Hello all,
A little challenge that's been bothering me today.
The docs for std.algorithm give an illustration of its use to calculate mean and
standard deviation in a single pass:
// Compute sum and sum of squares in one pass
r = reduce!(a + b, a + b * b)(tuple(0.0, 0.0), a);
On 2013-02-13 14:26, Marco Leise wrote:
template Julia(TReal)
{
struct ComplexStruct
{
float r;
float i;
...
Why aren't r and i of type TReal?
On 02/13/2013 02:26 PM, Marco Leise wrote:
You get both, 50% more speed and more precision! It is a
win-win situation. Also take a look at Phobos' std.math that
returns real everywhere.
I have to say, it's not been my experience that using real improves speed.
Exactly what optimizations are
On 02/13/2013 02:26 PM, Marco Leise wrote:
I compiled with LDC2 and these are the results:
D code serial with dimension 32768 ...
using floats Total time: 13.399 [sec]
using doubles Total time: 9.429 [sec]
using reals Total time: 8.909 [sec] // - !!!
You get both, 50% more speed
... where k represents the index count 1, 2, 3, ... However,
it's not evident to me how you could get reduce() to know this
counting value.
You would use zip and sequence to add indices to x, like this:
reduce!reducer(initial, zip(x, sequence!n))
Where calculating Q[k] is concerned, you
Am Wed, 13 Feb 2013 14:44:36 +0100
schrieb FG h...@fgda.pl:
On 2013-02-13 14:26, Marco Leise wrote:
template Julia(TReal)
{
struct ComplexStruct
{
float r;
float i;
...
Why aren't r and i of type TReal?
They are actual storage in memory,
reduce!reducer(MQ(x.front, 0), zip(x, sequence!n))
A small correction : you would need to use x.drop(1) instead of
x, because the first element of x is only used to compute the
initial value of 1. If you wanted k to have the same meaning as
the one in your formula, you would need to use
Am Wed, 13 Feb 2013 14:48:21 +0100
schrieb Joseph Rushton Wakeling joseph.wakel...@webdrake.net:
On 02/13/2013 02:26 PM, Marco Leise wrote:
You get both, 50% more speed and more precision! It is a
win-win situation. Also take a look at Phobos' std.math that
returns real everywhere.
I
On 02/13/2013 03:29 PM, Marco Leise wrote:
They are actual storage in memory, where every increase in
size hurts.
When I replaced with TReal, it sped things up for double.
On 2013-02-13 14:44, Joseph Rushton Wakeling wrote:
The docs for std.algorithm give an illustration of its use to calculate mean and
standard deviation in a single pass: [...]
However, this formula for standard deviation is one that is well known for being
subject to potentially fatal rounding
jerro:
reduce!reducer(MQ(x.front, 0), zip(x, sequence!n))
A small correction : you would need to use x.drop(1) instead of
x, because the first element of x is only used to compute the
initial value of 1. If you wanted k to have the same meaning as
the one in your formula, you would need to
On 02/13/2013 03:48 PM, FG wrote:
You can use reduce and put the division and subtraction into the reduce itself
to prevent overflows. You also won't end up with jaw-dropping tuples, sorry. :)
float[] a = [10_000.0f, 10_001.0f, 10_002.0f];
auto n = a.length;
auto avg =
Am Wed, 13 Feb 2013 15:00:21 +0100
schrieb Joseph Rushton Wakeling joseph.wakel...@webdrake.net:
Compiling with ldmd2 -O -inline -release on 64-bit Ubuntu, latest from-GitHub
LDC, LLVM 3.2:
D code serial with dimension 32768 ...
using floats Total time: 4.751 [sec]
using
On 02/13/2013 03:48 PM, FG wrote:
Typical thing with examples - they try to be terse and show off a mechanism like
reduce, without going into too much details and hence are unusable IRL.
My favourite -- in the tutorial for a really serious piece of scientific code
written in C:
int n =
On 02/13/2013 03:56 PM, Marco Leise wrote:
Ok, I get pretty much the same numbers as before with:
ldmd2 -O -inline -release
It's even a bit faster than my lng command line.
My experience has been that the higher -O values of ldc don't do much, but of
course, that's going to vary
Am Wed, 13 Feb 2013 15:45:13 +0100
schrieb Joseph Rushton Wakeling joseph.wakel...@webdrake.net:
On 02/13/2013 03:29 PM, Marco Leise wrote:
They are actual storage in memory, where every increase in
size hurts.
When I replaced with TReal, it sped things up for double.
Give me that stuff,
See enumerate():
http://d.puremagic.com/issues/show_bug.cgi?id=5550
I like this enumerate() thing. Is there any particular reason why
it isn't in phobos, or is it just that no one has added it yet?
I think with enumerate it becomes:
MQ(x.front, 0).enumerate(1).reduce!reducer()
I think the
Good point about choosing the right type of floating point numbers.
Conclusion: when there's enough space, always pick double over float.
