How are numbers converted to strings in Julia? Specifically, where in the
Julia source is this conversion performed?
I ask because my x86 assembly implementation of 64-bit integer to string
conversion is on average about 1.5x slower than Julia. Either Julia is
doing something very smart, or
Hi i followed all the steps required for downloading juno through light
table. However, Juno doesn't appear under plug ins. Can anyone assist me on
how I can remedy this or perhaps download Juno in another way
The Wikipedia article was redirected quite recently in fact, and after
reviewing the edit history and reading the comments on the talk page
https://en.wikipedia.org/wiki/Talk:Unum_(number_format), I can say (as an
experienced Wikipedia editor) that the case for restoring the article can
be made
As your code is not a reproducible example, there's no way to say what's
going wrong exactly. The error message is telling you that you are trying
to access an array out of bounds, so it would be necessary to know the
sizes of the arrays in your code to find the error.
Two other things:
-
Yes, sure.
On Friday, July 31, 2015 at 2:02:25 PM UTC+2, Tim Holy wrote:
I have a demo to answer your question, but I'd like to simply add it to
the
documentation on SharedArrays. May I use some of your code in writing up
the
demo? (MIT license, see
I've been translating some Matlab code for calculating a
spectral optimisation function that uses Matlab's zpk() function and I
can't find an equivalent in Julia packages. The DSP.jl looks like its a
start on a well appointed package but is a bit thin at this stage. Is there
anything out there
Hi,
I'm processing the data from an oscilloscope. The natural way to
encapsulate them is to have a structure in a form of
type Wfm
xorigin
xinc
len
ydata
properties::Dict{ASCIIString,Any}
end
, where X represents time and Y data are equidistantly spaced in time. To
access a
I added some info to the readme at https://github.com/tbreloff/Unums.jl. I
talk a little bit about how I'm intending to build the package, the
available types, etc. There is also a stub issue for continuing the
discussion of how unums fit into the world of numerical analysis:
In a FLAC (as in Free Lossless Audio Codec) package for version 0.4 I used
@enum to create a type to pass to calls to the C library libflac.
Metadata for a flac stream object.
Used in both encoding and decoding of a stream.
Each type of metadata object contains an indicator
of its `typ`, an
Hey Tom,
Well, I could change the setting to anyone with the link can edit - we
risk vandalism in that case, but as long as we keep the document link to
here the risk is minimal.
On Friday, 31 July 2015 15:43:06 UTC+2, Tom Breloff wrote:
I added some info to the readme at
Hi.
Regarding your questions, I am also having the same problems and no luck
with those {Any} and SharedArrays. Also,
I am having the same problems with those crashes, so I have to restart
Julia after finishing any parallel code.
I am looking forward to see Tim's comments about those type
Thanks for your help Tim !
On Friday, July 31, 2015 at 8:30:44 AM UTC-3, Tim Holy wrote:
If MyModule.jl is on your LOAD_PATH,
@everywhere import MyModule
should work. You can add
push!(LOAD_PATH,/my/code/repository)
to your .juliarc.jl file.
This has been deprecated
Depending how your filter is represented, zpkdata from Control.jl might be
a good choice - i'm not using it, but the code looks promising.
On Thursday, July 30, 2015 at 3:54:39 PM UTC-4, Jason Merrill wrote:
* Wrath of Kahan, 2
In the unum computation, squaring operations are fused with squareu, but
float Interval calculations are not fused. I also believe the check for the
ubounds containing zero in e[z] is erroneous in a
To answer my own question
for typ in instances(MetadataType) ... end
On Friday, July 31, 2015 at 9:41:13 AM UTC-5, Douglas Bates wrote:
In a FLAC (as in Free Lossless Audio Codec) package for version 0.4 I used
@enum to create a type to pass to calls to the C library libflac.
Metadata
A github wiki in the Unums.jl package would seem ideal. You get the anyone
can edit feature, with accountability of who made each edit (github wikis
are git repos, and to make edits people need to have a github account) and
easy reversal of eventual bad changes.
On Friday, July 31, 2015 at
I just now learned about this discussion and I see quite a few messages I
need to reply to.
I am very excited about a Julia version of unum arithmetic, and it does
seem like the ideal language for it. Alan Edelman, Deepak Vinchhi, and
Viral Shah proposed to create it with funding from A*STAR
There is no question that getting rigorous, tight bounds is going to cost
something compared to the guesswork of conventional floating point.
