<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress.com" -->
<rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	>

<channel>
	<title>cuda &amp;laquo; WordPress.com Tag Feed</title>
	<link>http://en.wordpress.com/tag/cuda/</link>
	<description>Feed of posts on WordPress.com tagged "cuda"</description>
	<pubDate>Sun, 29 Nov 2009 03:21:34 +0000</pubDate>

	<generator>http://en.wordpress.com/tags/</generator>
	<language>en</language>

<item>
<title><![CDATA[Intellisense y resaltado (highlight) de CUDA en Visual Studio 2005]]></title>
<link>http://cudagpu.wordpress.com/2009/11/25/intellisense-resaltado-highlight-de-cuda-en-visual-studio-2005/</link>
<pubDate>Wed, 25 Nov 2009 08:50:33 +0000</pubDate>
<dc:creator>cudagpu</dc:creator>
<guid>http://cudagpu.wordpress.com/2009/11/25/intellisense-resaltado-highlight-de-cuda-en-visual-studio-2005/</guid>
<description><![CDATA[Una de las características principales de los entornos de desarrollo es el CTRL+ESPACIO, y el resalt]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Una de las características principales de los entornos de desarrollo es el CTRL+ESPACIO, y el resaltado de sintaxis. Como somos unos perros y no tenemos tampoco una memoria infalible es la técnica perfecta para rellenar automáticamente los nombres de variable y que el entorno te ayude en los parámetros de las funciones.</p>
<p>Todo esto es muy bonito (sobretodo en Netbeans, que lo hace muy bien) , pero cuando estamos editando un fichero CUDA (extensión .cu normalmente) Visual Studio no hace ni caso y no sabe ni es capaz de resaltar ni darnos ayuda con Intellisense.</p>
<p>Pero hay una manera bastante sencilla de enseñar (en realidad engañar) a Visual Studio la sintaxis de CUDA y que trate a estos ficheros como al resto.</p>
<p><strong>Resaltado de sintaxis CUDA en Visual Studio 2005 (<a href="http://macudapfc.wordpress.com/2009/03/05/resaltado-de-la-sintaxis-de-cuda-en-visual-studio-2008/">fuente</a>)</strong></p>
<p>El primer paso es localizar el fichero que incluye las palabras que deberá resaltar Visual Studio, NVIDIA proporciona ya estos datos si hemos instalado toda la SDK de CUDA, en concreto se encuentran en la siguiente dirección:</p>
<p>C:\Documents and Settings\All Users\Datos de programa\NVIDIA Corporation\NVIDIA GPU Computing SDK\C\doc\syntax_highlighting\visual_studio_8\usertype.dat</p>
<p>Debemos copiar este fichero al directorio del IDE de Visual Studio, en concreto:</p>
<p>C:\Archivos de programa\Microsoft Visual Studio 8\Common7\IDE</p>
<p>En el caso de que este fichero ya exista, lo que hay que hacer es añadir el contenido del usertype.dat de CUDA al usertype.dat que ya tenga nuestro IDE.</p>
<p>Ahora abrimos Visual Studio y vamos a: Herramientas / Opciones / Editor de texto / Extensión de archivo. Allí escribimos en la extensión &#8220;cu&#8221; y la agregamos. Con esto Visual Studio ya es capaz de resaltar la mayoría de sintaxis de CUDA.</p>
<p><strong>Habilitar Intellisense para ficheros CUDA en Visual Studio</strong></p>
<p>En el caso de que queramos tener las ayudas de Intellisense (CTRL+ESPACIO) y esas cosas, debemos tocar el registro de Windows para engañar a Visual Studio.</p>
<ul>
<li>Cerrar Visual Studio si está abierto</li>
<li>Inicio / Ejecutar / regedit</li>
<li>Navegamos hasta la ruta: HKEY_CURRENT_USER\Software\Microsoft\VisualStudio\8.0\Languages\Language Services\C/C++</li>
<li>Editamos la clave: NCB Default C/C++ Extensions</li>
<li>Y al final añadimos: ;.cuh;.cu;</li>
<li>A mi me queda finalmente así: .cpp;.cxx;.c;.cc;.h;.hh;.hxx;.hpp;.inl;.tlh;.tli;.cuh;.cu;</li>
</ul>