Tested with GDC in win64. floats: 16.0s / doubles: 14.1s / reals: 11.2s.
I thought to myself: cool, I almost beat the 13.4s I got with C++, until I
changed
Am Wed, 13 Feb 2013 15:45:13 +0100
schrieb Joseph Rushton Wakeling joseph.wakel...@webdrake.net:
On 02/13/2013 03:29 PM, Marco Leise wrote:
They are actual storage in memory, where every increase in
size hurts.
When I replaced with TReal, it sped things up for double.
Oh this gets even
On 2013-02-13 16:26, Marco Leise wrote:
I'd still bet a dollar that with an array of values floats would
outperform doubles, when cache misses happen. (E.g. more or
less random memory access.)
I'll play it safe and only bet my opDollar. :)
On 02/13/2013 04:17 PM, FG wrote:
Good point about choosing the right type of floating point numbers.
Conclusion: when there's enough space, always pick double over float.
Tested with GDC in win64. floats: 16.0s / doubles: 14.1s / reals: 11.2s.
I thought to myself: cool, I almost beat the 13.4s
On Saturday, 2 February 2013 at 15:39:00 UTC, Namespace wrote:
I've never come across Appenders before. Could you please
explain them a little bit, and what each call in your modified
code does?
Thanks,
Josh
http://dlang.org/phobos/std_array.html#.Appender
And read Era's post.
Why
If some of you has a little of time to review code, I have
converted the very fast C memory mapped version of Ordered
words to D:
http://rosettacode.org/wiki/Ordered_words#Mmap
http://rosettacode.org/wiki/Ordered_words#Memory_Mapped_Version
The C version contains several pointers that get
Andrea Fontana:
Why Appender has no ~ operator itself?
auto app = appender!string();
app.put(a); // why not app ~= a?
Now it's in GIT head.
Bye,
bearophile
Am Wed, 13 Feb 2013 16:17:12 +0100
schrieb FG h...@fgda.pl:
Good point about choosing the right type of floating point numbers.
Conclusion: when there's enough space, always pick double over float.
Tested with GDC in win64. floats: 16.0s / doubles: 14.1s / reals: 11.2s.
I thought to myself:
On 02/13/2013 04:41 PM, Joseph Rushton Wakeling wrote:
On 02/13/2013 04:17 PM, FG wrote:
Good point about choosing the right type of floating point numbers.
Conclusion: when there's enough space, always pick double over float.
Tested with GDC in win64. floats: 16.0s / doubles: 14.1s / reals:
Hi everyone.
I'm new to D, coming from a Java/Python background. I've been
reading the excellent The D Programming Language book, and want
to now start playing around with D.
I'm having an issue with importing. When I have the following
file:
file ~/src/sample.d: =
On Tuesday, 12 February 2013 at 22:06:48 UTC, monarch_dodra wrote:
On Tuesday, 12 February 2013 at 21:41:14 UTC, bioinfornatics
wrote:
Some time fastq are comressed to gz bz2 or xz as that is often
a
huge file.
Maybe we need keep in mind this early in developement and use
std.zlib
While
On Wed, Feb 13, 2013 at 06:17:51PM +0100, Korey Peters wrote:
[...]
...and at the terminal:
me@ubuntu:~/src$ rdmd sample.d
/tmp/.rdmd-1000/rdmd-sample.d-94E53075E2E84D963426A11F2B81FDED/objs/sample.o:
In function `_Dmain':
sample.d:(.text._Dmain+0xa): undefined reference to
Thanks for your response, H.S.Teoh.
On Wednesday, 13 February 2013 at 17:47:09 UTC, H. S. Teoh wrote:
You need to specify both files on the command line, so that the
linker
knows where to find everything:
rdmd sample.d sample_a.d
Running this from the command line produces
Am Wed, 13 Feb 2013 18:10:47 +0100
schrieb Joseph Rushton Wakeling joseph.wakel...@webdrake.net:
Just to update on times. I was running another large job at the same time as
doing all these tests, so there was some slowdown. Current results are:
-- with g++ -O3 and using double rather
On Wed, Feb 13, 2013 at 06:57:52PM +0100, Korey Peters wrote:
Thanks for your response, H.S.Teoh.
On Wednesday, 13 February 2013 at 17:47:09 UTC, H. S. Teoh wrote:
You need to specify both files on the command line, so that the
linker
knows where to find everything:
rdmd sample.d
When you are comparing LDC and GDC, you should either use
-mcpu=generic for ldc or -march=native for GDC, because their
default targets are different. GDC will produce code that works
on most x86_64 (if you are on a x86_64 system) CPUs by default,
and LDC targets the host CPU. But this does
Hmm.
I moved my two sample files from ~/path/to/where/I/was/working
to ~/ and the import worked.
This makes me suspect a permissions issue. I'll carry on working
in ~/ for now, until I sort my stupidity out!
Thanks for your help.
On Wednesday, 13 February 2013 at 18:42:51 UTC, H. S. Teoh wrote:
On Wed, Feb 13, 2013 at 06:57:52PM +0100, Korey Peters wrote:
Thanks for your response, H.S.Teoh.
