HNHacker News
TopNewBestAskShowJobs

TinkersW

680 karma · joined May 19, 2020

submissionscomments
TinkersW··on Rust inadequate for text compression codecs?
If you are quoting the numbers google put out, recall that those numbers were after they profiled and manually disabled bounds checks in cases where it caused significant slowdown.. so in other words the numbers were garbage.
TinkersW··on Zen 5's AVX-512 Frequency Behavior
Nvidia calls them cores to deliberately confuse people, and make it appear vastly more powerful than it really is. What they are in reality is SIMD lanes.

So the H100(which costs vastly more than a Zen5..), has 14592 32 bit SIMD lanes, not cores.

A Zen 5 has 16x4(64) 32 bit SIMD lanes per core, so scale that by core count to get your answer. A higher end desktop Zen5 will have 16 cores, so 64x16 = 1024. The Zen5 also clocks much higher than the GPU, so you can also scale it up by perhaps 1.5-2x

While this is obviously less than the H100, the Zen5 chip costs $550 and the H100 cost $40k.

There is more to it than this, GPUs also have transcendental functions, texture sampling, and 16 bit ops(which are lacking in CPUs). While CPUs are much more flexible, and have powerful byte & integer manipulation instructions, along with full speed 64 bit integer/double support.

TinkersW··on Zen 5's AVX-512 Frequency Behavior
If you normalize for cost, the difference isn't nearly as large as you might think, around 4-5x.
TinkersW··on Zelensky leaves White House after angry meeting
I don't think that is why he won.

Dems basically handed this election to Trump; Biden running when he had memory issues, then handing it to Harris whose numbers were questionable against Trump(with the braindead excuse that only she could get the funds), along with generally not responding to issues voters obviously care about(border), trying to ignore issues like excessive wokeness/DEI instead of admitting it is a problem.

Trump is going to lie and obviously only cares about himself, some % of rubes will vote for that, but you don't have to make it so damn easy for him.

TinkersW··on Zelensky leaves White House after angry meeting
Russia already had, and even created more NATO membership on its borders(Finland just recently). That excuse is clearly false.
TinkersW··on Welcome to Ladybird, a truly independent web browser
I scanned through it, and it is actually pretty easy to read C++. Though they could use to learn what comments are.
TinkersW··on Teslas monitor everything – including you [video] from WIRED
No, but it is an argument for making it illegal for car manufacturers to monitor you.
TinkersW··on Zlib-Rs Is Not Only Safer but Now Outperforming Zlib C Implementations
zlib is really old and out of date, no real reason to use it except legacy. I think most optimization effort goes into newer compression libs like zstd.
TinkersW··on Mere weeks after Starship's breakup, the vehicle may soon fly again
It isn't about obligations, it is about the best interest of the US. Which is very clearly not handing everything to Russia for no good reason, and for nothing at all in return.
TinkersW··on Mere weeks after Starship's breakup, the vehicle may soon fly again
The second you say something like that first sentence you have already lost the argument.

Which is annoying as the US is going to face some serious consequences from Trumps idiocy, phrasing it the way you did, is handing the future to Trump and his kind.

TinkersW··on FFmpeg School of Assembly Language
It depends, for SIMD float-> scalar floats it is fast as they operate on the same registers. If pulling out of lane 0 you don't even need to do anything(just a type cast). For other lanes you need a shuffle.

For SIMD integer to scalar integer, it has to move into separate register, so there is some short penalty(3 cycles iir).

TinkersW··on FFmpeg School of Assembly Language
They said they prefer intrinsics which the article says are only about 10% slower(citation needed), you misunderstood and made a comparison against scalar.

Personally I'd say the only good reason to use assembly over intrinsics is having control over calling convention, for example the windows CC is absolute trash and wastes many SIMD registers.

TinkersW··on Greg K-H: "Writing new code in Rust is a win for all of us"
Ya it rather baffling, it would be a solid improvement, and they can easily ban the parts that don't work for them(exceptions/stl/self indulgent template wankery).

On a memory safety scale I'd put C++ about 80% of the way from C to Rust.

TinkersW··on Greg K-H: "Writing new code in Rust is a win for all of us"
It is unfortunate that it is written in such a unhinged way as there are probably some valid points mixed in with the insanity..
TinkersW··on Broken legs and ankles heal better if you walk on them within weeks
I smashed up my wrist in a bike crash, resting it didn't fix shit. When I lifted weights it would tweak out and hurt, so I avoided that for like a year.