<p>Ahora Visual Studio se cree que los ficheros .cu y .cuh son en lenguaje C y en la mayoría de los casos será capaz de darnos información sobre los parámetros de las funciones y rellenar variables.</p>
<ul></ul>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[zmarnowane szanse]]></title>
<link>http://evi219.wordpress.com/2009/11/24/zmarnowane-szanse/</link>
<pubDate>Tue, 24 Nov 2009 12:35:52 +0000</pubDate>
<dc:creator>evi219</dc:creator>
<guid>http://evi219.wordpress.com/2009/11/24/zmarnowane-szanse/</guid>
<description><![CDATA[Zawsze trzeba podejmować ryzyko. Tylko wtedy uda nam się pojąć, jak wielkim cudem jest życie, gdy bę]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p style="padding-left:60px;"><em>Zawsze trzeba podejmować ryzyko. Tylko wtedy uda nam się pojąć, jak wielkim cudem jest życie, gdy będziemy gotowi przyjąć niespodzianki, jakie niesie nam los.</em></p>
<p style="padding-left:60px;"><em>Bowiem każdego dnia wraz z dobrodziejstwami słońca Bóg obdarza nas chwilą, która jest w stanie zmienić to wszystko, co jest przyczyną naszych nieszczęść. I każdego dnia udajemy, że nie dostrzegamy tej chwili, że ona wcale nie istnieje. Wmawiamy sobie z uporem, że dzień dzisiejszy podobny jest do wczorajszego i do tego, co ma dopiero nadejść. Ale człowiek uważny na dzień, w którym żyje, bez trudu odkrywa magiczną chwilę. Może być ona ukryta w tej porannej porze, kiedy przekręcamy klucz w zamku, w przestrzeni ciszy, która zapada po wieczerzy, w tysiącach i jednej rzeczy, które wydaja się nam takie same. Ten moment istnieje naprawdę, to chwila, w której spływa na nas cała siła gwiazd i pozwala nam czynić cuda. Tylko niekiedy szczęście bywa darem, najczęściej trzeba o nie walczyć. Magiczna chwila dnia pomaga nam dokonywać zmian, sprawia, iż ruszamy na poszukiwanie naszych marzeń. I choć przyjdzie nam cierpieć, choć pojawią się trudności, to wszystko jest jednak ulotne i nie pozostawi po sobie śladu, a z czasem będziemy mogli spojrzeć wstecz z dumą i wiarą w nas samych.</em></p>
<p style="padding-left:60px;"><em>Biada temu, kto nie podjął ryzyka. Co prawda nie zazna nigdy smaku rozczarowań i utraconych złudzeń, nie będzie cierpiał jak ci, którzy pragną spełnić swoje marzenia, ale kiedy spojrzy za siebie – bowiem zawsze dogania nas przeszłość – usłyszy głos własnego sumienia: „A co uczyniłeś z cudami, którymi Pan Bóg obsiał dni twoje? Co uczyniłeś z talentem, który powierzył ci Mistrz? Zakopałeś te dary głęboko w ziemi, gdyż bałeś się je utracić. I teraz została ci jedynie pewność, że zmarnowałeś własne życie.”</em></p>
<p style="padding-left:60px;"><em>Biada temu, kto usłyszy te słowa. Bo uwierzył w cuda, dopiero gdy magiczne chwile życia odeszły na zawsze.</em></p>
<p style="padding-left:60px;"><em>[Paulo Coelho, Nad brzegiem rzeki Piedry usiadłam i płakałam...]</em></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[CUDA 2.2/2.3 의 Run time 비교]]></title>
<link>http://needq.wordpress.com/2009/11/23/cuda23/</link>
<pubDate>Sun, 22 Nov 2009 15:41:31 +0000</pubDate>
<dc:creator>needq</dc:creator>
<guid>http://needq.wordpress.com/2009/11/23/cuda23/</guid>
<description><![CDATA[이전 포스트에 언급했던 작업결과페이지의 변화로 이제 CUDA MB 의 작업시간도 다른작업들과같이 한눈에 확인할수있게 되었다. 그래서 이번에 이것을 이용해 CUDA 2.2 버전과 2]]></description>
<content:encoded><![CDATA[이전 포스트에 언급했던 작업결과페이지의 변화로 이제 CUDA MB 의 작업시간도 다른작업들과같이 한눈에 확인할수있게 되었다. 그래서 이번에 이것을 이용해 CUDA 2.2 버전과 2]]></content:encoded>
</item>
<item>
<title><![CDATA[Trash-talking CUDA]]></title>
<link>http://reconshmigurable.wordpress.com/2009/11/19/trash-talking-cuda/</link>
<pubDate>Thu, 19 Nov 2009 16:49:48 +0000</pubDate>
<dc:creator>reconshmigurable</dc:creator>
<guid>http://reconshmigurable.wordpress.com/2009/11/19/trash-talking-cuda/</guid>