Interval arithmetic has failed for the last fifty years because it requires
a LOT of work and expertise to keep the bounds from exploding; the ubox
The expression
e = :(xc = b)
returns in Julia 0.3
:(x c = b)
with e.args[2] equal to :b
but in Julia 0.4
:(x c = begin # none, line 1:
b
end)
with e.args[2] equal to quote b end.
I relied on such expressions to denote instrumental variables in formulas
for a
Unum format specifically eliminates NaN payloads as one of the worst
decisions made in the IEEE 754 standard. They break just about every rule
there is about good engineering. There are two kinds of NaN in unum math:
quiet and signaling. The other bit patterns are used to represent actual
Ah, that sounds good. So I should just fork Tom's package, copy the
document we have so far into the wiki, then make a pull request?
On Friday, 31 July 2015 17:01:57 UTC+2, Waldir Pimenta wrote:
A github wiki in the Unums.jl package would seem ideal. You get the
anyone can edit feature, with
I'd recommend not using FloatRange for this task. Instead, I'd create a
new Interval type that doesn't have discrete steps and instead represents
all real numbers between (and including) its endpoints. You can even use
an unused operator (..) to allow easy construction of intervals. Naively,
What IEEE 754 did with negative zero is a half-baked attempt to represent
inexactness. Consider, for example, that the square root of negative zero
is defined to be negative zero. Unums have the same representation of zero
as floats in that the sign bit can be 0 or 1, but the sign bit is
Zenna,
If unums are used without the ubox method and some of the other techniques
described in the book (like tightest-possible evaluation in the *g*-layer
for a well-defined set of functions), they will indeed fall prey to two of
the main problems of interval arithmetic: the dependency
Be careful. I, too, hoped that it would be sufficient to change the last
fraction bit to a ubit. But the relative error changes with every
calculation, so you *have* to let the fraction length float and be another
annotation field. And for more subtle reasons, you need to be able to vary
the
The chapter Permission to Guess explains how to round unums. The
guess function replaces an inexact unum with an exact one, either the one
closest to the midpoint if there is at least one more bit of fraction
precision available, or one of the endpoints if the ULP size is already as
small as
On Friday, July 31, 2015 at 12:51:57 PM UTC-4, Jason Merrill wrote:
On Thursday, July 30, 2015 at 3:54:39 PM UTC-4, Jason Merrill wrote:
* Wrath of Kahan, 2
In the unum computation, squaring operations are fused with squareu, but
float Interval calculations are not fused. I also believe
Many people are judging the merits of unum arithmetic based on an hour-long
PowerPoint presentation. I really wish they wouldn't do that. There's no
way that can get all the key concepts across in anything less than a
400-page book that anticipates the questions and criticisms and answers
It would be wonderful if someone else would create a super-concise
explanation of unums! Thank you, thank you! I can't seem to do it. I must
be getting long-winded in my old age.
I looked a
https://github.com/tbreloff/Unums.jl
and saw that it erroneously says the sign bit field can have
On Fri, Jul 31, 2015 at 2:46 PM, Diego Javier Zea diego...@gmail.com wrote:
Hi! I don't know how to generate parameter list of functions in order to
avoid a lot repetitive code:
julia for (param, special) in [ (:(),:()), (:(, y),:(+ y)) ]
@eval begin
function f(x
In the Wrath of Kahan, 2 example, the input to e[z] might be intended to
be a unum, but in the actual examples used in the book, q[x] returns a
ubound, not a unum, so I stand by my analysis that the presentation is
incorrect in an important way.
As a plug for Julia, it should be a lot easier
Hi! I don't know how to generate parameter list of functions in order to
avoid a lot repetitive code:
julia for (param, special) in [ (:(),:()), (:(, y),:(+ y)) ]
@eval begin
function f(x $(param), z)
x $(special) + z
end
end
end
The wiki in github projects is a separate git repository from the code one,
so I'm not sure you can add it to the upstream repo with a PR like that.
The easiest way is for Tom to activate the wiki in the repo settings, then
anyone with a github account can add and edit pages directly in the web
Thanks for the catch about the signbit... already changed.
