[julia-users] Re: Best way to convert Any of vectors into matrix

[Gunnar Farnebäck]

For maximum efficiency Diego's loop solution is likely best but for 
succinctness, try hcat(A...).

Den lördagen den 23:e augusti 2014 kl. 04:05:19 UTC+2 skrev Bradley Setzler:
>
> Good evening,
>
> I often have Any types filled with vectors (especially as the return of a 
> pmap), and need them as a matrix or DataFrame. For example, suppose I have,
>
> julia> A
> 3-element Array{Any,1}: 
> [1,2] 
> [3,4] 
> [5,6]
>
> But I want,
>
> julia> B
> 2x3 Array{Int64,2}: 
> 1 3 5 
> 2 4 6
>
> I came up with the following, which successfully constructs B from A, but 
> it's a bit inefficient/messy:
>
> B = A[1]
> for i=2:length(A)
> B = hcat(B,A[i])
> end
>
> Is there a better way to do this? Something like,
> julia> B = hcatForAny(A)
>
> Thanks,
> Bradley
>
>

Re: [julia-users] problem with array of arrays

[Tim Holy]

You need to pass a collection, not an element, as the second argument.

b = Vector{Int}[A[3]]
julia> findin(A, b)
1-element Array{Int64,1}:
 3

Your version was trying to match elements of A (vectors) to -1, 2, and 8, and 
of course none of those match.

--Tim

On Sunday, August 24, 2014 02:06:06 PM K Leo wrote:
> What is wrong with the following?
> 
> julia> A
> 4-element Array{Array{Int64,1},1}:
>   [-1,2,4]
>   [-1,2,7]
>   [-1,2,8]
>   [1,-2,10]
> 
> julia> findin(A, A[3])
> 0-element Array{Int64,1}

[julia-users] How to import module from another file in same directory?

[Andrei]

Let's say I have following project layout:

P.jl that contains module P -- main package module, exposes code from a.jl
and b.jl
a.jl that contains module A and
b.jl that contains module B -- some domain specific modules
u.jl that contains module U -- util functions

Now I want to use functions in U from modules A and B. In simplest case I
would just include("u.jl") inside of a.jl and b.jl, but this way functions
from U will be defined in both - A and B. So I really want to import U, not
include u.jl, but I can't do this since u.jl is not on the LOAD_PATH (and
messing with it manually looks somewhat bad to me).

Is there some standard way to tackle it?

(Note, that A, B and U are here just for code splitting, other ways to do
same stuff are ok too.)

[julia-users] question about performance hit because of build options on unix HPC

[Florian Oswald]

Hi!

I'm running julia release 0.3 on an hpc system that features two 6-core 
2.4GHz Intel Westmere processors (a total of 12 processor cores) per node. 
I noticed my program is significantly slower on that machine than on my 
mac, where I run this on 1 core (2.4GHz Intel Core i5). (runtime is 60 secs 
vs 90 secs)

I would expect the hpc cores to be slightly slower but was surprised it was 
that much. I was wondering if any of the options I set to build this are 
culprit for this? I do run the binary of release 0.3 on my mac, whereas I 
build from source with those options on the hpc:

OPENBLAS_TARGET_ARCH=NEHALEM 
OPENBLAS_DYNAMIC_ARCH = 0 
OPENBLAS_USE_THREAD=0 
JULIA_CPU_TARGET=core2

I can build julia without the final option. I cannot build it without the 
first three options.

I cannot run multiple nodes without the final option, since I get an error 

Target architecture mismatch. Please delete or regenerate 
sys.{so,dll,dylib}.

[julia-users] Re: Colors in REPL (OSX)

[Sean Garborg]

Not quite what you're asking for, but my approach has been to tweak the 
terminal's color scheme (or select/download a new one) so that none of the 
ANSI colors that command line programs reach for is a bad fit for my 
background color.

On Thursday, August 21, 2014 2:57:57 PM UTC-4, Rob J Goedman wrote:
>
> Just wondering, I used to change the color of the help prompt by editing 
> REPL.jl and re-making julia.
>
> With the release of 0.3 I'm trying to stick to the official version.
>
> Is there a way to change these colors (:yellow for help conflicts with my 
> background color), maybe when starting Julia REPL?
>
> Thanks,
> Rob J. Goedman
> goe...@icloud.com 
>
>
>
>  
>

[julia-users] Re: question about performance hit because of build options on unix HPC

[Valentin Churavy]

Hej,

so from what I gather if you set JULIA_CPU_TARGET=core2, Julia restricts 
itself to the cpu features defined 
here: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86.td?view=markup
Currently the only supported and tested options are i386, core2 and native, 
whereas native tries to detect the available features of your cpu and uses 
all of them, leading to better performance on more advanced architectures. 

The problem is that if you compile with target set to native you can't 
deploy your binaries to different machines with different processors. There 
is currently a PR that might help in that 
regard. https://github.com/JuliaLang/julia/pull/8074

The idea would be to build Julia with core2 and on each cluster node 
regenerate the sysimage with target=native.

Best Valentin

On Sunday, 24 August 2014 14:37:03 UTC+2, Florian Oswald wrote:
>
> Hi!
>
> I'm running julia release 0.3 on an hpc system that features two 6-core 
> 2.4GHz Intel Westmere processors (a total of 12 processor cores) per node. 
> I noticed my program is significantly slower on that machine than on my 
> mac, where I run this on 1 core (2.4GHz Intel Core i5). (runtime is 60 secs 
> vs 90 secs)
>
> I would expect the hpc cores to be slightly slower but was surprised it 
> was that much. I was wondering if any of the options I set to build this 
> are culprit for this? I do run the binary of release 0.3 on my mac, whereas 
> I build from source with those options on the hpc:
>
> OPENBLAS_TARGET_ARCH=NEHALEM 
> OPENBLAS_DYNAMIC_ARCH = 0 
> OPENBLAS_USE_THREAD=0 
> JULIA_CPU_TARGET=core2
>
> I can build julia without the final option. I cannot build it without the 
> first three options.
>
> I cannot run multiple nodes without the final option, since I get an error 
>
> Target architecture mismatch. Please delete or regenerate 
> sys.{so,dll,dylib}.
>
>
>
>
>

[julia-users] finding a .so in ccall

[Andreas Lobinger]

Hello,

maybe it's the weather, maybe it's a side effect, but what happens here:

julia> Duktape._jl_duktape
"/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"

julia> 
ccall((:duk_create_heap,"/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"),Ptr{Void},())
Ptr{Void} @0x0cfd5ee0

julia> ccall((:duk_create_heap,Duktape._jl_duktape),Ptr{Void},())
ERROR: type: anonymous: in ccall: first argument not a pointer or valid 
constant expression, expected DataType, got Type{(Any...,)}
 in anonymous at no file

i remember something about that the library needs to be found somehow, but 
that would not explain the sucessful first call ... for me the difference 
is just the string as literal or as variable.