Eventually I decided fuck this, and starting lifting weights even though it hurt, then I'd massage it which somehow reduced the pain. Repeat that, and very quickly the wrist fixed itself, zero issues now.

When I broke my collarbone years ago, the docs said just let it heal on its own, which it did just fine(I was hiking/running about 2 hours every day, maybe that helped).

TinkersW··on Google defends scrapping AI pledges and DEI goals in all-staff meeting
California banned affirmative action in 1996 precisely because it is discriminatory.

I can only hope the dems are smart enough to drop DEI for good, I don't think Trump is healthy for the US.

TinkersW··on OpenAI scrubs diversity commitment web page from its site
Your justification for discrimination against white guys is that a very tiny percent of them are legacy admissions and/or from wealthy families?

You know the rest of them are not benefiting from that in the slightest? Actually they are probably hurt by it under the current DIE platforms.

I agree that it would be great if hiring wasn't so focused on a small group of schools(ie excuses to hire people just like yourself).

If DIE is replaced it needs to be with a system based on family income not skin color etc.

TinkersW··on Resigning as Asahi Linux project lead
You just proved the C devs point..
TinkersW··on Don't "optimize" conditional moves in shaders with mix()+step()
I really only care about performance in a SIMD sense, as not using SIMD means you aren't targeting performance to begin with(GPU == SIMD, CPU SSE/AVX == SIMD)

Branches fall off vs cmov/select as your lane count increases, they are still useful but only when you are certain the probability all lanes agreeing is reasonably high.

TinkersW··on Don't "optimize" conditional moves in shaders with mix()+step()
A real branch is useful if you can realistically skip a bunch of work, but this requires all the lanes to agree, on a GPU that means 32 to 64 lanes need to all agree, also for something basic like a few arithmetic ops there is no point.
TinkersW··on Don't "optimize" conditional moves in shaders with mix()+step()
It is weird how long misinformation like this sticks around, the conditional move/select approach has been superior for decades on both CPU & GPU, but somehow some people still write the other approach as an "optimization".
TinkersW··on 21st Century C++
If you already used C++20 you aren't meaningfully behind, very little of interest has been introduced since then, and much of it isn't usable yet because of implementation issues.
TinkersW··on 21st Century C++
Just ban ranges lib, it is hot garbage anyway. The compilers are able to optimize lambdas fairly well nowadays(when inlined), I wouldn't be that concerned.
TinkersW··on 21st Century C++
Ya that is rather disingenuous, modules aren't ready, and likely won't be for another 5 years.

Also they are difficult to switch to, so I would expect very few established projects to bother.

TinkersW··on State Farm seeks 22% emergency rate hike
What is annoying is that this effects everyone in CA, even if you don't live in a fire prone area. They make you pay for the idiots living in these areas.
TinkersW··on Radiant Foam: Real-Time Differentiable Ray Tracing
Em? Seems reasonable, ray tracing a BVH is inherently slow AF because it diverges like mad(SIMD no like).
TinkersW··on Boom XB-1 First Supersonic Flight [video]
No chance, far too loud, and I doubt Boom will get approval either
TinkersW··on C is not suited to SIMD (2019)
The article seems to conflate C and C++ in its very first sentence, though in reality these are very different languages. In C++ you can use std::array, and you can make your SIMD containers use restrict to make it more like Fortran.

Anyway auto-vectorization will never compete with intrinsics for performance, so this all seems rather silly. Try expressing pshub in auto-vectorizer. Or a bunch of optimal load/swizzles to out perform gather.

TinkersW··on C is not suited to SIMD (2019)
This reads like jibberish.

C functions can't be vectorized? WTF are you talking about? You can certainly pass vector registers to functions.

Exp can also be vectorized, AVX512 even includes specific instructions to make it easier.( there is no direct exp instruction on most hardware,it is generally a sequence of instructions)

TinkersW··on DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL
AI sure, which is good, as I'd rather not have giant companies in the US monopolizing it. If they open source it and undercut OpenAI etc all the better

GPU: nope, that would take much longer, Nvidia/ASML/TSMC is too far ahead

← PreviousPage 4 of 13Next →