On Fri, Jul 31, 2015 at 2:30 PM, John Gustafson johngustaf...@earthlink.net
wrote:
It would be wonderful if someone else would create a super-concise
explanation of unums! Thank you, thank you! I can't seem to do it. I must
be
Actually, Jason, the book went through intense peer review repeatedly for
over a year before it hit the shelves. Horst Simon, the series editor,
vetted the manuscript and made sure William Kahan saw it as well. Kahan,
the guy behind the IEEE 754 Standard for floats and a Turing Award
Here is how you can represent the square root of 2 with a finite number of
symbols: 1.414…
The … means There are more decimals after the last one shown, not all
zeros and not all nines. If the trailing digits were all zeros, we would
instead write 1.414 and it would be exact. If the trailing
Thanks for the detailed reading, Jason. If the input is a unum, then the
only way for it to contain zero is to be exactly zero. And that was the way
the example was used. If the input to the function is a ubound, then you
are correct that the test would need to be rewritten.
On Friday, July
To me, there are 3 main criteria that I'm comparing floats vs unums:
1) Speed
2) Elegance
3) Correctness
Right now Floats win #1 and Unums win #2 and #3 (IMO). If unums are
optimized in hardware someday, I believe they will also win #1, especially
with the addition of some summary bits that can
It is indeed possible to do a shortcut to something very similar to a unum
representation, and Ulrich Kulisch has suggested this: Every number is a
pair of 32-bit floats where the last bit of the fraction is the ubit. The
two floats represent the endpoints of an interval. So there is no
Guys, this reminds me: There used to be a Wikipedia page on Unum
(arithmetic), but it was taken down for some reason and now searches just
direct to my Wikipedia page. Maybe it's time to revive it. Then we could
start building a concise explanation there.
On Friday, July 31, 2015 at 8:01:57 AM
The quick explanation for why unums/ubounds don't diverge like intervals
do is that when the result gets wider, it divides like an amoeba into a
collection of unums that alternate between exact and inexact values that
tile the interval
It seems like this will incur another problem:
Zenna, I had been wondering if there might be something in the tiling
nature that makes unums particularly well suited to solving problems posed
on higher dimensional surfaces?
On Friday, July 31, 2015 at 3:54:29 PM UTC-4, Zenna Tavares wrote:
The quick explanation for why unums/ubounds
What would be the first problem you address with this made hardware?
On Friday, July 31, 2015 at 3:39:01 PM UTC-4, John Gustafson wrote:
I discuss this in the book; there have to be strict bounds on how long a
computation remains in the *g*-layer (fused) or people would dump their
entire
I discuss this in the book; there have to be strict bounds on how long a
computation remains in the *g*-layer (fused) or people would dump their
entire calculation in there. I think i got most of the fused operations
that make sense, and I pointed out some that do not make sense. It is key
I'll try to answer this concisely.
Unum addition and multiplication are associative; they give bitwise
identical results if performed in the *g*-layer (fused operations) and if
not fused, will produce the same answer but possibly with different
accuracy. When (*a* + *b*) + *c* is not identical
Tom, I very much hope to soon be spending full time on unums. I have been
offered an appointment at A*STAR to come to Singapore for a year and
develop unum arithmetic. So absolutely I hope to help with the Julia
implementation once I've made that move! It was supposed to start next
month; just
Existing methods for evaluation of polynomials do not produce tight bounds.
There is indeed a mountain of papers written about doing it, all with the
flavor of our method sucks less because it usually gets somewhat tighter
bound, though still not as tight as possible. I was surprised to learn
Hello,
I am fairly new to Julia, but it has already been a fantastic help for my
work so far. I just upgraded to Windows 10 and installed Julia and Juno.
Unfortunately Juno couldn't connect with Julia. I got the error message:
Couldn't connect to Julia
INFO: Couldn't find Jewel package,
I have been trying to include a gist in an Escher web app:
function main(window)
vbox(
plaintext(Gist test),
Elem(:script, src = ~/.julia/v0.3/Patchwork/runtime/build.js),
Elem(:script,
display = inline,
src =
The wiki should be active now.
John: welcome to the thread! I hope you'll find the time to review the
implementation I'm designing as well as contribute to the wiki.
On Fri, Jul 31, 2015 at 1:49 PM, John Gustafson johngustaf...@earthlink.net
wrote:
Here is how you can represent the square
I would probably attempt an n-body calculation first. That would allow us
to check the hypothesis that uboxes form ellipsoidal clouds as the
computation progresses, which is why Kahan came up with a form of
arithmetic based on hyperellipsoids.
On Friday, July 31, 2015 at 12:51:21 PM UTC-7,
I have a demo to answer your question, but I'd like to simply add it to the
documentation on SharedArrays. May I use some of your code in writing up the
demo? (MIT license, see
https://github.com/JuliaLang/julia/blob/master/CONTRIBUTING.md#improving-documentation)
--Tim
On Thursday, July 30,
Ok, this did solve my problem.