Re: [julia-users] finding a .so in ccall

[Isaiah Norton]

You probably need to declare as const

within Duktape:

`const _jl_duktape = "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"`


On Sun, Aug 24, 2014 at 10:16 AM, Andreas Lobinger 
wrote:

> Hello,
>
> maybe it's the weather, maybe it's a side effect, but what happens here:
>
> julia> Duktape._jl_duktape
> "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"
>
> julia>
> ccall((:duk_create_heap,"/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"),Ptr{Void},())
> Ptr{Void} @0x0cfd5ee0
>
> julia> ccall((:duk_create_heap,Duktape._jl_duktape),Ptr{Void},())
> ERROR: type: anonymous: in ccall: first argument not a pointer or valid
> constant expression, expected DataType, got Type{(Any...,)}
>  in anonymous at no file
>
> i remember something about that the library needs to be found somehow, but
> that would not explain the sucessful first call ... for me the difference
> is just the string as literal or as variable.
>
>
>
>
>

Re: [julia-users] finding a .so in ccall

[Andreas Lobinger]

Thank you, i did and it worked. 

I tried to find thisin the documentation and actually more than once it's 
stated that the tuple needs to be const. But it wasn't obvious for me that 
there are different flavours of the string. The error message could be 
improved (imho).

On Sunday, August 24, 2014 4:18:53 PM UTC+2, Isaiah wrote:
>
> You probably need to declare as const
>
> within Duktape:
>
> `const _jl_duktape = "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"`
>
>
> On Sun, Aug 24, 2014 at 10:16 AM, Andreas Lobinger  > wrote:
>
>> Hello,
>>
>> maybe it's the weather, maybe it's a side effect, but what happens here:
>>
>> julia> Duktape._jl_duktape
>> "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"
>>
>> julia> 
>> ccall((:duk_create_heap,"/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"),Ptr{Void},())
>> Ptr{Void} @0x0cfd5ee0
>>
>> julia> ccall((:duk_create_heap,Duktape._jl_duktape),Ptr{Void},())
>> ERROR: type: anonymous: in ccall: first argument not a pointer or valid 
>> constant expression, expected DataType, got Type{(Any...,)}
>>  in anonymous at no file
>>
>> i remember something about that the library needs to be found somehow, 
>> but that would not explain the sucessful first call ... for me the 
>> difference is just the string as literal or as variable.
>>
>>
>>
>>
>>
>

[julia-users] Re: How to import module from another file in same directory?

[Valentin Churavy]

What you are looking for is described 
in 
http://julia.readthedocs.org/en/latest/manual/modules/#relative-and-absolute-module-paths

in P.jl you include all your submodules 
module P
 include("u.jl")
 include("a.jl")
 include("b.jl")

 using .A, .B

 export f, g
end

 u.jl 
module U
 g() = 5
 f() = 6
end

a.jl and b.jl both lokk like this

module A
 import ..U

 f = U.f
 g = U.g

 export f, g
end

so one dot as a prefix looks in the namespace of the current module and two 
dots as prefix looks in the namespace of the parent module.

Hope that helps

On Sunday, 24 August 2014 14:10:58 UTC+2, Andrei Zh wrote:
>
> Let's say I have following project layout: 
>
> P.jl that contains module P -- main package module, exposes code from a.jl 
> and b.jl
> a.jl that contains module A and
> b.jl that contains module B -- some domain specific modules
> u.jl that contains module U -- util functions
>
> Now I want to use functions in U from modules A and B. In simplest case I 
> would just include("u.jl") inside of a.jl and b.jl, but this way functions 
> from U will be defined in both - A and B. So I really want to import U, not 
> include u.jl, but I can't do this since u.jl is not on the LOAD_PATH (and 
> messing with it manually looks somewhat bad to me).  
>
> Is there some standard way to tackle it? 
>
> (Note, that A, B and U are here just for code splitting, other ways to do 
> same stuff are ok too.)
>

Re: [julia-users] Re: question about performance hit because of build options on unix HPC

[Florian Oswald]

Hi!
thanks for that, very helpful. So do I have to do anything to build a
CPU-specific binary after I get this PR? Do i understand this correctly
that i shoudl set MARCH=native now, and julia will figure out at runtime -
i.e. when I use it on whatever architecture - which are the best settings
for that machine? That sounds quite amazing.


On 24 August 2014 15:05, Valentin Churavy  wrote:

> Hej,
>
> so from what I gather if you set JULIA_CPU_TARGET=core2, Julia restricts
> itself to the cpu features defined here:
> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86.td?view=markup
> Currently the only supported and tested options are i386, core2 and
> native, whereas native tries to detect the available features of your cpu
> and uses all of them, leading to better performance on more advanced
> architectures.
>
> The problem is that if you compile with target set to native you can't
> deploy your binaries to different machines with different processors. There
> is currently a PR that might help in that regard.
> https://github.com/JuliaLang/julia/pull/8074
>
> The idea would be to build Julia with core2 and on each cluster node
> regenerate the sysimage with target=native.
>
> Best Valentin
>
> On Sunday, 24 August 2014 14:37:03 UTC+2, Florian Oswald wrote:
>>
>> Hi!
>>
>> I'm running julia release 0.3 on an hpc system that features two 6-core
>> 2.4GHz Intel Westmere processors (a total of 12 processor cores) per node.
>> I noticed my program is significantly slower on that machine than on my
>> mac, where I run this on 1 core (2.4GHz Intel Core i5). (runtime is 60 secs
>> vs 90 secs)
>>
>> I would expect the hpc cores to be slightly slower but was surprised it
>> was that much. I was wondering if any of the options I set to build this
>> are culprit for this? I do run the binary of release 0.3 on my mac, whereas
>> I build from source with those options on the hpc:
>>
>> OPENBLAS_TARGET_ARCH=NEHALEM
>> OPENBLAS_DYNAMIC_ARCH = 0
>> OPENBLAS_USE_THREAD=0
>> JULIA_CPU_TARGET=core2
>>
>> I can build julia without the final option. I cannot build it without the
>> first three options.
>>
>> I cannot run multiple nodes without the final option, since I get an
>> error
>>
>> Target architecture mismatch. Please delete or regenerate
>> sys.{so,dll,dylib}.
>>
>>
>>
>>
>>

