The shmem stuff in PTools.jl is quite old. Do check out the current master
which has support for SharedArrays - see
http://docs.julialang.org/en/latest/manual/parallel-computing/#shared-arrays-experimental-unix-only-feature-
which is definitely more usable.
On Thu, Feb 13, 2014 at 5:00 AM, Jim
Thanks to all of you.
Thank you Iain, Elliot and Eric. :D
I am now starting with Julia. The use of MPI is natural in Julia?
2014-02-12 16:46 GMT-03:00 Eric Davies iam...@gmail.com:
It looks like there's a reason for that, and it relates to the different
way cuBLAS handles memory. See more
You're not doing anything wrong at all – this is a parser bug. Would you be
willing to file an issue on github?
https://github.com/JuliaLang/julia/issues/new
On Thu, Feb 13, 2014 at 9:50 AM, Laszlo Hars laszloh...@gmail.com wrote:
I wanted (in a macro) to print out quoted string literals, like
When I'm learning a new language one of the programs I write is a simple
sieve of Eratosthenes; the julia code is shown below.
Julia is taking 70% longer than python 2.7.5 to execute the program - have
I coded something incorrectly?
Language Execution time (seconds) C 0.106 pypy 0.323
Did you code this up in Python too? There's a built-in Julia function
called primes (written in pure
Juliahttps://github.com/JuliaLang/julia/blob/master/base/primes.jl),
which implements a prime number sieve efficiently:
julia @time primes(1000);
elapsed time: 0.167461201 seconds (6581064
On Thursday, February 13, 2014 10:34:32 AM UTC-5, Stefan Karpinski wrote:
Did you code this up in Python too? There's a built-in Julia function
called primes (written in pure
Juliahttps://github.com/JuliaLang/julia/blob/master/base/primes.jl),
which implements a prime number sieve
Thanks for the quick reply. Yes these were all hand coded in Julia, python,
and C - same algorithm.
Version 0.1.2
02f3159-Linux-amd64 (2013-05-15 23:25:17)
I was timing from the shell, my function is much faster timed from within
julia:
julia @time myprimes(1000)
elapsed time:
Hmm. I guess there's probably a fair amount of overhead from all that
BitArray twiddling. I wonder if we should switch to using a full array of
integers in the algorithm for smaller N. The BitArray thing is mostly good
for large N where the list of primes is going to be much smaller than the
list
Matplotlib (PyPlot) seems pretty fast even for 10^5 points.
Hi all,
I've been suffering from a vague BoundsError() which turned out quite
hard to debug because of a lack of information. This is (a reduced
version of) the code which tripped the issue:
using DataFrames
using Gadfly
function main(args)
df = DataFrame(x = [1, 1], y = [0, 1])
On Thursday, February 13, 2014 10:57:20 AM UTC-5, Steven G. Johnson wrote:
for i in 1:sieveTo
Note that you can speed it up a bit more (by ~0.01s on my machine) by
changing this to
@inbounds for i in 1:sieveTo
to turn off bounds-checking in the loop (since all the array indices
That actually doesn't help at all on my system. We ought to be inlining the
array accesses, in which case LLVM can probably eliminate the bounds checks
automatically.
On Thu, Feb 13, 2014 at 11:01 AM, Steven G. Johnson
stevenj@gmail.comwrote:
On Thursday, February 13, 2014 10:57:20 AM
On Wednesday, February 12, 2014 3:42:15 PM UTC-6, Taylor Maxwell wrote:
You should be able to perform an anova or ancova with the GLM package.
https://github.com/JuliaStats/GLM.jl
You can use the lm() function and your variables with factors need to be
PooledDataArrays in the dataframe.
I believe if there is *any* code in a type block other than field
declarations, the type will not have any inner constructors. Not really a
bug, but maybe somewhat confusing.
On Wed, Feb 12, 2014 at 10:07 PM, Jake Bolewski jakebolew...@gmail.comwrote:
OK, makes sense. A bit strange that it
Can anyone explain why these produce two different types?
julia a=[a=1,b=2]
[b=2,a=1]
julia typeof([i for i=a])
*Array{Any,1}*
julia typeof([i for i=[a=1,b=2]])
*Array{(ASCIIString,Int64),1}*
Le jeudi 13 février 2014 à 14:09 -0800, Fil Mackay a écrit :
Can anyone explain why these produce two different types?
julia a=[a=1,b=2]
[b=2,a=1]
julia typeof([i for i=a])
Array{Any,1}
julia typeof([i for i=[a=1,b=2]])
Array{(ASCIIString,Int64),1}
I think you may be hitting
https://github.com/JuliaLang/julia/issues/3469
Presumably this will be fixed at around the same time the debugger lands.
--Tim
On Thursday, February 13, 2014 04:52:24 PM Tim Besard wrote:
Hi all,
I've been suffering from a vague BoundsError() which turned out quite
hard to debug because of
Thanks - makes sense, not a major issue as this wont happen with locals in
the real world..