Now I will proceed with trying to connect a big c++ project with julia and
see what happens there, in the very optimistic case where I proceed without
any problems I will let you know just for reference, otherwise I will
return with
the new errors and mistakes I
The output is C:\\Users\\Serge\\.Julia\\v0.3
ENV[HOME] gave me an error message (ERROR: key not found: HOME in
getindex at env.jl:57)
On Friday, 31 July 2015 23:18:17 UTC+1, Isaiah wrote:
What is the output of (in the REPL only, not Juno):
Pkg.dir()
ENV[HOME]
On Fri, Jul 31, 2015 at 3:52
Dear Tim:
Yes, I am.
Thank you.
There is a lot to learn about Julia, but it is all worth it.
Best wishes,
Jim
On 07/31/2015 07:09 PM, Tim Holy wrote:
Are you looking for permutedims?
--Tim
On Friday, July 31, 2015 04:05:09 PM jamesmna...@gmail.com wrote:
Hi All:
I have a 3-d array
Hi All:
I have a 3-d array (Float64), call it q, with dimensions (d1, d2, d3)
that needs to be to changed to an array, call it Q, with dimensions (d2,
d3, d1) to perform several computations.
I have tried using the reshape command but that does not respect the first
dimension (row) of the q
Are you looking for permutedims?
--Tim
On Friday, July 31, 2015 04:05:09 PM jamesmna...@gmail.com wrote:
Hi All:
I have a 3-d array (Float64), call it q, with dimensions (d1, d2, d3)
that needs to be to changed to an array, call it Q, with dimensions (d2,
d3, d1) to perform several
I had the same error after upgrading to Windows 10.
Uninstall Julia and afterwards go into file explorer and physically delete
the .Julia folder and all its contents. I also deleted the .julia_history
file. Next reinstall Julia and you should be good to go...Arch
On Friday, July 31, 2015 at
Great. I did it and it is working now. Thank you so much!
On Saturday, 1 August 2015 00:17:21 UTC+1, Arch Call wrote:
I had the same error after upgrading to Windows 10.
Uninstall Julia and afterwards go into file explorer and physically delete
the .Julia folder and all its contents. I
Speaking of going out on a limb: are you aware of Mark Kikgard's work on GPU
accelerated path rendering?
http://www.slideshare.net/mobile/Mark_Kilgard/gtc-2014-nvidia-path-rendering
There is obvious *thematic* overlap, with the promise of faster, more accurate
2D graphics using LESS power.
I know it looks scary, but these days, supercomputers can do tens of
petaflops of calculation and have hundreds of millions of processors. And
other than LINPACK and some trivial stuff (like collecting Monte Carlo
data), very few applications have been scaled to run on hundreds of
millions of
What is the output of (in the REPL only, not Juno):
Pkg.dir()
ENV[HOME]
On Fri, Jul 31, 2015 at 3:52 PM, Serge Santos serge.san...@gmail.com
wrote:
Hello,
I am fairly new to Julia, but it has already been a fantastic help for my
work so far. I just upgraded to Windows 10 and installed Julia
On Fri, Jul 31 2015, Christopher Fisher fishe...@miamioh.edu
wrote:
I was wondering if there is a function for enumerating all of
the permutations of size m from n elements, with repetitions
allowed. For example, 3 permutations of [1 0] would be [ 1 1
1;1 1 0;1 0 1;0 1 1;1 0 0; 0 1 0;0 0
On Fri, Jul 31 2015, Christopher Fisher fishe...@miamioh.edu
wrote:
Thank you Tamas. I added return a and restructured the code (I
think some formatting was lost when pasting). Based on the
examples I tested, it appears to work. Can you recommend an
efficient method of inputing an array
If MyModule.jl is on your LOAD_PATH,
@everywhere import MyModule
should work. You can add
push!(LOAD_PATH,/my/code/repository)
to your .juliarc.jl file.
This has been deprecated because of precompilation; it was felt that the
string version left it too ambiguous about whether you
Thank you Tamas. I added return a and restructured the code (I think some
formatting was lost when pasting). Based on the examples I tested, it
appears to work. Can you recommend an efficient method of inputing an array
of numbers. For example:
a = [10,12]
p = perm_matrix(a,2)
p = [10 10; 10
66 matches
Mail list logo