Do others get this warning? If so, what modification of Mike's macro avoids
it?
WARNING: deprecated syntax "{a,b, ...}".
Use "Any[a,b, ...]" instead.
On Thursday, September 3, 2015 at 8:12:50 PM UTC-4, Jeffrey Sarnoff wrote:
>
> tip o' hat to that
>
> this minor edit, omitting {Any,Any} (follow
I meant the remote machine/network may be firewalled to only accept
incoming ssh, http and other known ports.
On Tue, Sep 8, 2015 at 5:49 AM, greg_plowman via julia-users <
julia-users@googlegroups.com> wrote:
> Is port 9009 open on the remote machine? You could try with "tunnel=true"
>> if it is
That makes sense!
It was from someone else's code, and I was admittedly surprised that it worked.
Sent from my iPhone
> On 8 Sep 2015, at 10:27 am, Jameson Nash wrote:
>
> This was only "working" in Julia 0.3 because it gave a random answer instead
> of an appropriate error:
>
> julia> abstr
This was only "working" in Julia 0.3 because it gave a random answer
instead of an appropriate error:
julia> abstract AbstractFoo{T}
julia> immutable Foo{T,V} <: AbstractFoo{promote_type(T,V)} end
julia> promote_type(TypeVar(:T),TypeVar(:V))
V #nope
julia> super(Foo{Int,Float64})
AbstractFoo{Fl
>
> Is port 9009 open on the remote machine? You could try with "tunnel=true"
> if it is not open.
I think so.
After running addprocs() and before the wait error, netstat on the remote
machine outputs the following:
C:\Users\Greg>netstat -an
Active Connections
Proto Local Address
It is essentially a copy of siglongjump (link to OpenBSD source for this
funciton:
http://ftp.asia.edu.tw/OS/OpenBSD/src/lib/libc/arch/amd64/gen/sigsetjmp.S)
with slightly different arguments plus some extra work to restore the
floating-point state. The purpose of this function is to replace all
ca
It really depends on what you're trying to achieve. Low memory usage? Speed?
Also consider:
http://julia.readthedocs.org/en/release-0.3/stdlib/arrays/#Base.sub
That should work even better in 0.4. It'll give you an object that behaves
like what you asked for, but it won't actually free the memo
Maybe making TInd immutable here would be sensible unless you need to
directly modify the fields of it a lot. Even then it would probably be
faster to just swap it out for a new TInd with the new fields. That way you
can store the whole vector as a contiguous block of memory.
On Monday, Septemb
Just curious: if it was [true,true,false] or [false,true,true] would it
then be possible? The memory for a matrix is still inline, so these cases
should correspond to changing the dimensions or rebasing the initial
column.
On Tuesday, September 8, 2015 at 4:10:07 AM UTC+10, Tomas Lycken wrot
Note that if you need the equivalent of Python's raw strings you can just
use an empty macro
julia> macro raw_str(s) s end
julia> raw"3.2 3.6 z(x,y) = x@~\327@~exp(-x@+2@+-y@+2@+)"
"3.2 3.6 z(x,y) = x@~\\327@~exp(-x@+2@+-y@+2@+)"
julia> raw"grdmath $ DDX"
"grdmath \$ DDX"
julia>
On Monday,
No, that's not possible, at least not in general - to do that, you would
have to at least shift all the elements of the columns after the deleted
one to make the matrix memory layout consistent.
What you can do, if you don't need to use the matrix as an actual matrix,
is to use e.g. a Vector{Ve
@Viral and others that might know enough to answer this:
I know enough javascript to take a stab at doing the
random-video-on-startpage thing, but I'm having trouble building the site
locally from the github repo. Any advice would be most
appreciated:
https://github.com/JuliaLang/julialang.git
Do you ever modify an individual? If not, why not just write
world[p].ind = newpop
Also, did you profile to make sure that deepcopy is the bottleneck? If your
structures are that simple, it's surprising to me that deepcopy is taking a
long time.
Cédric
On Monday, September 7, 2015 at 1:08:50
Hello everybody,
I'm comparably new to Julia, but not completely new to programming. Yet,
I'm a biologist by training, so please excuse potentially dumb questions in
advance :)
I am working in evolutionary ecology, programming individual-based
simulations. I have now transferred a (very) simpl
Awesome, thanks for the examples. I'll take a look.
On Monday, 7 September 2015 11:37:24 UTC-4, Isaiah wrote:
>
> Yup, that's the hacky solution :)
>
> I modified my example after copying it and didn't fix the last part.
> Here's a working version:
>
>
> julia> err = ccall( (:midiOutGetDevCapsA,
This is great! Can you list the steps required to get this working?
Thanks,
Ista
On Sep 5, 2015 6:07 PM, "Maxwell Peterson"
wrote:
> I have just spent a few hours getting past this error and this thread is a
> top google result, so I am adding my fix:
>
> I had the same issue as Andrew. First, n
Yup, that's the hacky solution :)
I modified my example after copying it and didn't fix the last part. Here's
a working version:
julia> err = ccall( (:midiOutGetDevCapsA, :Winmm), Uint64, (Ptr{Uint64},
Ref{Mid
iOutCaps}, Uint32), 0, output_struct, sizeof(output_struct))
WARNING: convert(::Type{P
Isaiah - Thank you so much. I was able to work around the array issue in
Julia 3.9 by declaring an immutable type with one field for each character.