Re: [julia-users] needless loss of performance: immutable conflates two concepts

[Adam Strzelecki]

> That's exactly why we can't do it. Julia is for people who are sketching with 
> code as much as it is for people who are building robust systems. The default 
> has to be easy and obvious – and the most obvious, no fuss default is to let 
> you change things. If there are going to be limits put on programmers for 
> robustness, they need to be opt-in. Ideally it should be simple and 
> convenient to opt into, but still, optional.

Fair enough. Maybe then enforce these only for (public) modules?

--Adam

Re: [julia-users] REPL v0.3, matlab-like completion

[Adrian Cuthbertson]

I use rlwrap with a simple "one-liner" repl (inspired by previous
suggestions by Steven G Johnson):

# $HOME/bin/rlrepl.jl  (julia 0.3.0)
function rlrepl()
ok=true
push!(LOAD_PATH,pwd())
while ok
print("jul> ")
linerd =readline()
if length(linerd) == 0
return
end
line = strip(linerd)
if length(line) == 0
continue
end
try
expr=parse(line)
ans=eval(Main,expr)
if ans != nothing && line[length(line)] != ';'
display("text/plain", ans)
println()
end
catch(ex)
Base.showerror(STDOUT,ex)
println()
end
end
end
rlrepl()

Then in rlwrap's inputrc file (either $HOME/.inputrc or /etc/inputrc),
add/change the following:
"\e[A": history-search-backward
"\e[B": history-search-forward
"\e[5": history-search-backward
"\e[6": history-search-forward

Then, to invoke rlrepl.jl...

# OS/X Terminal and SSH session to Linux (Ubuntu):
rlwrap -pBlue -afoo -tdumb -H 'ju.hist' -l 'ju.log' -s 1 julia
--no-history-file ~/bin/rlrepl.jl

You forego all the niceties of the new repl, but for me (who can't loose
the up/dn arrow) this works just like all my other repl/shell activities
(and I have a session log as well).

Hope that helps some of those other up/dn arrow history-search addicts out
there :).

Regards, Adrian









On Wed, Aug 20, 2014 at 3:04 PM, Carlos Becker 
wrote:

> Hi Tim, yes, I do know about Ctrl-R, but it is much lazier to use the
> up/down arrows, particularly when navigating history (otherwise you have to
> use Ctrl-R repeatedly, and Ctrl-Shift-R to go back).
>
> El miércoles, 20 de agosto de 2014 00:39:33 UTC+2, Tim Holy escribió:
>
>> You know about the Ctrl-r shortcut though, right?
>> --Tim
>>
>> On Tuesday, August 19, 2014 10:57:19 PM Carlos Becker wrote:
>> > Hi all,
>> >
>> > I think this is a typically asked question, but I don't know whether it
>> is
>> > possible now in julia v0.3.
>> > Namely to make the up/down arrows search completion history in the
>> REPL. If
>> > so, I will be happy to document it in the docs in the FAQ section.
>> >
>> > Thanks.
>> >
>> > --
>> > Carlos
>>
>>

[julia-users] Re: Best way to convert Any of vectors into matrix

[Bradley Setzler]

Hi Gunnar,

Correct me if I'm wrong, but hcat does not work for Any types. I'm looking 
for a version of hcat that works with Any.

We have A:

julia> A 
3-element Array{Any,1}: 
[1,2] 
[3,4] 
[5,6] 

We want B:

julia> B 
2x3 Array{Int64,2}: 
1 3 5 
2 4 6 

So we try hcat(A) but it fails to return B:

julia> hcat(A) 
3x1 Array{Any,2}: 
[1,2] 
[3,4] 
[5,6]


Thanks,
Bradley



On Sunday, August 24, 2014 4:01:25 AM UTC-5, Gunnar Farnebäck wrote:
>
> For maximum efficiency Diego's loop solution is likely best but for 
> succinctness, try hcat(A...).
>
> Den lördagen den 23:e augusti 2014 kl. 04:05:19 UTC+2 skrev Bradley 
> Setzler:
>>
>> Good evening,
>>
>> I often have Any types filled with vectors (especially as the return of a 
>> pmap), and need them as a matrix or DataFrame. For example, suppose I have,
>>
>> julia> A
>> 3-element Array{Any,1}: 
>> [1,2] 
>> [3,4] 
>> [5,6]
>>
>> But I want,
>>
>> julia> B
>> 2x3 Array{Int64,2}: 
>> 1 3 5 
>> 2 4 6
>>
>> I came up with the following, which successfully constructs B from A, but 
>> it's a bit inefficient/messy:
>>
>> B = A[1]
>> for i=2:length(A)
>> B = hcat(B,A[i])
>> end
>>
>> Is there a better way to do this? Something like,
>> julia> B = hcatForAny(A)
>>
>> Thanks,
>> Bradley
>>
>>

[julia-users] Re: Best way to convert Any of vectors into matrix

[Patrick O'Leary]

You need to splat (...) A: hcat(A...), in Gunnar's email, vs. hcat(A), in 
your email. This will supply each element of A as a separate argument to 
hcat().

Much like the prefix dot for elementwise computation, it's hard to see in a 
proportional typeface.

