pletnes

71 ms·

841.
by pletnes 9y ago
The most impressive is how easy it is to use. Basically just reuse of the numpy API.
842.
by pletnes 9y ago
Same story for me, give or take a few percent. I have to recommend dask for python though, it made out of memory errors largely disappear for me. It allows parallel processing of disk-size datasets with the convenience of in-memory datasets
843.
by pletnes 9y ago
One common misconception (or should I say, overgeneralization) is repeated in the article: threads are always unsuited to CPU intensive work. For instance, most numpy operations release the GIL, meaning that you can perform heavy computatio
844.
by pletnes 9y ago
Looks to me like yet another reason to use sqlite instead of «flat files». E.g postgres seems like a many orders of magnitude increase in complexity. Possibly related: We’ve been using hdf5 for a lot of data storage at work (raw image data)
845.
by pletnes 9y ago
I guess they want you to run your analyses on AWS directly.
846.
by pletnes 9y ago
Technically speaking, you don’t need to know the frequency response in detail, or whether it’s exactly linear. If the glass «sings» at a given frequency after being whacked by a short impulse, that’s a resonant frequency.
847.
by pletnes 9y ago
Absolutely, the people are the problem. Those are clearly not Harley bikes though.
848.
by pletnes 9y ago
This is so true. I’ll be happy to see Harley bankrupt and biking forbidden. Not to mention the insane velocities and traffic hazard the bikers pose. Source: have lived near a road.
849.
by pletnes 9y ago
It’ll be bundled with other atlassian products. Less hassle to integrate with those, one less service to log into, one less monthly payment, etc. Slack is possibly better for someone who doesn’t use atlassian products. Also, many people use
850.
by pletnes 9y ago
You might want to try sympy or sage as a mathematica replacement.
851.
by pletnes 9y ago
How does that work? Download DB file, read/write, upload DB file (if written)? What about concurrency?
852.
by pletnes 9y ago
The W series are workstations. Solid, fast, heavy, expensive.
853.
by pletnes 9y ago
Repairability is a good thing. That being said, the 5 year old macbook pro I replaced had way better battery life than the then-new HP 840 G3 I am now using day to day.
854.
by pletnes 9y ago
There’s dask.array which works on numpy arrays instead of dataframes. Otherwise, your argument holds.
855.
by pletnes 9y ago
That’s easy with python, too, in a lot of number crunching cases. Numpy with MKL will use all your cores, as will e.g dask and other libraries built on numpy. Farming out embarassingly parallel work to threads or processes is also easy.
856.
by pletnes 9y ago
Sure, I’ve seen C++ experts beat by numpy, i.e BLAS/LAPACK. A lot of people don’t use BLAS from C++ either, which would improve the performance on that side, too.
857.
by pletnes 9y ago
The author is a scientist analyzing his data. I never met anyone in that crowd using C#. Are there even any good data science/numerics libs out there? C++ has a lot of number crunching libs, python even more.
858.
by pletnes 9y ago
There’s a project to plug different JIT compilers into CPython, so there’s hope. https://github.com/Microsoft/Pyjion/blob/master/README.md Also, I’ve more than once seen cpython beat C++/Fortran sin
859.
by pletnes 9y ago
If you’re doing data science, do consider the anaconda python distribution with conda environments instead. Intel MKL linked numpy gives a lot better performance, and conda can install binary dependencies for libs that are tricky to compile
860.
by pletnes 9y ago
Several posters indicate that they’ve stuck to python 2.7 even for small side projects until now. I cannot understand why? Python 3 seems to have been technically superior for a few years, and side projects must surely be good for learning
861.
by pletnes 9y ago
The @ operator is the same as np.matmul which is different from np.dot for matrices of rank >=3.
862.
by pletnes 9y ago
That’s true, but I can mention several online stores that do this. Even on a good network the site remains blank for seconds at times. If they at least served me a logo I’d know the site wasn’t dead!
863.
by pletnes 9y ago
It’s useful for certain theoretical proofs in analytic math - or so I’ve heard. Not all math is well suited to be coded into reliable software.
864.
by pletnes 9y ago
Yes, Cramer’s rule (referred to in the article) does not give numerically stable results AFAIK - which is yet another can of worms. Good catch.
865.
by pletnes 9y ago
Also, it seems the author solved a triangular system. LAPACK has special routines for that. Were they used? LAPACK is typically optimized for larger systems, not 5 unknowns. Also, 5 is not a great number for vectorized operations - it might
866.
by pletnes 9y ago
The same can be said for gcc - basically, optimization heuristics are sometimes incorrect.
867.
by pletnes 9y ago
They’ll be sending fake signals, i.e time shifted. Since the speed of light is enormous, we’re probably talking micro- to millisecond disturbances. No worries for regular NTP servers I’d imagine.
868.
by pletnes 9y ago
Not for us vi/vim guys.
869.
by pletnes 9y ago
Tmux is awesome. Although I never understood why they used ctrl-b instead of ctrl-a as the default though - incompatible with screen and hard to reach with one hand.
870.
by pletnes 9y ago
Python does: foo in bar

More ›