Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
iskander
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
50 ms
·
481.
▲
by
iskander
13y ago
Is Bryan still working on Copperhead? Also, do you know if the DPH folks ever managed to iron out a version of higher order flattening which gives a predictable performance gain?
482.
▲
by
iskander
13y ago
PyCUDA lets you define new kernels by plopping CUDA source as a triple-quoted string in the middle of your Python code. It's not the most elegant thing in the world but you quickly get used to it. It seems on the surface that Continuum
483.
▲
by
iskander
13y ago
Vasily's approach to CUDA really revolutionized how I think about GPU programming and I'm glad the continuum folks are giving ILP on the GPU a broader audience. Can anyone testify to the quality of continuum's CUDA wrapper? I
484.
▲
by
iskander
13y ago
Is this a novelty account? How long has news.yc had these? Is there a policy of quickly banning these things? It certainly doesn't help the quality of discourse here.
485.
▲
by
iskander
13y ago
>cython, weave (C++ inline), numexpr (JIT for a limited set of NumPy functions), direct access to the BLAS (provided via a SciPy module), swig, f2py,.... I agree with your comment, I just want to add a few more Python performance options
486.
▲
by
iskander
13y ago
> weather is nice People keep saying this all over the thread but it rains so damn much in Amsterdam! The Dutch don't seem to mind, throw on some rain gear and bike 10km to work. Almost no one does that in the US, I think it's
487.
▲
by
iskander
13y ago
But they have awesome bike lanes! See: http://www.newyorker.com/online/blogs/evanosnos/2011/03/the-... The only trouble is that the lanes are often under-utilized, encouraging all sorts of other peo
488.
▲
by
iskander
13y ago
You probably know this, but for the sake of HN: Beijing used to be a lot more like Amsterdam. This wasn't a choice on the part of the population: cars were simply unavailable and affordable. You still have some pretty sweet separated
489.
▲
by
iskander
13y ago
>every teen here rides mountain bike made of carbon fibre Every teen in Belgium can afford a $2000 bike?
490.
▲
by
iskander
13y ago
From my experience in Amsterdam (lived there for two summers), there aren't two disjoint sets: cyclists and motorists. Everyone has a bike and many people have a car. It's usually more pleasant to bike somewhere.
491.
▲
by
iskander
13y ago
>After several weeks on the job...Bob had a TopSecret clearance. Top Secret clearance usually takes a few months. I'm comfortable with a world in which someone who worked for an NSA subcontractor for several months being stuck with
492.
▲
by
iskander
13y ago
Similarly, Siu from Continuum wrote some nice Python bindings ( http://www.llvmpy.org/ ) and LLVM already ships with OCaml bindings ( http://llvm.org/docs/tutorial/OCamlLangImpl1.html )
493.
▲
by
iskander
13y ago
>I'm going to greatly simplify things down for you with this example (so experts out there... don't shoot me :-p). A GPU can execute 1 program, but that one program can do 512 operations at once. CUDA & OpenCL programs can
494.
▲
by
iskander
13y ago
>last 65,000 years...62,000 years before the ancient Egyptians built the pyramids at Giza. A very minor nitpick, but the Giza pyramids were built around 2500BC, not 1000BC.
495.
▲
by
iskander
13y ago
This was one of the many projects that got gutted by Plus, I'm surprised that it's still being hosted at all.
496.
▲
by
iskander
13y ago
"Array of structs" -> "Struct of arrays"? I guess if you generalize this transformation sufficiently you end up doing whole-program flattening like NESL and DPH. Or is there a less intrusive way to pull that off?
497.
▲
by
iskander
13y ago
>The U.S. has stripped persons of their citizenship for things as mundane as fraud, so this is hardly a made-up case for Snowden. For naturalization fraud. That is, if the original basis for citizenship was found to be invalid. Aside f
498.
▲
by
iskander
13y ago
That's good news! Did that change with Kepler? Any links to more info?
499.
▲
by
iskander
13y ago
I wish they'd at least give us a computed goto...but alas, every switch is destined for predicated branching.
500.
▲
by
iskander
13y ago
>In addition, the reason it's hard to express structured data and nonuniform computation is because those aren't things GPU architecture excels at. If you write simple GPU kernels (like those that Theano/Thrust/Copper
501.
▲
by
iskander
13y ago
How do you support higher order procedures? Do you do control flow analysis & defunctionalization or are these "real" higher order functions?
502.
▲
by
iskander
13y ago
>Harlan already has native support for rich data structures, including trees and ragged arrays. Very soon the language will support higher order procedures. Structured data and nonuniform computations are very hard to express right now o
503.
▲
by
iskander
13y ago
Sure, it's alex dot rubinsteyn atop google's excellent email service.
504.
▲
by
iskander
13y ago
You're probably ready to write your own JIT -- I'd recommend making one for Python ;-) I think after the low-hanging fruit above, there's lot of weird interpreter/compiler folklore on old usenet posts, Forth VM designs,
505.
▲
by
iskander
13y ago
Language implementation sometimes gets discussed up on lambda-the-ultimate . Also, searching for anything and everything Mike Pall has written about VM design is probably worthwhile. Mozilla developers also have some pretty interesting blog
506.
▲
by
iskander
13y ago
The ponderousness of these data structures isn't just about memory consumption or having to use boxed numbers. As far as performance goes, PyObjects infect everything in the interpreter. For example, when you're calling a Python f
507.
▲
by
iskander
13y ago
>Rearrange your instruction bodies based on execution / transition frequencies to improve instruction cache performance. Do you mean...group all the frequent operations together so they overlap on cache lines? It's hard to tell
508.
▲
by
iskander
13y ago
>It may be theoretically true, but in practice, as nice as Python may be to use, it has proved a very difficult language to speed up. I disagree with you. Python isn't much harder to speed up than Lua and in some ways it's bett
509.
▲
How fast can we make interpreted Python?
(phi-node.com)
167 points
by
iskander
13y ago
|
80 comments
510.
▲
by
iskander
13y ago
I worked at a school with a high percentage of foreign students on F-1 visas. The visa-granting process is much more arbitrary than you imagine. The supposed "strict procedure" is actually filtered through many layers of several p
More ›