On Sunday, August 24, 2014 10:36:29 AM UTC-5, Bradley Setzler wrote:
>
> Hi Gunnar,
>
> Correct me if I'm wrong, but hcat does not work for Any types. I'm looking 
> for a version of hcat that works with Any.
>
> We have A:
>
> julia> A 
> 3-element Array{Any,1}: 
> [1,2] 
> [3,4] 
> [5,6] 
>
> We want B:
>
> julia> B 
> 2x3 Array{Int64,2}: 
> 1 3 5 
> 2 4 6 
>
> So we try hcat(A) but it fails to return B:
>
> julia> hcat(A) 
> 3x1 Array{Any,2}: 
> [1,2] 
> [3,4] 
> [5,6]
>
>
> Thanks,
> Bradley
>
>
>
> On Sunday, August 24, 2014 4:01:25 AM UTC-5, Gunnar Farnebäck wrote:
>>
>> For maximum efficiency Diego's loop solution is likely best but for 
>> succinctness, try hcat(A...).
>>
>> Den lördagen den 23:e augusti 2014 kl. 04:05:19 UTC+2 skrev Bradley 
>> Setzler:
>>>
>>> Good evening,
>>>
>>> I often have Any types filled with vectors (especially as the return of 
>>> a pmap), and need them as a matrix or DataFrame. For example, suppose I 
>>> have,
>>>
>>> julia> A
>>> 3-element Array{Any,1}: 
>>> [1,2] 
>>> [3,4] 
>>> [5,6]
>>>
>>> But I want,
>>>
>>> julia> B
>>> 2x3 Array{Int64,2}: 
>>> 1 3 5 
>>> 2 4 6
>>>
>>> I came up with the following, which successfully constructs B from A, 
>>> but it's a bit inefficient/messy:
>>>
>>> B = A[1]
>>> for i=2:length(A)
>>> B = hcat(B,A[i])
>>> end
>>>
>>> Is there a better way to do this? Something like,
>>> julia> B = hcatForAny(A)
>>>
>>> Thanks,
>>> Bradley
>>>
>>>

[julia-users] Re: Announcement: Playground.jl

[Adam Smith]

Nice work! It seems to work on a Mac as well. There is a warning that gets 
printed every time you use the playground command, but it still seems to 
work.

$ playground . activate
Warning: requiring "Playground" did not define a corresponding module.
playground> echo "foo"
foo
playground> julia
   _
   _   _ _(_)_ |  A fresh approach to technical computing
  (_) | (_) (_)|  Documentation: http://docs.julialang.org
   _ _   _| |_  __ _   |  Type "help()" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.3.0 (2014-08-20 20:43 UTC)
 _/ |\__'_|_|_|\__'_|  |  Official http://julialang.org/ release
|__/   |  x86_64-apple-darwin13.3.0

julia> using Gadfly
ERROR: Gadfly not found
 in require at loading.jl:47

julia>

playground> exit



On Saturday, August 23, 2014 4:01:45 PM UTC-4, Rory Finnegan wrote:
>
> Hi everyone,
>
> I've published my Playground.jl 
>  package to create julia 
> sandboxes like python virtual environments, if anyone wants to give it a 
> try.  So far I've tested it on Funtoo and Linux Mint, but I'm looking for 
> people to try it out on other platforms (like Windows and OSX).
>
> Cheers,
> Rory
>

[julia-users] Re: Best way to convert Any of vectors into matrix

[Bradley Setzler]

Hi Patrick,

Thank you for showing that the dots (...) in Gunnar's solution are 
necessary, it works now:

julia> hcat(A...) 
2x3 Array{Int64,2}: 
1 3 5 
2 4 6

Best,
Bradley



On Sunday, August 24, 2014 10:52:13 AM UTC-5, Patrick O'Leary wrote:
>
> You need to splat (...) A: hcat(A...), in Gunnar's email, vs. hcat(A), in 
> your email. This will supply each element of A as a separate argument to 
> hcat().
>
> Much like the prefix dot for elementwise computation, it's hard to see in 
> a proportional typeface.
>
> On Sunday, August 24, 2014 10:36:29 AM UTC-5, Bradley Setzler wrote:
>>
>> Hi Gunnar,
>>
>> Correct me if I'm wrong, but hcat does not work for Any types. I'm 
>> looking for a version of hcat that works with Any.
>>
>> We have A:
>>
>> julia> A 
>> 3-element Array{Any,1}: 
>> [1,2] 
>> [3,4] 
>> [5,6] 
>>
>> We want B:
>>
>> julia> B 
>> 2x3 Array{Int64,2}: 
>> 1 3 5 
>> 2 4 6 
>>
>> So we try hcat(A) but it fails to return B:
>>
>> julia> hcat(A) 
>> 3x1 Array{Any,2}: 
>> [1,2] 
>> [3,4] 
>> [5,6]
>>
>>
>> Thanks,
>> Bradley
>>
>>
>>
>> On Sunday, August 24, 2014 4:01:25 AM UTC-5, Gunnar Farnebäck wrote:
>>>
>>> For maximum efficiency Diego's loop solution is likely best but for 
>>> succinctness, try hcat(A...).
>>>
>>> Den lördagen den 23:e augusti 2014 kl. 04:05:19 UTC+2 skrev Bradley 
>>> Setzler:

 Good evening,

 I often have Any types filled with vectors (especially as the return of 
 a pmap), and need them as a matrix or DataFrame. For example, suppose I 
 have,

 julia> A
 3-element Array{Any,1}: 
 [1,2] 
 [3,4] 
 [5,6]

 But I want,

 julia> B
 2x3 Array{Int64,2}: 
 1 3 5 
 2 4 6

 I came up with the following, which successfully constructs B from A, 
 but it's a bit inefficient/messy:

 B = A[1]
 for i=2:length(A)
 B = hcat(B,A[i])
 end

 Is there a better way to do this? Something like,
 julia> B = hcatForAny(A)

 Thanks,
 Bradley

Re: [julia-users] Re: question about performance hit because of build options on unix HPC

[Jameson Nash]

if you set MARCH=native, julia (and all dependencies) will compile to only
be runnable on the current machine (or other identical CPUs).

what kind of code are you running? if you are doing large linear algebra
operations, OPENBLAS_USE_THREAD=0 could be a likely culprit. I thought the
i5 Mac's were mostly dual or quad core.

