Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
bjourne
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
35 ms
·
871.
▲
Mold: A Modern Linker (superfast linker)
(github.com)
3 points
by
bjourne
2y ago
|
0 comments
872.
▲
by
bjourne
2y ago
Then why does ANC keep winning elections in SA?
873.
▲
by
bjourne
2y ago
OpenCL is a pretty decent hpc spec. It's too bad it never reached critical mass.
874.
▲
What does -fheinous-gnu-extensions option do?
(stackoverflow.com)
2 points
by
bjourne
2y ago
|
0 comments
875.
▲
by
bjourne
2y ago
No. My experience is that English is a great language for hiding ignorance behind technobabble and buzzwords. You claim that "conceptual innovation can’t possibly stem from conceptual confusion." It sounds reasonable, but honestly
876.
▲
by
bjourne
2y ago
I would use Python as the extension language since it is omnipresent. ELisp has lots of terrible misfeatures that makes building large software in it difficult. I would also add better support for the services model. E.g., rather than addin
877.
▲
by
bjourne
2y ago
Are bcrypt password hashes difficult to crack? I signed up for IA over 10 years ago with a much weaker password than those I use today.
878.
▲
by
bjourne
2y ago
As others have stated this idea is gratuitous breakage. Hope it won't become reality.
879.
▲
by
bjourne
2y ago
Same with South Korean airline crashes: https://www.salon.com/2013/07/10/cnn_asks_if_koreas_hierarch... But who knows if any of this is true or if "hierarchical cultures cause airline crashes" is ju
880.
▲
by
bjourne
2y ago
Will American judges play ball with this nonsense? Will they realize what absolutely nutty precedents they will set should they find TikTok guilty?
881.
▲
Targeted Intel OneAPI DPC++ Compiler Optimization Rules Out Spec Submissions
(phoronix.com)
1 points
by
bjourne
2y ago
|
0 comments
882.
▲
by
bjourne
2y ago
"The end of Dennard scaling and the impending repeal of Moore's law" is very overused in very many papers. :)
883.
▲
by
bjourne
2y ago
Nah, for plotly. But matplotlib is part of my toolbox for figures that plotly doesn't handle well, such as image grids.
884.
▲
by
bjourne
2y ago
I use matplotlib/seaborn/plotly-express to generate the figure and write a png. That way I can easily refine the figure once it goes from "trivial" to "complex": customize dpi, axis titles and ticks, latex in l
885.
▲
by
bjourne
2y ago
Actually, in the micro benchmark the mutexes were five times slower than the spinlocks. In a real hpc application I reduced the runtime by 30-40% just by replacing them with spinlocks.
886.
▲
SYCL: A Portable Alternative to CUDA
(chsasank.com)
2 points
by
bjourne
2y ago
|
0 comments
887.
▲
by
bjourne
2y ago
WSL can do a lot more than just run Linux binaries.
888.
▲
by
bjourne
2y ago
The binary is only one part of the puzzle (and largely solved by WSL). Installation/uninstallation and desktop integration is just as much of a hassle.
889.
▲
by
bjourne
2y ago
I made a benchmark on this last year when I didn't know how slow pthread mutexes were: https://stackoverflow.com/questions/76965112/why-are-pthread... For my use case, the mutex wait amounted to roughly 40% o
890.
▲
by
bjourne
2y ago
Top sports books do that already. Lesser sports books don't have enough liquidity to react in real-time. In fact a simple winning strategy is to check what odds Pinnacle gives and see if any bookie offers better odds (arbitrage betting
891.
▲
Unix buffering delays output to stdout, ruins your day
(turnkeylinux.org)
3 points
by
bjourne
2y ago
|
0 comments
892.
▲
by
bjourne
2y ago
Read up on disparate impact. It could very well be argued that legacy admission is effectively ethnic discrimination.
893.
▲
by
bjourne
2y ago
> I get annoyed by legacy admissions as much as the next guy, but this strikes me as problematic. An institutions' membership or selection criteria is pretty fundamental to their right to exist. Eh, it's pretty fundamental to K
894.
▲
Who is Marcellus Williams: Execution in Missouri despite evidence of innocence
(innocenceproject.org)
359 points
by
bjourne
2y ago
|
466 comments
895.
▲
by
bjourne
2y ago
This q has been asked so much already. Write an xor network from scratch in numpy. Then write one for mnist. Then cifar10/100. Then one rnn for text gen. Then you know more fundamentals than most practitioners.
896.
▲
Practical Deep Learning for Coders
(course.fast.ai)
2 points
by
bjourne
2y ago
|
0 comments
897.
▲
by
bjourne
2y ago
Java dependencies are too complicated for make. See https://www.oreilly.com/library/view/managing-projects-with/...
898.
▲
by
bjourne
2y ago
My point is that build systems must be able to deal with tools with insanely stupid interfaces like pdflatex. Btw, WAF's strategy of dealing with pdflatex is to rerun the command "until output files stop changing". That'
899.
▲
by
bjourne
2y ago
I guess you aren't keen on Java then? Complex dynamic dependency graphs aren't difficult for humans to handle or many build tools other than make.
900.
▲
by
bjourne
2y ago
I've been using WAF for ages so naturally I wonder how this system compares to WAF? My experience with build systems is that they all get the easy parts rights. You can compile C and C++ code and they successfully scan header files for
More ›