twtw

18 ms·

121.
by twtw 8y ago
1/2 second for this single file, not for the whole project. The bit you quoted about removing includes was about the changes throughout the project.
122.
by twtw 8y ago
I always used to use various dusty old netbooks for this kind of stuff, and would run many applications on the same one. A dedicated raspberry pi in a nice little case for each is certainly modern and cool, but it doesn't give me quite
123.
by twtw 8y ago
Or just use a programming language instead of a calculator with various features bolted on. [Speaking of cvxopt, originally written in matlab] "Fortunately, it has since been rewritten in a real programming language." - Stephen Bo
124.
by twtw 8y ago
No offense intended, but ... that isn't the most glowing recommendation. I read it as "This book is great, really helped me understand the concepts. In fact, it helped so much I failed the class."
125.
by twtw 8y ago
I'm not sure I follow the reasoning in the article. It says if ray tracing flops, AMD will take the lead, but then goes on to say that AMD believes ray tracing will be important and is working on it.
126.
by twtw 8y ago
But the question is could it be made to be good for that domain? See https://arxiv.org/abs/1702.04024 [cited by this paper] for promising results doing linear algebra using aws lambda.
127.
by twtw 8y ago
Is it accurate to say that rank-n tensor is to n-D array of numbers as linear transformation is to matrix?
128.
by twtw 8y ago
I would have appreciated some historical context for the z80 in this post. 8,500 transistors (equiv to ~2800 nand gates in depletion-load nmos logic * ) and not really designed to be a compiler target. I'd say the z80 is pretty impress
129.
by twtw 8y ago
Hasn't this worked via HSAIL for years?
130.
by twtw 8y ago
Gets energy from the power plant to your light bulb without heating up stuff along the way quite as much.
131.
by twtw 8y ago
> 20 year old (!) This paper is close to 20 years old too. The (2018) in the title is wrong.
132.
by twtw 8y ago
Tangential, but the formatting on this is awful. It's like a high school lab report - every equation is a different size and aligned differently.
133.
by twtw 8y ago
This appears to be a GCN backend intended for use with OpenMP and friends, so not really comparable to cuda (more comparable maybe to the llvm nvptx backend).
134.
by twtw 8y ago
It's shown in the 8th image in the rotating set of images at the bottom. Some text in one of the images references Tetsuo Kogawa, so I guess the circuit is from here: https://anarchy.translocal.jp/radio/micro/
135.
by twtw 8y ago
Sure, that's how it is now, but it definitely hasn't always been that way. The original public companies definitely were "early stage, pre-revenue," though you can argue that those were long enough ago to be irrelevant.
136.
by twtw 8y ago
> you can move it off your property And where should I put it? I don't want it on the sidewalk, and I don't want it in mystreet. I can't put it in the trash because it's someone else's property. That's essen
137.
by twtw 8y ago
Is that moving goalposts, or just recognizing another problem? It seems like a problem, so is it a bad thing to solve it once we've (according to you) solved hunger and malnutrition?
138.
by twtw 8y ago
> You can't just steal bikes because they've been left on public land What is the anticipated solution when you leave your bike on my private land? I've seen loads of bird scooters left on private property, and I don&#x27
139.
by twtw 8y ago
> I think the VC model itself is a great development from the last century. The fact that someone can raise millions (more than most people earn in a lifetime!) with an idea enables a lot of innovation that would be hard to nurture other
140.
by twtw 8y ago
> comparing GPU to naive implementations (and sometimes, implementations written in dynamic languages) instead of to highly tuned BLAS or MKL implementations In your experience, are these comparisons using naive GPU implementations as we
141.
by twtw 8y ago
It's looking like that "something cross-platform" will be cuda, given that AMD's latest approach is to implement a compiler to convert cuda to "hip."
142.
by twtw 8y ago
So that's why Xeon Phi has been so successful! /s Real talk, though. Why do you think larabee is the "right" hardware?
143.
The AMD keynote at CES 2019: looking ahead(anandtech.com)
4 points by twtw 8y ago|0 comments
144.
by twtw 8y ago
Addresses.
145.
by twtw 8y ago
It's interesting to see how people react so strongly to this phrase. By default, I assume "boring" is a verb. Maybe it says something about the reader whether they see a verb or an adjective.
146.
by twtw 8y ago
What would "from whole cloth" even mean for this?
147.
by twtw 8y ago
I wonder what world you live in (career, culture, etc) that makes you feel that a post with title "coolest things I learned in 2018" should come with "academic substantiation." Also, the whole point of the words for emot
148.
by twtw 8y ago
Re 2. I like this with the second interpretation mentioned on the page. Avoid boring (v.) other people, rather than avoid people who you seem to be boring (adj).
149.
by twtw 8y ago
Those two comments come up frequently, but it's worth noting that in the whole preceding thread 'cperciva was being a pretty massively arrogant jerk. That thread is a classic example of why the most capable people aren't nece
150.
by twtw 8y ago
The alternatives you mention are things that could happen in the future. The alternative 'p1necone brought up is the status quo.

More ›