Any of the processors list in the link above should work, so you can pick
any newer processor whose feature set is compatible with all of your
machines, such as nehalem or corei7-avx (sandybridge)


On Sun, Aug 24, 2014 at 10:50 AM, Florian Oswald 
wrote:

> Hi!
> thanks for that, very helpful. So do I have to do anything to build a
> CPU-specific binary after I get this PR? Do i understand this correctly
> that i shoudl set MARCH=native now, and julia will figure out at runtime -
> i.e. when I use it on whatever architecture - which are the best settings
> for that machine? That sounds quite amazing.
>
>
> On 24 August 2014 15:05, Valentin Churavy  wrote:
>
>> Hej,
>>
>> so from what I gather if you set JULIA_CPU_TARGET=core2, Julia restricts
>> itself to the cpu features defined here:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86.td?view=markup
>> Currently the only supported and tested options are i386, core2 and
>> native, whereas native tries to detect the available features of your cpu
>> and uses all of them, leading to better performance on more advanced
>> architectures.
>>
>> The problem is that if you compile with target set to native you can't
>> deploy your binaries to different machines with different processors. There
>> is currently a PR that might help in that regard.
>> https://github.com/JuliaLang/julia/pull/8074
>>
>> The idea would be to build Julia with core2 and on each cluster node
>> regenerate the sysimage with target=native.
>>
>> Best Valentin
>>
>> On Sunday, 24 August 2014 14:37:03 UTC+2, Florian Oswald wrote:
>>>
>>> Hi!
>>>
>>> I'm running julia release 0.3 on an hpc system that features two 6-core
>>> 2.4GHz Intel Westmere processors (a total of 12 processor cores) per node.
>>> I noticed my program is significantly slower on that machine than on my
>>> mac, where I run this on 1 core (2.4GHz Intel Core i5). (runtime is 60 secs
>>> vs 90 secs)
>>>
>>> I would expect the hpc cores to be slightly slower but was surprised it
>>> was that much. I was wondering if any of the options I set to build this
>>> are culprit for this? I do run the binary of release 0.3 on my mac, whereas
>>> I build from source with those options on the hpc:
>>>
>>> OPENBLAS_TARGET_ARCH=NEHALEM
>>> OPENBLAS_DYNAMIC_ARCH = 0
>>> OPENBLAS_USE_THREAD=0
>>> JULIA_CPU_TARGET=core2
>>>
>>> I can build julia without the final option. I cannot build it without
>>> the first three options.
>>>
>>> I cannot run multiple nodes without the final option, since I get an
>>> error
>>>
>>> Target architecture mismatch. Please delete or regenerate
>>> sys.{so,dll,dylib}.
>>>
>>>
>>>
>>>
>>>
>

Re: [julia-users] Re: Colors in REPL (OSX)

[Rob Goedman]

Thanks Sean, yes, for now I just use the Terminal preferences to help out.

Regards,
Rob

Sent from Rob Goedman's iPad Mini


> On Aug 24, 2014, at 06:58, Sean Garborg  wrote:
> 
> Not quite what you're asking for, but my approach has been to tweak the 
> terminal's color scheme (or select/download a new one) so that none of the 
> ANSI colors that command line programs reach for is a bad fit for my 
> background color.
> 
>> On Thursday, August 21, 2014 2:57:57 PM UTC-4, Rob J Goedman wrote:
>> Just wondering, I used to change the color of the help prompt by editing 
>> REPL.jl and re-making julia.
>> 
>> With the release of 0.3 I'm trying to stick to the official version.
>> 
>> Is there a way to change these colors (:yellow for help conflicts with my 
>> background color), maybe when starting Julia REPL?
>> 
>> Thanks,
>> Rob J. Goedman
>> goe...@icloud.com
>> 
>> 
>> 
>> 

Re: [julia-users] needless loss of performance: immutable conflates two concepts

[Stefan Karpinski]

That's certainly a possibility – perhaps part of the planned paranoid mode
that lints code etc.


On Sun, Aug 24, 2014 at 11:11 AM, Adam Strzelecki  wrote:

> > That's exactly why we can't do it. Julia is for people who are sketching
> with code as much as it is for people who are building robust systems. The
> default has to be easy and obvious – and the most obvious, no fuss default
> is to let you change things. If there are going to be limits put on
> programmers for robustness, they need to be opt-in. Ideally it should be
> simple and convenient to opt into, but still, optional.
>
> Fair enough. Maybe then enforce these only for (public) modules?
>
> --Adam

Re: [julia-users] why sum(abs(A)) is very slow

[Adam Smith]

I've run into this a few times (and a few hundred times in python), so I 
made an @iterize macro. Not sure how useful it is, but you can put it in 
front of a bunch of chained function calls and it will make iterators 
automatically to avoid creating any temp arrays:

A = randn(1<<26)
@time sum(abs(A))
@time @iterize sum(abs(A))
@time sumabs(A)

println(sum(abs(A)))
println(@iterize sum(abs(A)))
println(sumabs(A))

println(sum(A))
println(@iterize sum(A))

println(sum(ceil(floor(abs(A)
println(@iterize sum(ceil(floor(abs(A)

Output:
elapsed time: 0.367873796 seconds (537878296 bytes allocated, 2.48% gc time)
elapsed time: 0.107278414 seconds (577616 bytes allocated)
elapsed time: 0.045590637 seconds (639580 bytes allocated)
5.3551932868680775e7
5.3551932868672036e7
5.3551932868678436e7
658.6904827808266
658.6904827808266
2.4537098e7
2.4537098e7

The macro is in a gist: Iterize.jl 


I had tried using @devec, but that actually made it about 100x slower.