<description><![CDATA[Found lurking on HPCWire.com, in an interview with Steve Wallach of Convey Computer: &#8230;I call p]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Found lurking on <a href="http://hpcwire.com">HPCWire.com</a>, in an interview with <a href="http://en.wikipedia.org/wiki/Steve_Wallach">Steve Wallach</a> of <a href="http://www.conveycomputer.com/">Convey Computer</a>:</p>
<p><em>&#8230;I call programs that don&#8217;t take into consideration legacy systems and that are obscenely difficult to integrate, &#8220;pornographic&#8221; programs &#8212; you can&#8217;t always describe them exactly, but you know them when you see them. In 1984, I converted a FORTRAN program from CDC to ANSI FORTRAN to see what they were doing and it was awful. In the contemporary world, CUDA is the new pornographic programming language.<br />
</em></p>
<p><a href="http://bit.ly/pv2da">http://bit.ly/pv2da</a></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[CUDA en C# o .NET]]></title>
<link>http://cudagpu.wordpress.com/2009/11/19/cuda-en-csharp-o-net/</link>
<pubDate>Thu, 19 Nov 2009 08:40:51 +0000</pubDate>
<dc:creator>cudagpu</dc:creator>
<guid>http://cudagpu.wordpress.com/2009/11/19/cuda-en-csharp-o-net/</guid>
<description><![CDATA[Parece que hay posibilidades de hacer programas en CUDA en otros lenguajes distintos a C++. En concr]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Parece que hay posibilidades de hacer programas en CUDA en otros lenguajes distintos a C++. En concreto acabo de encontrar un proyecto que funciona en C#, ese <em>gran idioma</em> que se inventó Microsoft que está entre Java y C++.</p>
<p>Pues bien, podemos aprovechar las bondades que ofrece C# respecto a C++, sobretodo cuando vayamos a hacer una interfaz gráfica, que Visual Studio maneja decentemente en proyectos .NET.</p>
<p><strong>Configurar Visual Studio 2005 para CUDA con .NET</strong></p>
<p>Lo primero es ir a la siguiente <a href="http://www.hoopoe-cloud.com/Solutions/CUDA.NET/Default.aspx">dirección web de Hoopoe</a>, ahí podemos bajar un DLL (y algunos proyectos de ejemplo), concretamente el fichero <a href="http://www.hoopoe-cloud.com/files/cuda.net/2.3/cuda.net2.3.7_win.zip">es éste</a>.</p>
<p>Descomprimimos todo y copiamos lo que hay dentro de la carpeta \bin a donde nosotros queramos. Lo ideal es tener todo CUDA junto, así que yo lo puse en C:\CUDA\CUDA.NET</p>
<p>A continuación abrimos el proyecto más sencillo: transpose.csproj</p>
<p>Lo más seguro es que tengamos un problema de referencia porque no encuentra CUDA.NET, es fácil arreglar esta dependencia:</p>
<ul>
<li>Sobre la carpeta Propieties del proyecto hacemos doble click</li>
<li>Abrimos la pestaña &#8220;Rutas de acceso de referencia&#8221;</li>
<li>Escribimos el nombre de la carpeta donde hemos puesto el .dll anteriormente: C:\CUDA\CUDA.NET</li>
<li>Agregar carpeta</li>
</ul>
<p><strong>Actualizar el PATH</strong></p>
<p>Ahora ya debería compilar, sin embargo yo he tenido otro problema. El compilador nvcc devuelve -1, y si abrimos una consola nosotros e intentamos compilar a mano el problema es que no encuentra el compilador de C++ cl.exe. Para arreglar esto debemos añadir al PATH de Windows una ruta más:</p>
<ul>
<li>Abrir panel de control</li>
<li>Sistema</li>
<li>Opciones avanzadas</li>
<li>Variables de entorno</li>
<li>En variables de sistema buscamos: Path</li>
<li>Modificarla añadiendo al final lo siguiente: un punto y coma &#8216;;&#8217; y la ruta donde tengamos el compilador de C++, en nuestro caso será: C:\Archivos de programa\Microsoft Visual Studio 8\VC\bin</li>
<li>Si usas otro Visual Studio cambia el número 8 por 9 o 10.</li>
</ul>
<ul>