On Fri, Feb 14, 2014 at 9:19 AM, Milan Bouchet-Valat nalimi...@club.frwrote:
Le jeudi 13 février 2014 à 14:09 -0800, Fil Mackay a écrit :
Can anyone explain why these produce two different types?
Thanks for all the advice, everyone. I've just finished a parallel sparse
matrix vector multiplication library written in straight julia for shared
memory machines, using Amit Murthy's SharedArrays. You can find it at
https://github.com/madeleineudell/ParallelSparseMatMul.jl
For a matrix with 1E9
Nice!
On Thu, Feb 13, 2014 at 6:49 PM, Madeleine Udell
madeleine.ud...@gmail.comwrote:
Thanks for all the advice, everyone. I've just finished a parallel sparse
matrix vector multiplication library written in straight julia for shared
memory machines, using Amit Murthy's SharedArrays. You
This was submitted as an issue on
github: https://github.com/JuliaLang/julia/issues/5800
The problem was that the 3 first ending quotes is interpreted as the end of
the literal, and that leaves a single double quote that causes parser
errors later on.
Ivar
kl. 16:24:18 UTC+1 torsdag 13.
Very impressive! It's pretty cool to have native-Julia performance matching
MKL! And linear performance up to 40 threads...not too shabby...
I was actually hoping that sparse matvec multiplication would be one of the
first major uses of ShareArrays, and you've just done it. This will certainly
Hey there, I'm trying to use Gadfly's Geom.binrect to plot a matrix, but I
can't figure out how to do it without going through a lot of rigamarole to
generate a DataFrame like is used in the
examplehttps://github.com/dcjones/Gadfly.jl/blob/master/doc/geom_rectbin.md
docs.
I have, say, a 10x10
There's actually a special function spy to make plotting matrices
simpler, where spy(M) returns a plot. All that function does is basically
call findnz on the matrix and pass the result to x, y, and color in the
regular plot function.
Special handling of matrix arguments is something to
Great, spy() is exactly what I wanted! Is it documented anywhere, or did I
just miss it?
-E
On Thu, Feb 13, 2014 at 4:47 PM, Daniel Jones danielcjo...@gmail.comwrote:
There's actually a special function spy to make plotting matrices
simpler, where spy(M) returns a plot. All that function
I'm looking for a Lyapunov/Sylvester solver for
AXB + CXD = F
There is one in SDE.jl, but it just reduces to kronecker product for the
full problem. I'm pretty sure that LAPack has a solver built-in. Is it
easy to access that?
Sheehan
Hmm, that picture didn't seem to embed nicely. Here's another shot at it.
On Thu, Feb 13, 2014 at 5:08 PM, Elliot Saba staticfl...@gmail.com wrote:
Also, is there a way to get the axes to line up a little better? I want
to tell spy() that it should start a 1 and not 0, as it seems to want
Thanks, Jiahao! It looks like you've already made a great dent in the
iterative solvers wishlist. I'm planning on using the
ParallelSparseMatMul library along with some iterative solvers
(possibly just LSQR) to implement iterative solvers for nonnegative
least squares, lasso, elastic net, etc
I just tagged a new version that should fix the scales and let spy take
arguments like plot (so you can change the axis labels, etc).
You didn't miss spy in the documentation, it's not there, though it should
be.
On Thursday, February 13, 2014 5:08:52 PM UTC-8, Elliot Saba wrote:
Hmm, that
OK I filed an issue.
Cheers,
Sheehan
On 14 Feb 2014, at 12:12 pm, Jiahao Chen jia...@mit.edu wrote:
I'm pretty sure that LAPack has a solver built-in. Is it easy to access
that?
We currently don't wrap xTRSYL, but if you'd like to file an issue, we
can take it from there.
Thanks,
Have you already noticed that error messages (on julia
0.3.0-prerelease+1419) for sqrt(-1) and sqrt(-1.0) are different?
Here:
julia sqrt(-1)
ERROR: DomainError
julia sqrt(-1.0)
ERROR: DomainError
sqrt will only return a complex result if called with a complex argument.
try sqrt(complex(x))
in
I would say the deployment strategy of Julia is the same as Python. Install
a binary distribution as a prereq of your own package..?
On Thursday, February 13, 2014 7:42:13 PM UTC+11, Gour wrote:
On Thu, 13 Feb 2014 00:07:54 -0500
Stefan Karpinski ste...@karpinski.org javascript: wrote:
It seems to return the more verbose error on every type but Int64 (and
Complex).
On Thursday, 13 February 2014 20:42:59 UTC-6, Paulo Castro wrote:
Have you already noticed that error messages (on julia
0.3.0-prerelease+1419) for sqrt(-1) and sqrt(-1.0) are different?
Here:
julia sqrt(-1)
33 matches
Mail list logo