On Saturday, August 23, 2014 8:15:44 AM UTC-4, Stefan Karpinski wrote:
>
> On Sat, Aug 23, 2014 at 7:23 AM, > wrote:
>
>> >To do any of that justice, you end up with a language that looks 
>> basically like Haskell. So why not just use Haskell?
>>
>> Because I don't know anything about it (yet), except the name and the 
>> fact that you often associated it with lazy evaluation.
>>
>> Because (#2), this could be a way to make sumabs and the likes obsolete 
>> in *Julia*. :)
>>
>
> We do really want to get rid of things like sumabs., so it's certainly 
> worth considering. I know I've thought about it many times, but I don't 
> think it's the right answer – you really want to preserve eager evaluation 
> semantics, even if you end up moving around the actual evaluation of things.
>  

[julia-users] Does Julia have something similar to Python's documentation string?

[Ivar Nesje]

Not yet. 

Re: [julia-users] finding a .so in ccall

[Milan Bouchet-Valat]

Le dimanche 24 août 2014 à 07:31 -0700, Andreas Lobinger a écrit :
> Thank you, i did and it worked. 
> 
> I tried to find thisin the documentation and actually more than once
> it's stated that the tuple needs to be const. But it wasn't obvious
> for me that there are different flavours of the string. The error
> message could be improved (imho).
Please file an issue on GitHub, maybe it's possible to fix this.


Regards

> On Sunday, August 24, 2014 4:18:53 PM UTC+2, Isaiah wrote:
> You probably need to declare as const
> 
> 
> within Duktape:
> 
> 
> `const _jl_duktape
> = "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"`
> 
> 
> On Sun, Aug 24, 2014 at 10:16 AM, Andreas Lobinger
>  wrote:
> Hello,
> 
> maybe it's the weather, maybe it's a side effect, but
> what happens here:
> 
> julia> Duktape._jl_duktape
> "/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"
> 
> julia>
> 
> ccall((:duk_create_heap,"/home/lobi/juliarepo/duktape-0.11.0/src/duktape.so"),Ptr{Void},())
> Ptr{Void} @0x0cfd5ee0
> 
> julia>
> ccall((:duk_create_heap,Duktape._jl_duktape),Ptr{Void},())
> ERROR: type: anonymous: in ccall: first argument not a
> pointer or valid constant expression, expected
> DataType, got Type{(Any...,)}
>  in anonymous at no file
> 
> i remember something about that the library needs to
> be found somehow, but that would not explain the
> sucessful first call ... for me the difference is just
> the string as literal or as variable.
> 
> 
> 
> 
> 
> 

[julia-users] Re: Does Julia have something similar to Python's documentation string?

[Job van der Zwan]

Any plans? Discussions on Github worth reading through?

I personally am really charmed by the godoc 
 approach - could 
something like that work for Julia? (so figuring out a sensible idiomatic 
way to document functions and modules that makes the documentation easy to 
read in plaintext, but also easy to be turned into pretty formatted 
documentation by tools)

On Sunday, 24 August 2014 23:07:30 UTC+2, Ivar Nesje wrote:
>
> Not yet. 

Re: [julia-users] Multivariate Normal versus Multivariate Normal Canon in Distributions package

[John Myles White]

Can you clarify a bit what you mean by “same result”?

It would also be great to see a simpler example that doesn’t involve the 
Bayesian updating steps, which hopefully aren’t essential to hitting an error.

 — John

On Aug 23, 2014, at 4:33 PM, asim  wrote:

> 
> 
> I am trying to use the multivariate normal canonical form to draw random 
> numbers. According to my understanding, the following two functions should 
> give the same result over a large number of draws, but I am getting different 
> results. What am I doing wrong. 
> 
> function ablRegCoeffPostDraw(xtx::Array{Float64, 2}, xty::Array{Float64, 1},
>errorVariance::Float64, priorMean::Array{Float64, 1}, 
> priorPrec::Array{Float64, 2})
> 
> postPrec=priorPrec+xtx/errorVariance
> 
> postMean=priorPrec*priorMean+xty/errorVariance
>   
> postMean=postPrec\postMean
>   
> rand(MvNormal(postMean, inv(postPrec)))
>   
> end
> 
> function ablRegCoeffPostDraw1(xtx::Array{Float64, 2}, xty::Array{Float64, 1},
>errorVariance::Float64, priorMean::Array{Float64, 1}, 
> priorPrec::Array{Float64, 2})
> 
> postPrec=priorPrec+xtx/errorVariance
> 
> postMean=priorPrec*priorMean+xty/errorVariance
>   
> postMean=postPrec\postMean
>   
> potential=postPrec*postMean
>   
> rand(MvNormalCanon(potential, postPrec))
>   
> end

[julia-users] Re: Does Julia have something similar to Python's documentation string?

[Jason Knight]

Happy reading: https://github.com/JuliaLang/julia/issues/3988 :)

On Sunday, August 24, 2014 5:03:04 PM UTC-5, Job van der Zwan wrote:
>
> Any plans? Discussions on Github worth reading through?
>
> I personally am really charmed by the godoc 
>  approach - could 
> something like that work for Julia? (so figuring out a sensible idiomatic 
> way to document functions and modules that makes the documentation easy to 
> read in plaintext, but also easy to be turned into pretty formatted 
> documentation by tools)
>
> On Sunday, 24 August 2014 23:07:30 UTC+2, Ivar Nesje wrote:
>>
>> Not yet. 
>
>

Re: [julia-users] Re: Does Julia have something similar to Python's documentation string?

[Stefan Karpinski]

I really like godoc – that's basically what I want plus a convention that the 
doc strings are markdown.

> On Aug 24, 2014, at 6:03 PM, Job van der Zwan  
> wrote:
> 
> Any plans? Discussions on Github worth reading through?
> 
> I personally am really charmed by the godoc approach - could something like 
> that work for Julia? (so figuring out a sensible idiomatic way to document 
> functions and modules that makes the documentation easy to read in plaintext, 
> but also easy to be turned into pretty formatted documentation by tools)
> 
>> On Sunday, 24 August 2014 23:07:30 UTC+2, Ivar Nesje wrote:
>> Not yet.

Re: [julia-users] Is it type-stable?

[Mohammed El-Beltagy]

That is no longer valid under in Julia 0.3. The code only runs 3x faster 
and the type stable code looks a bit more complex. 

