[julia-users] Re: Problems with NA on Array{Any} using square brackets notation

This seems to be a consequence of the promote_rule for NA, as defined in https://github.com/JuliaStats/DataArrays.jl/blob/fc8a8a1f7713c61c76a3562f17fba8c0d193c82b/src/natype.jl#L36-L37 It even has a TODO comment, so it is obvious that the author did not like his solution, but had some reason to

[julia-users] Re: Converting Python to Julia and submitting them to a library

Also worth mentioning, is the Julian convention of adding a bang to the method name if it's *mutating*, i.e. if it changes its arguments, since arguments are passed by reference. Since multiplybias in Johan's example mutates the neuron, it might be nice to name the function multiplybias!instead:

Re: [julia-users] Re: Problems with NA on Array{Any} using square brackets notation

Le mercredi 21 mai 2014 à 00:30 -0700, Ivar Nesje a écrit : > This seems to be a consequence of the promote_rule for NA, as defined > in > > > > https://github.com/JuliaStats/DataArrays.jl/blob/fc8a8a1f7713c61c76a3562f17fba8c0d193c82b/src/natype.jl#L36-L37 > > > > It even has a TODO comment,

Re: [julia-users] Serialize slow for string

Samuel, rewriting in C is almost never necessary, because Julia is as fast as C. It's just that some implementations are faster, and some slower. I don't want to dissuade you in any way from improving the performance of serialize, but you might consider looking at HDF5/JLD. --Tim On Tuesday, M

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

On Wednesday, May 21, 2014 07:33:01 AM Andreas Noack Jensen wrote: >> Vectorized code is always easier to understand > > That is a strong statement. I have vectorised MATLAB code with repmat and > meshgrids that is completely unreadable, but would be fairly easy to follow > if written as loops. I r

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

That's a great feature Andreas! Do you mean like this: a = rand(5) b = zeros(5,5) b += Diagonal(a) Because I get a no method +(Array{Float64,2},Diagonal{Float64}) ... I'm on 0.2 and a fairly old commit right now, so that might be the reason. Julia Version 0.2.0 Commit 05c6461 (2013-11-16 23:44

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Yes, that only works on the 0.3-prerelease versions. Hopefully it won't be too long until a final version is released. Ivar kl. 13:28:46 UTC+2 onsdag 21. mai 2014 skrev Oliver Lylloff følgende: > > That's a great feature Andreas! > > Do you mean like this: > > a = rand(5) > b = zeros(5,5) > b +=

[julia-users] Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

I have a few `BitArray`s representing various conditions, that I'd like to assemble to a mask matrix. Basically, I want to achieve something like this: Rs = linspace(4,8,65) Zs = linspace(-3,3,136)' # yes, transposed inside_R_box = minR .< Rs .< maxR inside_Z_box = minZ .< Zs .< maxZ So far so

Re: [julia-users] Converting Python to Julia and submitting them to a library

Le mardi 20 mai 2014 à 04:41 -0700, ccsv.1056...@gmail.com a écrit : > I have several libraries I wrote in python that I want to convert to > Julia that are for neural networks and exponential smoothing. > > My questions are what do I do about the classes? I have not seen > classes in julia. > >

[julia-users] Re: Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

Doesn't inside_R_box * inside_Z_box gives what you want ?

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Hello colleague, On Wednesday, May 21, 2014 7:33:01 AM UTC+2, Andreas Noack Jensen wrote: > > Please consider b += Diagonal(c) instead of diagm. Diagonal(c) only stores > the diagonal elements but works like diagm(c) for matrix arithmetic > > Vectorized code is always easier to understand > > > T

[julia-users] Re: Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

Indeed it does - thanks! - but I don't understand *why* it works... Is it some funkiness with the way * is implemented for BitArrays, or is it a logical consequence of the usual matrix multiplication rules? Would you care to explain the logic behind it to me? =) // T On Wednesday, May 21, 2014

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

But is M += ((A*B) + (d*C) + (E .* M0)) any slower in Julia than in Matlab? Avoiding the temporaries is a lot less trivial as it might look like and I am not sure that Matlab does this. Am Mittwoch, 21. Mai 2014 14:09:57 UTC+2 schrieb Andreas Lobinger: > > Hello colleague, > > On Wednesday, M

Re: [julia-users] Re: Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

The problem here is that .& isn't defined as an operator, so there's no broadcasting form of & at all. At a guess I'd say that's just an oversight, in which case making a github issue is the right thing. On the other hand, the broadcasting versions of .* etc are all defined on generic AbstractArr

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

I agree that vectorizing is very nice in some circumstances, and it will be great if Julia can get to the point of not introducing so many temporaries. (Like Tobi, I am not convinced this is an easy problem to solve.) But there are plenty of cases where loops are clearer even when vectorization

Re: [julia-users] Re: Why do macros use the @ sign?