On Wednesday, 13 February 2013 at 17:47:09 UTC, H. S. Teoh
wrote:
You need to specify both files on the command line, so that
On 2013-02-13 18:39, monarch_dodra wrote:
In any case, I am now parsing the 6Gig packed into 1.5Gig in about 53 seconds
(down from 61). I also tried doing a dual-threaded approach (1 thread to unzip,
1 thread to parse), but again, the actual *parse* phase is so ridiculously fast,
that it changes
On Wednesday, 13 February 2013 at 19:33:00 UTC, jerro wrote:
This solves the issue:
rdmd --force sample
Hi jerro,
That definitely helped. There's still some things I haven't
figured out yet about D's importing, but this has got me going.
Thank you.
Hi!
I'm learning to use D collections properly, and I'm looking for a
sorted data structure with logarithmic access time (i.e., a
binary search tree will do, but a hash table would not help). As
far as I can see, std.container.RedBlackTree is exactly what I
need. However, I am not sure if I
P.S. More on C++ version:
Personally, I don't see why at all we should call the copy
constructor more than once per element. I mean, if we intend to
build a generic data structure, we sure need an internal node
object with some extra bytes (internal references and counters)
per each element,
On Thursday, February 14, 2013 00:33:26 Ivan Kazmenko wrote:
P.S. More on C++ version:
Personally, I don't see why at all we should call the copy
constructor more than once per element. I mean, if we intend to
build a generic data structure, we sure need an internal node
object with some
On Wed, 13 Feb 2013, jerro wrote:
I created self-contained sample.d, ran it with rdmd, then moved class A to
sample_a.d and tried to run it with rdmd again. I could reproduce the issue
that way. It seems that rdmd caches the dependency list. Using --chatty flag
confirms that rdmd does not run
You can check if disabling the GC just before the insert process
improves the performance. You may see 3x performance improvement.
Disabling is safe provided you re-enable, this can be done
reliably with scope(exit) or something similar.
import core.memory;
// ...
void main ()
{
Should the following work?
import std.traits;
mixin template Foo()
{
void foo(T)(T t) if (isSomeString!T) {}
}
class A
{
void foo()(int i){}
mixin Foo;
}
void main()
{
auto a = new A;
a.foo(hello);
}
Error: template hello.A.foo does not match any function template
On 2013-02-14 01:09, Rob T wrote:
You can check if disabling the GC just before the insert process improves the
performance. You may see 3x performance improvement. Disabling is safe provided
you re-enable, this can be done reliably with scope(exit) or something similar.
How did you know? It
Please file a bug report on this.
Done.
On Wed, 13 Feb 2013 18:22:02 -0500, Ivan Kazmenko ga...@mail.ru wrote:
Hi!
I'm learning to use D collections properly, and I'm looking for a sorted
data structure with logarithmic access time (i.e., a binary search tree
will do, but a hash table would not help). As far as I can see,
On Wed, Feb 13, 2013 at 03:53:51PM -0800, Brad Roberts wrote:
On Wed, 13 Feb 2013, jerro wrote:
I created self-contained sample.d, ran it with rdmd, then moved
class A to sample_a.d and tried to run it with rdmd again. I could
reproduce the issue that way. It seems that rdmd caches the
And a related question:
class A
{
void foo(int i){}
void foo(Tuple!(int) i){}
}
class B: A
{
override void foo(int i){}
}
int main()
{
auto b = new B;
b.foo(tuple(5));
}
This fails to compile. Why can't B use A's tuple overload of
foo()? If I do this:
class B: A
{
On Wednesday, 13 February 2013 at 23:22:03 UTC, Ivan Kazmenko
wrote:
Hi!
-
Ivan Kazmenko.
Keep in mind that C++ and D have very different philosophies
regarding copy construction.
C++ has strong ownership, so for example, whenever you copy a
string/vector, or pass it by value, the
On Thursday, 14 February 2013 at 00:29:51 UTC, cal wrote:
Should the following work?
import std.traits;
mixin template Foo()
{
void foo(T)(T t) if (isSomeString!T) {}
}
class A
{
void foo()(int i){}
mixin Foo;
}
void main()
{
auto a = new A;
a.foo(hello);
}
Error:
On Thursday, 14 February 2013 at 05:49:33 UTC, cal wrote:
And a related question:
class A
{
void foo(int i){}
void foo(Tuple!(int) i){}
}
class B: A
{
override void foo(int i){}
}
int main()
{
auto b = new B;
b.foo(tuple(5));
}
This fails to compile. Why can't B use A's
On Thursday, 14 February 2013 at 00:25:15 UTC, FG wrote:
On 2013-02-14 01:09, Rob T wrote:
You can check if disabling the GC just before the insert
process improves the
performance. You may see 3x performance improvement. Disabling
is safe provided
you re-enable, this can be done reliably with
52 matches
Mail list logo