On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
>
> I think Michael is suggesting that it would help to have an automated tool 
> to do this. This stuff comes up in subtle ways in complex code. While 
> editing some of the t-SNE code someone posted recently, I found that just 
> removing a single type-unstable call to maximum made the code 100x faster. 
> It was a bit too complex to read through code_typed or anything else, so 
> automated tools would make a big difference.
>
>  -- John
>
> On Dec 16, 2013, at 10:46 AM, Ivar Nesje > 
> wrote:
>
> You might use code_typed and look at the output and see if it looks 
> curious.
>
> kl. 19:28:05 UTC+1 mandag 16. desember 2013 skrev John Myles White 
> følgende:
>>
>> I believe Leah was working on a tool to do some of this. 
>>
>>  -- John 
>>
>> On Dec 16, 2013, at 10:22 AM, Michael Fox <415...@gmail.com> wrote: 
>>
>> > After reading about type stability in the FAQ and this excellent post 
>> -- 
>> http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
>>  
>> -- I want to make sure I'm doing it right. 
>> > 
>> > Is there any way to ask Julia if a particular function has achieved 
>> type stability? This would be a good feature for julialint. 
>>
>>
>

Re: [julia-users] Is it type-stable?

[John Myles White]

Indeed, looks like a potential performance regression in 0.3 for sumofsins2:

function sumofsins2(n::Integer)  
r = 0.0  
for i in 1:n  
r += sin(3.4)  
end
return r  
end

 — John

On Aug 24, 2014, at 5:06 PM, Mohammed El-Beltagy  
wrote:

> That is no longer valid under in Julia 0.3. The code only runs 3x faster and 
> the type stable code looks a bit more complex. 
> 
> On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
> I think Michael is suggesting that it would help to have an automated tool to 
> do this. This stuff comes up in subtle ways in complex code. While editing 
> some of the t-SNE code someone posted recently, I found that just removing a 
> single type-unstable call to maximum made the code 100x faster. It was a bit 
> too complex to read through code_typed or anything else, so automated tools 
> would make a big difference.
> 
>  -- John
> 
> On Dec 16, 2013, at 10:46 AM, Ivar Nesje  wrote:
> 
>> You might use code_typed and look at the output and see if it looks curious.
>> 
>> kl. 19:28:05 UTC+1 mandag 16. desember 2013 skrev John Myles White følgende:
>> I believe Leah was working on a tool to do some of this. 
>> 
>>  -- John 
>> 
>> On Dec 16, 2013, at 10:22 AM, Michael Fox <415...@gmail.com> wrote: 
>> 
>> > After reading about type stability in the FAQ and this excellent post -- 
>> > http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
>> >  -- I want to make sure I'm doing it right. 
>> > 
>> > Is there any way to ask Julia if a particular function has achieved type 
>> > stability? This would be a good feature for julialint. 
>> 
> 

Re: [julia-users] Is it type-stable?

[Iain Dunning]

Mohammed, the last post was over 6 months ago - what precisely is not valid?

On Sunday, August 24, 2014 8:06:50 PM UTC-4, Mohammed El-Beltagy wrote:
>
> That is no longer valid under in Julia 0.3. The code only runs 3x faster 
> and the type stable code looks a bit more complex. 
>
> On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
>>
>> I think Michael is suggesting that it would help to have an automated 
>> tool to do this. This stuff comes up in subtle ways in complex code. While 
>> editing some of the t-SNE code someone posted recently, I found that just 
>> removing a single type-unstable call to maximum made the code 100x faster. 
>> It was a bit too complex to read through code_typed or anything else, so 
>> automated tools would make a big difference.
>>
>>  -- John
>>
>> On Dec 16, 2013, at 10:46 AM, Ivar Nesje  wrote:
>>
>> You might use code_typed and look at the output and see if it looks 
>> curious.
>>
>> kl. 19:28:05 UTC+1 mandag 16. desember 2013 skrev John Myles White 
>> følgende:
>>>
>>> I believe Leah was working on a tool to do some of this. 
>>>
>>>  -- John 
>>>
>>> On Dec 16, 2013, at 10:22 AM, Michael Fox <415...@gmail.com> wrote: 
>>>
>>> > After reading about type stability in the FAQ and this excellent post 
>>> -- 
>>> http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
>>>  
>>> -- I want to make sure I'm doing it right. 
>>> > 
>>> > Is there any way to ask Julia if a particular function has achieved 
>>> type stability? This would be a good feature for julialint. 
>>>
>>>
>>

Re: [julia-users] Is it type-stable?

[Mohammed El-Beltagy]

The blog post by John 
http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
demonstrated that massive improvements can be realized my having type 
stable code. The reasoning was the type stable code, resulted in much more  
efficient compiler generated code. This was very nicely empirically 
demonstrated. It worked for Julia 0.2, but now it does not for 0.3.  

On Monday, August 25, 2014 3:19:34 AM UTC+3, Iain Dunning wrote:
>
> Mohammed, the last post was over 6 months ago - what precisely is not 
> valid?
>
> On Sunday, August 24, 2014 8:06:50 PM UTC-4, Mohammed El-Beltagy wrote:
>>
>> That is no longer valid under in Julia 0.3. The code only runs 3x faster 
>> and the type stable code looks a bit more complex. 
>>
>> On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
>>>
>>> I think Michael is suggesting that it would help to have an automated 
>>> tool to do this. This stuff comes up in subtle ways in complex code. While 
>>> editing some of the t-SNE code someone posted recently, I found that just 
>>> removing a single type-unstable call to maximum made the code 100x faster. 
>>> It was a bit too complex to read through code_typed or anything else, so 
>>> automated tools would make a big difference.
>>>
>>>  -- John
>>>
>>> On Dec 16, 2013, at 10:46 AM, Ivar Nesje  wrote:
>>>
>>> You might use code_typed and look at the output and see if it looks 
>>> curious.
>>>
>>> kl. 19:28:05 UTC+1 mandag 16. desember 2013 skrev John Myles White 
>>> følgende:

 I believe Leah was working on a tool to do some of this. 

