This depends a lot on the input sizes. For full length convolutions, the
fft approach should be more accurate because of less additions. For very
short kernels this does not hold anymore. But in practice these kinds of
errors are mostly negligable.
Am Dienstag, 4. März 2014 16:42:33 UTC+1
There was quite some discussion on this topic in
https://groups.google.com/forum/?fromgroups=#!searchin/julia-users/Int32/julia-users/Rte_I6htLRc/VJG5DWVcZbQJ
I had the feeling that we almost convinced Stefan :-)
The good thing is that Array{Int32} operations do not promote to
Array{Int64}. But
Should the user be able to specify the option in code? Winston has an ini
file that is parsed on load. Otherwise it should be possible to solve this
using global variables.
Similar issues arise for the sparse matrix which currently is CSC while CSR
would be needed in my application. Is there some deeper discussion on the
data order I have missed (I have search the ML and Github)?
I know that column major order comes from the Matlab background. But in
this regards
Cool. When having CSR it would be similar useful to have regular row major
order arrays.
I am converting some algorithms from Python/Numpy to Julia and have no idea
how to handle that Julia arrays are column major.
I have a matrix on disk (very large 10GB) and it is stored row major.
This is the natural representation in my application.
Now I want to solve a linear system of
Am Mittwoch, 19. Februar 2014 17:52:12 UTC+1 schrieb Stefan Karpinski:
Yeah, I pretty much agree with all of this. Once field access is
overloadable, we can just create a Direction type and define
(automatically) the following methods:
getfield(::Type{Direction}, ::Field{:NORTH}) =
Yes it would be nice if it would be its own type and if we would have an
abstract type Enum that the concrete enums subclass from. An enum should
be further
compatible with Cint and not only through conversions but also when be used
as field type (to ensure compatibility with C structs).
I
One quite nice approach I have found in Gtk.jl is to define an Enum as:
baremodule Direction
const NORTH = 0
const EAST = 1
const WEST = 2
const SOUTH = 3
end
IMHO it behaves exactly how an enum should work
- The enum entries are in the namespace of the enum name (like in C# and
new
pwd()
Am Dienstag, 18. Februar 2014 01:01:28 UTC+1 schrieb J Luis:
I searched he docs (and Googled) but found no found no way to print the
path. The obvious attempt failed
```
julia path()
ERROR: path not defined
```
have
set the ImakeMagick path to the Windows path and have it inherit from
julia.bat but still get an error saying
julia using Images
WARNING: ImageMagick utilities not found. Install for more file format
support.
Terça-feira, 18 de Fevereiro de 2014 0:04:44 UTC, Tobias Knopp escreveu
.
As I said before, I think this could help Julia newcomers when wrapping
their C/C++ libraries.
Thanks.
--
Carlos
On Mon, Feb 10, 2014 at 5:20 PM, Tobias Knopp
tobias...@googlemail.comjavascript:
wrote:
Am Montag, 10. Februar 2014 14:17:02
Am Montag, 10. Februar 2014 18:27:03 UTC+1 schrieb Carlos Becker:
1) I think that there would be a few issues with the lack of julia's
exported symbols (non-DLLEXPORTed symbols in julia.h)
The ones related to arrays and basic types are exported now, but others
are not, and therefore
Carlos, the code that you showed can be completely written in Julia. It
would be helpful if you could give us more insight what you want to
achieve. Is there a specific API that you want to wrap? You said that the
API returns a double pointer but the length of the memory is not know (if I
get
julia eval(parse(1+2))
3
julia symbol(x)
:x
May be an implicit conclusion: Has similarities with Matlab then it has to
work all out of the box ;-)
Am Mittwoch, 5. Februar 2014 17:26:39 UTC+1 schrieb Steven G. Johnson:
On Tuesday, February 4, 2014 8:21:48 PM UTC-5, Steven Siew wrote:
I think that it should be made absolutely clear to
No there don't seem to be a cleanup method. And multiple julia environments
are currently not possible.
But from my perspective it would be great to work towards putting all the
globals into a struct so that multiple Julia threads per process become
feasible.
Am Mittwoch, 5. Februar 2014
It already works. But as this is on julia-users I just wanted to mention
that this is only a prototype and getting all the dependencies working is
non-trivial (at least on windows and OSX). The code lives here:
https://github.com/tknopp/Julietta.jl
Am Dienstag, 4. Februar 2014 13:13:22 UTC+1
I have also kind of reinvented the wheel for my Gtk based terminal but this
was on purpose to get something quickly working and later replace it with
the proper solution.
But its still not entirely clear what the best solution is. The REPL.jl
package also seems to have some kind of client
Isn't the LICENSE.md file in Julia pretty clear? Julia is MIT licensed and
repl-readline.chttps://github.com/JuliaLang/julia/blob/master/ui/repl-readline.cis
GPL. I don't see the problem. If I where using libjulia, I can use it in
a commercial program. One is of course not allowed to ship fftw
If you want to track master, you will have to do Pkg.checkout(Gtk). 0.5
is pretty fresh though
Am Sonntag, 26. Januar 2014 20:26:50 UTC+1 schrieb Andreas Lobinger:
Hello colleague,
On Sunday, January 26, 2014 6:22:13 PM UTC+1, Tim Holy wrote:
I don't get those errors at all. Are you on
No. Giving types in function definitions does not give you any speedup as
the function are always compiled for concrete types.
When you define composite types it is however important to give concrete
types for optimal performance.
Am Donnerstag, 23. Januar 2014 00:41:41 UTC+1 schrieb Patrick
janvier 2014 à 00:17 -0800, Tobias Knopp a écrit :
@Stefan: Is there a good reason to promote Int32 operations to the
native mashine type? Sorry, if this has already discussed in depth.
But for me it feals wrong to automatically upcast integer operations.
I think that the type should keep
But why isn't float32 promoting to float64 on basic arithmetics then?
As Markus pointed out, when we get SIMD support (in the form of the @simd
macro or by autovectorization of llvm) there can be a factor of two between
both integer computations.
P.S.: Hope you don't get this wrong. Its just
Would be interesting to see some use cases where Java-like OO better fits
than Julias OO. In C++ one can use both and usually choses based on whether
the dispatching can be done at runtime or at compile time (i.e. classes
with virtual function for runtime decisions and templates for compile
201 - 225 of 225 matches
Mail list logo
