brigade

19 ms·

211.
by brigade 4y ago
A nuance: they might not be able to revoke existing OGL 1.0 licenses for anyone that has already used it to publish derivates, but they can likely cancel the offer , even for existing OGL 1.0 content, for anyone that hasn't yet relied
212.
by brigade 4y ago
Vectorizing text handling is likely to use generalized register permutes (and I spy some _mm512_shuffle_epi8 here), which are the bugaboo in length-agnostic SIMD. Fundamentally, the maximum index you can read from in a register depends on
213.
by brigade 4y ago
> But I don't think there's much argument to be had here that the M1/M2 Apple cores are just bigger > You pretty much can make 2 cores fit inside of the M1 core My entire point has simply been debunking this. I'm p
214.
by brigade 4y ago
I tried gemm-benchmark on my M1 Max, and it took 22W to hit 2.2 Tflops with AMX (accelerate) or 36W to hit 270 GFlops with NEON (OpenBLAS) So that's actually just about as power-efficient for fp32 as a 3090, which according to wikipedi
215.
by brigade 4y ago
I mean, I was trying to be fair by counting L2 and associated logic. If you only count down to L1 (and no, Zen4's 15-cycle latency L2 is not comparable to Apple's L1 that achieves a 3-4 cycle latency; Zen4's combined L2+L
216.
by brigade 4y ago
Well yeah, area budgets per-core have never shrunk linearly with transistor density; it serves a wider range of use cases to balance beefing up cores with adding more of them. Like, Intel 7 is 20x denser than Intel 32nm, but a Sandy Bridge
217.
by brigade 4y ago
M1, M2, and Zen 4 are all on TSMC 5nm TSMC only just started mass production of 3nm like last week, nothing has shipped yet.
218.
by brigade 4y ago
Area-wise I'm pretty sure Apple's P-cores are in line with Intel and AMD - including L2 cache and other shared logic, M1's P-core was 4mm^2 per core, M2's P-core is about 5mm^2 [1], Zen 3 and 4 were about 4mm^2 [2], and
219.
by brigade 4y ago
> RISC-V is designed with great care taken to weight all decisions as not to hamper any scope of implementation, from the lowest power microcontrollers to the fastest supercomputers, and everything in between. RISC-V was consistently des
220.
by brigade 4y ago
Intel doesn't publish or even have a "target operating temperature" or "working temperature"; I have no clue what you're pulling that out of.
221.
by brigade 4y ago
OSHA didn't consider a properly adjusted laptop stand acceptable? Well, I only use one to reuse the laptop as a second display...
222.
by brigade 4y ago
Yeah, these days it's likely for only the frontend to stall during misprediction recovery. Already decoded µops from before the branch still make their way through the backend as normal. If you have enough of those (long dependency cha
223.
by brigade 4y ago
Tcase is different than Tjunction. Tjunction has been around 100 since like Yonah at least.
224.
by brigade 4y ago
It’s also a decade older and benchmarks at half the performance. I’d bet that 90% of the issue is that make, /bin/sh, and autotools all use fork() on macOS instead of posix_spawn() (I think that’s still the case…), and fork is slo
225.
by brigade 4y ago
The right solution is not thinking about it at all. The only extant malicious chargers anyone has shown evidence of have been proof-of-concepts whose concept hasn't worked in nearly a decade. If you're worried about purely theoret
226.
by brigade 4y ago
Of these, I think the only one that can work solely because of an overly "flexible in what you receive" is the USB condoms, which shouldn't be encouraged in the first place. Specifically, that people wanted to charge at more
227.
by brigade 4y ago
The CONSTRAINED UNPREDICTABLE handling - it's effectively implementation defined, but ARM constrains the implementations to always choose one of 4 possibilities: - Whether subsequent non-faulting elements are loaded as normal, or ignor
228.
by brigade 4y ago
I think we all know exactly how this policy is enforced: a regex for banned URLs. If whatever regex they come up with matches the link Carmack posted, it's banned. If not, it's banned if it personally catches the negative attent
229.
by brigade 4y ago
For this batch... most of the memory safety issues look to be in kernel code or adjacent, followed by Safari/Webkit. The Android Linux kernel is still pure C (for now), and I'm pretty sure new Chrome/Blink code is still prima
230.
by brigade 4y ago
Decoding the “changes” can be dependent on the “changes” in a previous frame, notably that the prediction for motion vectors can come from previous frames’ motion vectors. Which I frames obviously don’t have. Besides that the only way to lo
231.
by brigade 4y ago
Earlier this year, used Teslas were selling for upwards of $10k over what Tesla sold them for new, and Tesla kept bumping up the MSRP. It was the worst used/new value ratio across the industry, probably since there are no dealers you
232.
by brigade 4y ago
Personally, it’s not that surprising that comics written for a smartphone sized infinite scroll canvas are now outearning those designed for book-sized canvasses. Same thing happened with mobile games over consoles, and I haven’t seen anyon
233.
by brigade 4y ago
Just as I don’t need research either to know that dark mode inherently increases eye strain thanks to the afterimages caused by using it for more than a couple dozen seconds at a time. I have my own two eyes after all, and plenty of other e
234.
by brigade 4y ago
That's as useful as saying DNG can be powered by lossless JPEG. No standard lossless JPEG decoder can do anything useful with DNG files.
235.
by brigade 4y ago
Not sure about your comparison specifically, but in general the hard edges in comics should be an ideal case for AV1's spatial prediction. JPEG-XL eschewed that in favor of being inherently progressive; its splines promised to make up
236.
by brigade 4y ago
Yes, they recognize now that WebP was a mistake, which is why they won’t single-handedly force any more image formats upon the web. (Safari learned this lesson earlier when they got stuck being the only browser to support JPEG2000, and Mi
237.
by brigade 4y ago
Also cameras. Which either aren’t interested (no one really cares about non-RAW formats on DSLRs), or are going AVIF once hardware encoders land (Android phones)
238.
by brigade 4y ago
Sure: https://www.nytimes.com/interactive/2018/12/10/business/loca... It's simple - an app asks for background location permissions, then uploads all the datapoints and timestamps the OS gives
239.
by brigade 4y ago
Authorization isn't "I authorize my account to be debited $X", it's "I authorize my account to be debited $X by Party Y" Errors under Reg E include any instance where Party Y is not who receives the funds, incl
240.
by brigade 4y ago
The "good reason" is that it's an extension of how banks have always treated fraudulent debits until you show knowledge of the Reg E dispute process, so it makes it outside of first-level customer support. And it won't b

More ›