  -- John 

 On Dec 16, 2013, at 10:22 AM, Michael Fox <415...@gmail.com> wrote: 

 > After reading about type stability in the FAQ and this excellent post 
 -- 
 http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
  
 -- I want to make sure I'm doing it right. 
 > 
 > Is there any way to ask Julia if a particular function has achieved 
 type stability? This would be a good feature for julialint. 


>>>

Re: [julia-users] Is it type-stable?

[John Myles White]

Yes, I believe there’s been a compiler regression for this specific example.

 — John

On Aug 24, 2014, at 5:54 PM, Mohammed El-Beltagy  
wrote:

> The blog post by John 
> http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
> demonstrated that massive improvements can be realized my having type stable 
> code. The reasoning was the type stable code, resulted in much more  
> efficient compiler generated code. This was very nicely empirically 
> demonstrated. It worked for Julia 0.2, but now it does not for 0.3.  
> 
> On Monday, August 25, 2014 3:19:34 AM UTC+3, Iain Dunning wrote:
> Mohammed, the last post was over 6 months ago - what precisely is not valid?
> 
> On Sunday, August 24, 2014 8:06:50 PM UTC-4, Mohammed El-Beltagy wrote:
> That is no longer valid under in Julia 0.3. The code only runs 3x faster and 
> the type stable code looks a bit more complex. 
> 
> On Monday, December 16, 2013 8:48:40 PM UTC+2, John Myles White wrote:
> I think Michael is suggesting that it would help to have an automated tool to 
> do this. This stuff comes up in subtle ways in complex code. While editing 
> some of the t-SNE code someone posted recently, I found that just removing a 
> single type-unstable call to maximum made the code 100x faster. It was a bit 
> too complex to read through code_typed or anything else, so automated tools 
> would make a big difference.
> 
>  -- John
> 
> On Dec 16, 2013, at 10:46 AM, Ivar Nesje  wrote:
> 
>> You might use code_typed and look at the output and see if it looks curious.
>> 
>> kl. 19:28:05 UTC+1 mandag 16. desember 2013 skrev John Myles White følgende:
>> I believe Leah was working on a tool to do some of this. 
>> 
>>  -- John 
>> 
>> On Dec 16, 2013, at 10:22 AM, Michael Fox <415...@gmail.com> wrote: 
>> 
>> > After reading about type stability in the FAQ and this excellent post -- 
>> > http://www.johnmyleswhite.com/notebook/2013/12/06/writing-type-stable-code-in-julia/
>> >  -- I want to make sure I'm doing it right. 
>> > 
>> > Is there any way to ask Julia if a particular function has achieved type 
>> > stability? This would be a good feature for julialint. 
>> 
> 

[julia-users] Re: Announcement: Playground.jl

[Rory Finnegan]

Awesome! Thanks for letting my know it works on Mac.  I'll see if I can fix 
that warning message.

On Sunday, 24 August 2014 11:59:47 UTC-4, Adam Smith wrote:
>
> Nice work! It seems to work on a Mac as well. There is a warning that gets 
> printed every time you use the playground command, but it still seems to 
> work.
>
> $ playground . activate
> Warning: requiring "Playground" did not define a corresponding module.
> playground> echo "foo"
> foo
> playground> julia
>_
>_   _ _(_)_ |  A fresh approach to technical computing
>   (_) | (_) (_)|  Documentation: http://docs.julialang.org
>_ _   _| |_  __ _   |  Type "help()" for help.
>   | | | | | | |/ _` |  |
>   | | |_| | | | (_| |  |  Version 0.3.0 (2014-08-20 20:43 UTC)
>  _/ |\__'_|_|_|\__'_|  |  Official http://julialang.org/ release
> |__/   |  x86_64-apple-darwin13.3.0
>
> julia> using Gadfly
> ERROR: Gadfly not found
>  in require at loading.jl:47
>
> julia>
>
> playground> exit
>
>
>
> On Saturday, August 23, 2014 4:01:45 PM UTC-4, Rory Finnegan wrote:
>>
>> Hi everyone,
>>
>> I've published my Playground.jl 
>>  package to create julia 
>> sandboxes like python virtual environments, if anyone wants to give it a 
>> try.  So far I've tested it on Funtoo and Linux Mint, but I'm looking for 
>> people to try it out on other platforms (like Windows and OSX).
>>
>> Cheers,
>> Rory
>>
>

[julia-users] Re: Function with state calling stateless version, or the other way around?

[Jay Kickliter]

Thanks. I have to admit I really don't understand Julia's tasks. Can you 
explain how they could be used in this context?

As an experiment today, I did reverse most of the functions, and the code 
is much cleaner now. I think I may have only needed to write down the 
problem (my original post) to see the solution.

On Saturday, August 23, 2014 8:27:43 PM UTC-6, Abe Schneider wrote:
>
> It may be worth having both around. If you wrap the stateless versions 
> with something that keeps state, users get the choice of which one to use.
>
> For example, you could use Tasks to create generators that produce the 
> next value. The Task will keep the state for you, which allows you to keep 
> your DSP code simple and stateless.
>

[julia-users] ANN: BayesNets.jl

[Mykel Kochenderfer]

BayesNets.jl : This library supports 
representation, inference, and learning in Bayesian networks. See the 
documentation here 
.
 
Only some very basic algorithms have been implemented so far, but I look 
forward to working with others who would like to contribute.

[julia-users] PyPlot problems, no GUI backend for matplotlib

[Stelios Georgiou]

Hello, 

I can't get PyPlot to work in julia or IJulia notebook, this is the error I 
get. When I test matplotlib also I get no visuals. I can save images and 
view them after but cannot see live figures. Do you have any ideas? 

( I use homebrew python, with packages installed with pip ) 
( Julia is installed from standalone .dmg installer )  
( Using OS X 10.9 )

*julia> *
*using PyPlot*
*WARNING: No working GUI backend found for matplotlib.*

*INFO: Loading help data...*
ipython (2.2.0)

matplotlib (1.3.1)

[julia-users] Re: ANN: PGF/TikZ packages

[Mykel Kochenderfer]

@Kaj Wiik: Thanks for letting me know of this issue. It is fixed with 
v0.0.2 of TikzPictures. Feel free to file an issue if you encounter any 
other trouble.