites tend to be the same people who feel entitled to run their code on my machine as part of delivering their site, who prioritize their brand identity over my accessibility settings, who want to build their complex web app UI with no respect for the conventions of my chosen operating system. My browser should not be an ally to them; it should be taking my side in those conflicts.

wtallis··on 245TB Micron 6600 ION Data Center SSD Now Shipping
> There are many computer components competing for the same resources (fabs, wafers).

That's a bit off the mark. Processors, RAM, and flash memory each require their own specialized fabs. TSMC makes processors but not DRAM or NAND flash. Kioxia makes flash only. Samsung has all three types of fabs. Micron does DRAM and NAND, in different fabs.

The increase in SSD prices is not because there are many components competing for the same constrained resources, it's that they are complementary goods being subject to the same dynamics in parallel because the servers that are causing this demand spike need all of those components. Where we do see competition for the same fab capacity is in the mix of DRAM types, where GPUs want HBM, server CPUs want DDR5, laptops and phones (and the occasional server CPU) want LPDDR. And competition among different types of processors for TSMC's fab capacity.

wtallis··on 245TB Micron 6600 ION Data Center SSD Now Shipping
A drive's write endurance rating is derived at least in part from the JEDEC standard data retention requirements: 1 year at 30C for consumer drives, 3 months at 40C for enterprise drives, IIRC. Thus, a drive that has reached the end of its rated write endurance can be expected to have those retention characteristics. A drive that hasn't been subjected to that much wear will have significantly longer retention.
← PreviousPage 6 of 34Next →