<li>El punto y coma &#8216;;&#8217; sirve para separar distintos directorios que añadimos al PATH, es importante no olvidarse ponerlo</li>
</ul>
<p>Con todo esto ya compila perfectamente, recomiendo añadir al final del fichero Program.cs de su método Main la línea: Console.ReadLine(); para que no se cierre la consola al instante.</p>
<p><strong>Revisando el proyecto</strong></p>
<p><strong></strong>Lo que tenemos es un fichero en C# compilado con Visual Studio y otro en CUDA que se compila con nvcc y C++. Hay que darse cuenta de que la compilación del fichero CUDA se realiza manualmente, en Propierties/Eventos de generación se encuentran las compilaciones de los ficheros CUDA. Si queremos tener más, habrá que añadir ahí sus compilaciones. No es demasiado complicado.</p>
<p>Por otra parte, las instrucciones que llaman a los kernels CUDA y reservan memoria cambian ligeramente respecto a las que conocemos de C++. Existe un objeto de la clase CUDA que es al que se le hacen las llamadas:</p>
<p>En el método Main se puede ver fácilmente como utilizarlo, son autoexplicables así que lo mejor es verlo uno mismo.</p>
<p>En general este método supongo que servirá para otros proyectos .NET.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Shared Memory Parallel Ray Tracer using OpenMP and CUDA]]></title>
<link>http://jestermax.wordpress.com/2009/11/18/shared-memory-parallel-ray-tracer-using-openmp-and-cuda/</link>
<pubDate>Thu, 19 Nov 2009 00:14:13 +0000</pubDate>
<dc:creator>jestermax</dc:creator>
<guid>http://jestermax.wordpress.com/2009/11/18/shared-memory-parallel-ray-tracer-using-openmp-and-cuda/</guid>
<description><![CDATA[So since I&#8217;ve been doing my Masters in Computer Science at Dalhousie, I&#8217;ve been working ]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>So since I&#8217;ve been doing my Masters in Computer Science at Dalhousie, I&#8217;ve been working on some neat projects. I&#8217;ll post them separately since they&#8217;re in separate research areas though.</p>
<p>For a parallel algorithms course, I&#8217;ve decided to play around with the ray tracer that Adam Stirtan and I created for an undergrad project. Right now I&#8217;m intending on using OpenMP to parallelize the ray tracing aspect of the program. If I have time left over then I&#8217;m hoping to use NVidia&#8217;s CUDA to get the GPU on my graphics card to work on the post processing shading effects. I&#8217;ve gotten some &#8220;hello world&#8221; style code compiled and running so my next step is to dig in a bit further with that and see what I can come up with.</p>
<div id="attachment_150" class="wp-caption alignnone" style="width: 310px"><a href="http://jestermax.wordpress.com/files/2009/11/js_rt_par_nov17_2009_a.png"><img class="size-medium wp-image-150" title="JS_RT_Par_Nov17_2009_A" src="http://jestermax.wordpress.com/files/2009/11/js_rt_par_nov17_2009_a.png?w=300" alt="" width="300" height="234" /></a><p class="wp-caption-text">Blind insertion of OpenMP to a ray tracer</p></div>
<p>Last night I made a simple stab at using OpenMP (Shared memory API for parallel programming) for using multiple threads for ray tracing. I wasn&#8217;t expecting much since I was just blindly adding some annotations to the project, but as you can see from the image it looks kinda neat. I&#8217;m assuming it chose to use 4 threads (since I didn&#8217;t specify a number).</p>
<p>Getting back to CUDA programming, trying to get a project set up isn&#8217;t really a fun thing. Thankfully somebody out there has made a project creation macro:</p>
<p>http://forums.nvidia.com/index.php?showtopic=65111</p>