Hopefully encouraging less "fancy metaprogramming stuff" means simpler library implementations which actually feel like they're written in the language. It's a real pain when you have to learn a new DSL for every library implementation you want to understand (I say this having read a lot of C++ te

Re: [julia-users] Re: Why do macros use the @ sign?

Yes, exactly. Macros are an escape value that allow you to do special things in a sense beyond what the language supports. That's great when you need it, but it's potentially confusing and often the implementation is less than beautiful and tends to be much less composable than using functions.

[julia-users] Gadfly.draw() gives GLib-GObject-WARNING's and does not draw correctly

Hi! I tried to do this as shown on the gadfly homepage: julia> p = plot(x=[1,2,3], y=[4,5,6]) Plot(...) julia> draw(PNG("myplot.png", 12cm, 6cm), p) which gave me a long list of warnings of the type (process:13547): GLib-GObject-WARNING **: plugin pointer (0x194c33c70) for type 'BasicEngine

[julia-users] Re: Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

You can also use: bitbroadcast(&, inside_R_box, inside_Z_box) to explicitly ask for a broadcasting operation which returns a BitArray. This is exactly what is being done under the hood by inside_R_box .* inside_Z_box, but the bitbroadcast version allows you to be more explicit, and to use othe

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Hello colleague, On Wednesday, May 21, 2014 2:55:34 PM UTC+2, Tim Holy wrote: > > I agree that vectorizing is very nice in some circumstances, and it will > be > great if Julia can get to the point of not introducing so many > temporaries. > (Like Tobi, I am not convinced this is an easy probl

Re: [julia-users] Re: dot function problem

And to MATLAB's credit, it shouldn't transpose in this case either. Note that MATLAB's dot() will do additional argument checking that the transpose-multiplication syntax won't do, so you should test both if speed matters. On Wednesday, May 21, 2014 12:59:50 AM UTC-5, Andreas Noack Jensen wrote

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

On Wednesday, May 21, 2014 8:27:36 AM UTC-5, Andreas Lobinger wrote: > Therefore i'm expecting high-level SIMD language constructs. > Well, there's this... https://github.com/JuliaLang/julia/pull/5355 ...and this work in progress... https://github.com/JuliaLang/julia/pull/6271 ...and if you n

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Julia does have high level SIMD constructs. But again, does Matlab avoid temporaries in vectorized expressions? The fact that for loops are slower than vectorized expressions in Matlab does not mean that the vectorized code is optimal and any better than what we have in Julia. Am Mittwoch, 21.

[julia-users] How do I zip files?

Is there a built-in method for zipping files or must I use an external program? The only hits in the docs were for zipping lists.

[julia-users] Re: Gadfly.draw() gives GLib-GObject-WARNING's and does not draw correctly

Hi Florian, Issues like this are usually due to incompatible combinations of Glib/Cairo/Pango etc. So it really depends on what OS you are running this on, and how you installed Cairo and Pango. If you are on OSX, you may get some ideas from https://github.com/JuliaLang/Homebrew.jl/issues/35

[julia-users] Re: How do I zip files?

https://github.com/fhs/ZipFile.jl On Wednesday, 21 May 2014 14:55:28 UTC+1, RecentConvert wrote: > > Is there a built-in method for zipping files or must I use an external > program? The only hits in the docs were for zipping lists. >

Re: [julia-users] How do I zip files?

