tavert

9 ms·

31.
by tavert 9y ago
> if your project is Windows-only, you can get a second compiler’s opinion on your code, and Clang’s warnings might find bugs GCC (mingw.org and mingw-w64 triples) has existed and been able to do this, including and especially cross-comp
32.
by tavert 9y ago
https://github.com/ray-project/ray/issues/631
33.
by tavert 9y ago
Try parquet if your data is tabular, pyarrow and related tools are getting parquet up to a pretty comparable speed to hdf5, with arguably more flexibility and a better multithreading story.
34.
by tavert 9y ago
See the COIN-OR Optimization Services (OS) project for a solver-independent xml format called OSiL. It's functionally equivalent to AMPL .nl format but more human-readable. Unfortunately the osil readers for solvers have more performan
35.
by tavert 9y ago
Have you contributed to Ipopt? Just wondering if I'd recognize you from the mailing list.
36.
by tavert 9y ago
JuMP and Pyomo are both exactly that. Pyomo is as near as you can get to an AMPL clone with python syntax. JuMP is an optimization DSL using Julia macros.
37.
by tavert 9y ago
> I wish there were a lingua franca in this domain and sort of have support from all major solver implementations JuMP is kind of that in terms of modeling software - http://www.juliaopt.org/ Pyomo is also pretty useful
38.
by tavert 9y ago
> does anyone know an updated publication for these? Hans Mittelmann (also a wonderful individual) has you covered: http://plato.la.asu.edu/bench.html
39.
by tavert 9y ago
John Forrest is a hell of a character too if you ever talk to him in person. He's retired to a villa in Italy but still works on Clp, Cbc, and various consulting projects for fun: http://fastercoin.com/f2_files/abo
40.
by tavert 9y ago
Embed-ability, distribution. Using Jupyter from R, Julia, etc generally has an awkward dependency on python for the notebook server.
41.
by tavert 9y ago
> between Mountain View and the SJC airport is a vast isotropic region that’s been rendered devoid of natural landmarks Stealing that description. I'd extend it up the peninsula too, pretty much to Millbrae. Stanford and the bridges
42.
by tavert 9y ago
Also the namesake of a really good bay area metal band https://kowloonwalledcity.bandcamp.com/
43.
by tavert 9y ago
Highly unlikely. You'd only deprecate them if they serve the exact same purpose, but they don't. Many users will switch, but not those who want a posix runtime environment for a win32 executable that can run on machines that don&#
44.
by tavert 9y ago
It'll be more absolute dollar value of growth, but wouldn't be higher in percentage, and taxes are going to come out as some percentage of any money that was put in pre-tax at the time you withdraw. If you didn't work the ent
45.
by tavert 9y ago
You can pay the tax bill separately, which is a better move than letting it come out of the rollover amount if you have cash to afford it.
46.
by tavert 9y ago
Julia marginally makes sense in some academic environments. Outside of that? Not really.
47.
by tavert 9y ago
> It makes it so you always know where the source code for any package is. Where is a useful question to answer, and nice to have consistent across an ecosystem. But What is also a very important question to answer, and tying What to be
48.
by tavert 9y ago
> Berkeley (especially on the incredibly physically-accurate simulator MuJuCo) Huh? MuJoCo was primarily developed by Emo Todorov and his lab at University of Washington, and now his company.
49.
by tavert 9y ago
Have tried for C, C++, Fortran, Julia, and a bit of R. Too many pythonisms (and requiring a fairly heavyweight miniconda setup to run) for it to work naturally when you actively don't want to be using any python. I think the lackluster
50.
by tavert 9y ago
None of that requires conda to be able to do. Can also build in a chroot, without needing an unconventionally configured compiler, if you want to make your binaries redistributable - which is a requirement some users will have but not all.
51.
by tavert 9y ago
Not that big. A few minutes of following the instructions and you can build it in a docker container, post it as a tarball that anyone can use without needing the rest of miniconda to install.
52.
by tavert 9y ago
That sounds like a problem with RHEL, not a problem with btrfs
53.
by tavert 9y ago
Not very quietly, SolarCity as a separate brand has been wound down already, it's marketed as Tesla Energy now.
54.
by tavert 9y ago
What reasons? openSUSE is defaulting to btrfs. snapper is really sweet.
55.
by tavert 9y ago
Short term focused work on Celeste, by at least one of the core devs, came at a very significant opportunity cost of not having an up-to-date, usable Julia debugger.
56.
by tavert 9y ago
Stefan is one of the co-creators of the Julia language, not really an operations research person. He must have misread the results in the JuMP paper where Pyomo outperforms CVX and Yalmip on several problem classes.
57.
by tavert 9y ago
Any that are both open-source and C++11 compliant? Guess you can still build g++-4.6 with just C, then a newer g++ from that, but it's a bit of a pain.
58.
by tavert 9y ago
> What? Numpy does this. And Numpy needs an ad-hoc grafted on type system(s) to do it. > Show me an example of this system code that's nice in julia and clunky in python and I'll show you a superior implementation in python.
59.
by tavert 9y ago
Kind of, though C++ is a difficult language to use from anything other than C++ given there's no standard ABI. Either you need C++ glue-code wrappers around whatever language's FFI you actually want to use, or you embed Clang.
60.
by tavert 9y ago
I frequently wish the Jupyter notebook server were a C library instead of requiring Python. Would make deployment and installation of notebooks for non-Python languages easier. That's kind of the opposite of this though.

More ›