Kind of an ugly solution, but szPname is limited to 32 characters, so it's
not a big deal.
I think it would be more elegant to use a large bitstype
The problematic symlink is most likely to do with ~/.julia/v0.3/.cache. You
might be able to just remove that then do Pkg.init again. Failing that,
make a copy of the list of currently installed packages you have from
~/.julia/v0.3/REQUIRE, remove ~/.julia entirely and re-install those
packages
I have recently installed (then re-installed) Julia 0.3.11, and I am having
problems with packages. Should I be concerned?
*_*
*_** _ **_**(_)**_** | A fresh approach to technical
computing*
*(_)** | **(_)* *(_)**| Documentation: http://docs.julialang.or
Sorry about the last post - I'm new, so it was caught in the moderation
filter. I actually wrote it before Tony and Isaiah posted.
Isaiah - That looks promising, thanks! It's been a long time since I've
used C, so I completely missed that szPname isn't a pointer.
On Sunday, 6 September 2015 17:
Update: I submitted a PR for libunwind and got the following answer from
*cbergstrom*
(https://github.com/pathscale/libunwind/pull/5#issuecomment-138237585):
I don't want to turn you away from pathscale/libunwind, but do you know if
> Julia can use the Apple libunwind?
I really don't know wha
Here seems to be a safer implementation... but I'm still not quite
confident it's optimally written. Opinions welcome!
function subdivide(v0,v1,filter::Function,middle::Function,doer::Function)
c = Channel(100)
results = Any[doer(v0),doer(v1)]
filter(v0,v1) || return results # not
Thank you.
On Mon, Sep 7, 2015 at 1:48 PM, Yichao Yu wrote:
> On Mon, Sep 7, 2015 at 12:05 AM, Michele Zaffalon
> wrote:
> > Why does
> > input_len::Uint = sizeof(bytes_input)
> > work in the body of the function but not at the REPL?
>
> We don't have typed global yet.
>
> https://github.com/Ju
On Mon, Sep 7, 2015 at 12:05 AM, Michele Zaffalon
wrote:
> Why does
> input_len::Uint = sizeof(bytes_input)
> work in the body of the function but not at the REPL?
We don't have typed global yet.
https://github.com/JuliaLang/julia/issues/8870
>
> Inside the function, it is converted to
> _var0
Is port 9009 open on the remote machine? You could try with "tunnel=true"
if it is not open.
On Mon, Sep 7, 2015 at 4:32 PM, Greg Plowman wrote:
> Hi,
>
> I'm trying to use addprocs() to add remote workers on another windows
> machine.
>
> I'm using a ssh server for windows (Bitvise) with a modi
Hi,
I'm trying to use addprocs() to add remote workers on another windows
machine.
I'm using a ssh server for windows (Bitvise) with a modified Cluster
Manager, and have successfully used this method in another environment.
So I know that it works, although one difference is Window 7 (works) v
Numpy gives the same results as above, but interestingly R gives the
mathematically correct result. It appears that they use a 2-pass approach
for computing the mean:
https://github.com/wch/r-source/blob/6338928e70891145221c781a19d2bb6b17868fb5/src/main/summary.c#L429-L437
-s
On Monday, 7 Septe
The problem is that the computation of the mean is inaccurate: although the
relative error is small, when you perform the subtraction for computing the
standard deviation, as the numbers are almost the same, the relative error
is massively amplified (this is typically known as *catastrophic
can
Aren't you locally creating state on each of the worker processes?
0.3.0 is a year out of date, you should update to 0.3.11. Depending which
BLAS implementation is being used there, you probably need to add
USE_BLAS64=0 when you build. Where is the actual make command getting run
there, and what flags are being used? On master when you set
USE_SYSTEM_BLAS we n
Still no luck - I've looked at the windows sdk, and I'm sure I've got all
the type sizes correct. I've added the stdcall convention to the ccall with
no luck, along with initializing szPname to C_NULL, which seems to be the
more correct approach from the c source I've looked at. I've tried movin
julia> versioninfo()
Julia Version 0.3.0
Platform Info:
System: Linux (x86_64-pc-linux-gnu)
CPU: Intel(R) Core(TM) i7-4550U CPU @
1.50GHz
WORD_SIZE: 64
BLAS: libblas
Hmm, you could try (in pseudo-code):
str = load("my-extension.jl")
ex = parse(str)
eval(ApproxFun, ex)
That should evaluate all of the contents of my-extension.jl in the scope
of the module ApproxFun, so no need for imports or fully qualifying.
However, I don't know whether this is a good idea (o
On Sunday, September 06, 2015 11:46:05 PM Diego Javier Zea wrote:
> Then, what is the best way to test if a number is equal to 0.0?
Not to do it. Much better is to see whether it's much smaller than something
finite, e.g., sta < 100*eps(ave). ave provides the scale against which sta
should be me
*@Jameson*, I was able to go a little further by modifying libunwind's
source code in order to include instead of .
However, there is some point where non-trivial assembly code needs to be
ported:
https://github.com/pathscale/libunwind/blob/vanilla_pathscale/src/x86_64/setcontext.S#L95
I have
36 matches
Mail list logo