Assuming you're referring to compression, the ZipFile.jl package is probably what you want. There's also GZip.jl for (un)compressing single files (in .gz format). (Probably obvious, but zipping lists is an entirely different beast!) Cheers, Kevin On Wednesday, May 21, 2014, RecentConvert wro

Re: [julia-users] Hygienic macros could be both better and simpler

On Tuesday, May 20, 2014 4:53:38 PM UTC-4, Jameson wrote: > > Can you open a pull request? That's the best way to send a diff and have > it evaluated. > Thanks for the direction. I don't know much about Git, but I can probably figure out how to make a pull request. I am more of a Mercurial an

Re: [julia-users] Hygienic macros could be both better and simpler

Slightly OT, but github has an excellent chapter on pull requests in their help section: https://help.github.com/articles/using-pull-requests // T On Wednesday, May 21, 2014 4:30:02 PM UTC+2, David Moon wrote: > > On Tuesday, May 20, 2014 4:53:38 PM UTC-4, Jameson wrote: >> >> Can you open a pul

Re: [julia-users] Re: Gadfly.draw() gives GLib-GObject-WARNING's and does not draw correctly

Hi Avik, thanks for that. not quite sure what I should take from this discussion. Do you think upgrading to mavericks going to solve my problem? On 21 May 2014 15:02, Avik Sengupta wrote: > Hi Florian, > > Issues like this are usually due to incompatible combinations of > Glib/Cairo/Pango etc.

Re: [julia-users] Hygienic macros could be both better and simpler

The learning curve from mercurial should be pretty shallow. Or you could use hg-git. Once you have a branch / fork pushed to github, the pull request is created on the web interface with a few clicks. It's fine to make pull requests for WIP, to gather feedback along the way. Pushing additional comm

Re: [julia-users] Hygienic macros could be both better and simpler

If you want to just make the changes and post a diff, that would be fine too. We can apply the change if it makes sense. On Wed, May 21, 2014 at 11:10 AM, Jameson Nash wrote: > The learning curve from mercurial should be pretty shallow. Or you could > use hg-git. Once you have a branch / fork p

Re: [julia-users] It takes a long time to make a lot of Nothing (Sets are slow)

What version of Julia are you using? In current Julia, Array(Nothing,n) actually takes up no storage and returns instantly no matter how large n is: julia> @time a = Array(Nothing, typemax(Int)) elapsed time: 3.645e-6 seconds (112 bytes allocated) 9223372036854775807-element Array{Nothing,1}: not

Re: [julia-users] Serialize slow for string

Serializing strings was indeed unreasonably slow. I just pushed a commit that should be a significant improvement. On Wed, May 21, 2014 at 7:17 AM, Tim Holy wrote: > Samuel, rewriting in C is almost never necessary, because Julia is as fast as > C. It's just that some implementations are faster,

Re: [julia-users] Optim.jl: unexpected results when using gradient-based methods

Just to extend on what John said, also think that if you can restructure the code to devectorize it and avoid using global variables, you'll see *much* better performance. The way to avoid globals is by using closures, for example: function foo(x, data) ... end ... data_raw = readcsv(file

Re: [julia-users] Re: Why do macros use the @ sign?

What bothers me about the @ is that it is too easy an answer for why if, while, for, try, yieldto, etc. are hard-wired into the language instead of being implemented as macros. Nobody wants to write "@if x==y ; do this @else do that @end"! But the real answer has got to be more interesting th

Re: [julia-users] Re: Why should computer scientists and computational statisticians invest in Julia instead of R?

I am sympathetic to the need of being able to delete vertices or edges from a graph. However, Graphs.jl (and many other packages) is still very young, and it takes some time to provide a complete set of functionalities (especially when one have to make a tradeoff between efficiency, generality,

[julia-users] Re: Accessing the loop variable (and pi as float)

Hi Hans, There was a huge slowdown in your code with : local i::Int. If you change it to i = 0 (therefore i::Int64), the return type is always Int64. Therefore, Julia can generate efficient code. Use @time to check it. This change makes your code 10 time faster on my machine. On Wednesday, M

[julia-users] Element-wise `and` for matrixes of booleans (i.e. BitArrays), and broadcasting

I believe that the intention is for & to be broadcasting, probably just noone has gotten around to it. Please file an issue!

