Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
brrrrrm
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
39 ms
·
481.
▲
by
brrrrrm
5y ago
Unrelated to the content, anyone else having trouble reading this article? After three reloads from what seems to be way too many interactive ads, the page crashed entirely. I got about one paragraph in before this: https://post
482.
▲
by
brrrrrm
5y ago
The documentation say this is based on JackTrip, code here: https://github.com/jacktrip/jacktrip
483.
▲
by
brrrrrm
5y ago
A hypergraph can always be viewed as a bipartite graph
484.
▲
by
brrrrrm
5y ago
If you swap out “code” in the url for “raw” you get the raw text itself without a need to use sed https://jott.live/raw/blas_test.cc
485.
▲
by
brrrrrm
5y ago
for those curious about running their own matmul benchmarks, I wrote a script a while back that works with both linux and MacOS that should make comparison easy. https://jott.live/code/blas_test.cc I saw ~1.2tflops on
486.
▲
by
brrrrrm
5y ago
Is Duff’s device still relevant? Compilers can typically do such unroll logic for you these days.
487.
▲
by
brrrrrm
5y ago
The author, Nadav, has a lot of performance gems out there. Another good one explaining how to implement a fast matrix multiply: https://gist.github.com/nadavrot/5b35d44e8ba3dd718e595e40184...
488.
▲
by
brrrrrm
5y ago
Where? Variables are a language semantic not an implementation detail. A temporary variable is just a new symbolic name. You may be thinking of temporary storage
489.
▲
Festivus
(en.wikipedia.org)
4 points
by
brrrrrm
5y ago
|
0 comments
490.
▲
by
brrrrrm
5y ago
Even in "private mode" (not enabled by default), if I search "squid game mask" it hits maps.googleapis.com with my entire user agent string. How is this private? Perhaps a bug?
491.
▲
by
brrrrrm
5y ago
> You don't need the extension or sign up to use the search engine, just type into the search bar on the top of the page. I'm not sure this is true when you're using Chrome. I use safari and thought the same thing at firs
492.
▲
Open source Texas hold'em GTO solver
(github.com)
3 points
by
brrrrrm
5y ago
|
0 comments
493.
▲
Building stuff from source
(jott.live)
53 points
by
brrrrrm
5y ago
|
36 comments
494.
▲
by
brrrrrm
5y ago
Yea I tested it SIMD (8 ops per instruction) and it's hitting 100gflops, so that'd be 12.5k MIPS, right?
495.
▲
by
brrrrrm
5y ago
Ah, it's actually 100gflops (not 100Gfmas) because the fmla instruction is fused. I just ran it on CPU, single core (no pinning), just using the aarch64 fmla instruction. Using pure BLAS (multithreaded, with the MM ISA extension), I ge
496.
▲
by
brrrrrm
5y ago
~100G fp32 FMA/s per core on my M1 air last I checked.
497.
▲
by
brrrrrm
5y ago
“polish is illegal, allowing you to stay in idea-space” I’ve been noticing this design trend in other sites too and I really like it
498.
▲
by
brrrrrm
5y ago
I’ve seen this style before in Python and noticed that it gets kind of dicey once arguments become optional. Suddenly call sites don’t show all the available options and you have to grok the function implementation to get a sense of how to
499.
▲
Tesla shares surpass $1T in value
(nytimes.com)
59 points
by
brrrrrm
5y ago
|
98 comments
500.
▲
by
brrrrrm
5y ago
> if you didn’t need 100 database shards scattered around the globe, were you even doing real work? Is this actually a prevalent mentality at Google?
501.
▲
by
brrrrrm
5y ago
thanks for adding this context! One caveat is that the PyTorch strides used in the post are defined by elements rather than byte-wise.
502.
▲
by
brrrrrm
5y ago
Great question! einsum is almost exactly broadcast + sum. as_strided lets you do “funky” strides to enable all the various convolutions (sliding windows aren’t easily represented with einsum).
503.
▲
Trace, matrix multiplication and every type of convolution in a single line
(twitter.com)
3 points
by
brrrrrm
5y ago
|
0 comments
504.
▲
“as_strided” and “sum” are all you need
(jott.live)
76 points
by
brrrrrm
5y ago
|
10 comments
505.
▲
by
brrrrrm
5y ago
Loop refers to "for loop" (but in a limited context where all control flow and indexing operations are composed strictly of loop variables). The wiki on loop optimization[1] has some good links to literature on the topic. [1] ht
506.
▲
Loop_tool: A toolkit for loop-based computation
(github.com)
15 points
by
brrrrrm
5y ago
|
3 comments
507.
▲
by
brrrrrm
5y ago
a couple more: - at any time while stirring a cup of coffee, there will be a point on the top that is right where it started. (if we pretend coffee stirring is 2-dimensional, Brouwer's Fixed Point theorem) - a drunk man will eventually
508.
▲
by
brrrrrm
5y ago
Instagram uses Python
509.
▲
by
brrrrrm
5y ago
are there really any alternatives though?
510.
▲
Show HN: Pre-flop poker ranges
(ranges.glitch.me)
3 points
by
brrrrrm
5y ago
|
1 comments
More ›