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Le mercredi 21 mai 2014 14:55:34 UTC+2, Tim Holy a écrit : > > I agree that vectorizing is very nice in some circumstances, and it will > be > great if Julia can get to the point of not introducing so many > temporaries. > (Like Tobi, I am not convinced this is an easy problem to solve.) > > B

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Just to add one thing. Julia should definitely not use "fast for loops" as a selling point. If one has access to unboxed values, it's done. No hard work. This is a selling point in comparison to matlab, python and r, but not in general. A very smart automatic devectorizer is much much much hard

[julia-users] Re: Accessing the loop variable (and pi as float)

I don't think this slowness is because of type instability, as I understand the concept. On a 64-bit machine `Int === Int64` is true. Also, the following is also slow and results in lots of allocation: local i::Int64 as does simply declaring local i However, if one completely leaves off the

Re: [julia-users] pathtracer in julia still 3x slower than python

I think that your next big win is going to come from using Immutable Arrays for your matrix / vector ops. One of the first toy projects I did with Julia was port the "business card raytracer" to Julia ( https://github.com/jakebolewski/rays/tree/master/juliarays) and currently it is only 14% slo

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Hello colleague, On Wednesday, May 21, 2014 3:41:53 PM UTC+2, Tobias Knopp wrote: > > Julia does have high level SIMD constructs. > But again, does Matlab avoid temporaries in vectorized expressions? > I dont't know, but my matlab experience (and i belong to the group of profile users and my uni

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

I would be really sad if Julia were to become a language in which " i just write code that expresses what i want to calculate". Declarative languages can be really great (I'm a huge fan of domain-specific systems like JAGS), but they're a real pain to work with if you want to build something out

[julia-users] function isabstract(::DataType)

Is there a way to check if a given DataType is an abstract type? I've come up with a crude method below but thought there should be a better way. function isabstract(t::DataType) try t() catch exception return (typeof(exception) == ErrorException && strin

Re: [julia-users] Re: Why should computer scientists and computational statisticians invest in Julia instead of R?

If an arrant beginner can chime in with some non-technical opinion, permit me to do so here just this once-- The idea of "R vs Julia" is, if I may say so, not the way forward. These languages can and should help each other win. R, as of now, *needs* a speed language, as we all know; an effectiv

Re: [julia-users] function isabstract(::DataType)

A quick search turns up isleaftype, which seems to do exactly the opposite of what you want: isleaftype(AbstractArray) => false isleaftype(Array{Int, 2}) => true Hope that helps. On 21 May 2014 20:15, Stephen Chisholm wrote: > Is there a way to check if a given DataType is an abstract type?

Re: [julia-users] function isabstract(::DataType)

That's it. Thanks! On Wednesday, 21 May 2014 16:21:48 UTC-3, Mike Innes wrote: > > A quick search turns up isleaftype, which seems to do exactly the > opposite of what you want: > > isleaftype(AbstractArray) => false > isleaftype(Array{Int, 2}) => true > > Hope that helps. > > > On 21 May 2014 2

[julia-users] Re: Accessing the loop variable (and pi as float)

Good catch. I have no idea the reason behind this behaviour. It would be interesting to know why. On Wednesday, May 21, 2014 7:58:02 PM UTC+2, Peter Simon wrote: > > I don't think this slowness is because of type instability, as I > understand the concept. On a 64-bit machine `Int === Int64` is

[julia-users] Re: suggestion of OSX julia/ijulia setup

I'm on a 2012 Macbook Air and I run julia v0.3 nightly builds. Building Julia myself on this machine takes forever, so I just download the nightly build from http://status.julialang.org/download/osx10.7+ about once a week. I then drag it into /Applications and rename it from its commit hash ver

Re: [julia-users] Re: Why should computer scientists and computational statisticians invest in Julia instead of R?

I know that people really hate this approach, but can't you just write files to disk and then do I/O to get data from Julia into R? I've done a lot of that over the last two years. -- John On May 21, 2014, at 12:18 PM, Travis Porco wrote: > If an arrant beginner can chime in with some non-te

[julia-users] Shameless plug for DotPlot.jl - plots in your terminal

I saw Drawille for python on Hacker News yesterday, and realized I'd always wanted to do something just like that. Julia is capable of supporting a much more elegant syntax for plotting than the python or javascript versions of Drawille, so I went ahead an

Re: [julia-users] Shameless plug for DotPlot.jl - plots in your terminal

Hah! I saw that too and thought about the same thing. This is really cool. On Wed, May 21, 2014 at 4:51 PM, Adam Smith wrote: > I saw Drawille for python on > Hacker News yesterday, and realized I'd always wanted to do something just > like that. Julia is c

[julia-users] Re: Call julia.bat under Windows 8.1

Hi Tobias. Sorry for my late reply! I installed the prerealse 0.3 but it dosen't work! I try to use julia under Windows 7 and will see what happends...! :-) Regards, Günter Am Sonntag, 18. Mai 2014 10:56:44 UTC+2 schrieb Günter Faes: > > Dear Community, > > Sorry if this question was already as

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

Reading your posts I get the impression that Matlab outperforms Julia in vectorized code. I think it would be really great if you could give some numbers to let us know how far Julia lags behind. It seems they have an outstanding JIT if it can transform vectorized expressions into optimum code

Re: [julia-users] Releasing SharedArray Memory

Hi Tim, thanks for answering! I don't do any explicit passing, all memory management is done through a pmap(data, x->myFunction(x,capitals)) I was able to solve(?) the problem (has been running for a while, way longer than the time it took to crash before) by calling instead @everywhere capit

Re: [julia-users] Re: suggestion of OSX julia/ijulia setup

On newer OSes (10.9 for example), I rarely have to do a full compile / build with Julia. The first time takes a while for all the deps, but over 9/10 times all that is needed is: - git pull - make clean - make Sometimes a "make cleanall" is required, which would be slow but was also rare (when

[julia-users] How do I get the reduced start up time for Julia 0.3-prerelease when installing from a binary package?

I was excited to learn that Julia 0.3 will have a much smaller startup time than 0.2 does. If I understood correctly, the reason Julia was slow to start is that it compiles a large portion of the standard library upon starting and the fix was to precompile the library. I installed the Julia 0.3

Re: [julia-users] Re: how to apply vector on diagonal of matrix ?

I suspect they still generate temporaries, but probably have a better garbage collector and can re-use the same temporary across iterations. #5227 should take care of the first, and Jeff has posted his hopes that the second will also be implemented (someday). I suspect both of those would be eas

[julia-users] Gadfly: plotting histogram of a integer variable x

Hi guys, I'm having a problem when plotting data like this: data = int(round(dropna(outcome[:,11]))) p = plot(x=data, Geom.histogram) Here, data is an Array{Int64,1}. But the plot I get after running this have the bars spread, ignoring the space between integers. Is this expected? How can I mak

[julia-users] Re: JuliaCon Question Thread

Will the videos of JuliaCon be uploaded to JuliaLanguage channel on Youtube? I think we should stick with ir, and keep posting interesting thing there, so people have where to look for good quality videos about the language.

Re: [julia-users] It takes a long time to make a lot of Nothing (Sets are slow)

I was using the prerelease 3.0 for OSX I'll look up the date I downloaded it later, I'm at a different machine now. I cannot compile my own because I have the incorrect Fortran. When profiling the Graphs.js maximal_cliques function (with a slightly improved Set to avoid the copy and delete for

Re: [julia-users] Re: Accessing the loop variable (and pi as float)

I think this is because the loop might have zero iterations, and in that case `i` might not be initialized. Variables that might be used uninitialized get pessimized. We could improve this by representing such variables internally as a pair of a fast variable and a boolean flag. On May 21, 2014 3:4

Re: [julia-users] How do I get the reduced start up time for Julia 0.3-prerelease when installing from a binary package?

1. The newest 0.3 binaries are supposed to contain the sys.dylib file (it's sys.so on linux, and so I have started to just call it that everywhere for simplicity) and thereby gain the accelerated startup time. I'm not sure why this would be failing for you. 2. You can write a base/userimg.jl file,

Re: [julia-users] Re: Accessing the loop variable (and pi as float)

Added issue #6914 to describe this performance issue. --Pete On Wednesday, May 21, 2014 6:48:31 PM UTC-7, Jeff Bezanson wrote: > > I think this is because the loop might have zero iterations, and in that > case `i` might not be initialized. Variables that might be used > uninitialized get pessi

Re: [julia-users] Shameless plug for DotPlot.jl - plots in your terminal

Would it make sense to merge this functionality into ASCIIPlots? To me that seems like a better name, and John Myles White is likely to be willing to transfer the repository if you want to be the maintainer. That package started from code posted on the mailing list, and the author thought it was

Re: [julia-users] How do I get the reduced start up time for Julia 0.3-prerelease when installing from a binary package?

Regarding #2: The Ubuntu nightlies may not have sys.so, since I don't think JULIA_CPU_TARGET was being set. This has been fixed, and the faster startup should start showing up in about 24 hours, when the n