<p>Using this, you can create a project that has the NVidia extra build step. This a crucial aspect of writing a CUDA program as you have to run it though an NVidia program so it can parse out the extension keywords. On that note, you have to rename anything using the C for CUDA extensions to a .cu file. Since this will then turn off your syntax highlighting I suggest you edit the text editor options so that the .cu extension will be opened in a C++ context. It&#8217;s really easy to do but if anybody is actually reading this and doesn&#8217;t know how, just post for help.</p>
<p>For some initial test code to make sure your program is compiling and running properly, I suggest you check out this page:</p>
<p>http://llpanorama.wordpress.com/2008/05/21/my-first-cuda-program/</p>
<p>That&#8217;s all for now.</p>
<p>Keep fit, and have fun</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Nvidia's next generation CUDA architecture]]></title>
<link>http://anarchia3lites.wordpress.com/2009/11/13/nvidias-next-generation-cuda-architecture/</link>
<pubDate>Fri, 13 Nov 2009 11:33:07 +0000</pubDate>
<dc:creator>jagenaw</dc:creator>
<guid>http://anarchia3lites.wordpress.com/2009/11/13/nvidias-next-generation-cuda-architecture/</guid>
<description><![CDATA[Hey everyone, I know that this statement would seem a bit &#8220;deja vu&#8221;, however, it has bee]]></description>
<content:encoded><![CDATA[Hey everyone, I know that this statement would seem a bit &#8220;deja vu&#8221;, however, it has bee]]></content:encoded>
</item>
<item>
<title><![CDATA[CUDA on Octave 3.x in linux 64 bits]]></title>
<link>http://scinesur.wordpress.com/2009/11/12/cuda-on-octave-3-x-in-linux-64-bits/</link>
<pubDate>Thu, 12 Nov 2009 18:58:18 +0000</pubDate>
<dc:creator>yonesur</dc:creator>
<guid>http://scinesur.wordpress.com/2009/11/12/cuda-on-octave-3-x-in-linux-64-bits/</guid>
<description><![CDATA[Hello all, I did all this work in august, but forgot to create a blog to post it. I&#8217;m sure the]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Hello all,</p>
<p>I did all this work in august, but forgot to create a blog to post it. I&#8217;m sure there are people out there interested in this stuff (I was but didin&#8217;t find anything!).</p>
<p>This are the instructions to have the Matlab CUDA plugin that Nvidia provides ( <a href="http://developer.nvidia.com/object/matlab_cuda.html">http://developer.nvidia.com/object/matlab_cuda.html</a> ) working in Octave in linux 64 bits.</p>
<p>In case you didn&#8217;t know, Octave supports since 3.2 single precision floats, so I recommend that for interfacing with CUDA.</p>
<p>Instructions for compiling Matlab CUDA plugin for Octave in Linux:</p>
<p>1. Install CUDA SDK and Toolkit following instructions from Nvidia site. In ubuntu 9.04, I had to install also:</p>
<blockquote><p>sudo apt-get install build-essential libglut3-dev<br />
sudo aptitude install libxi-dev<br />
sudo apt-get install libxmu-dev</p></blockquote>
<p>Remember the last warning in the installation:</p>
<blockquote><p>* Please make sure your PATH includes /usr/local/cuda/bin<br />
* Please make sure your LD_LIBRARY_PATH<br />
*   for 32-bit Linux distributions includes /usr/local/cuda/lib<br />
*   for 64-bit Linux distributions includes /usr/local/cuda/lib64</p></blockquote>
<p>2. Compile a program called mexext which produces &#8220;mex&#8221; as string output and put it in /usr/bin. I did</p>
<blockquote><p>#include &#60;stdio.h&#62;</p>
<p>int main()<br />
{<br />
printf(&#8220;mex\n&#8221;);<br />
return 0;<br />
}</p></blockquote>
<p>And compiled it, but I suppose a simple script with &#8220;echo mex&#8221; would work.</p>
<p>3. Use this Makefile, instead of the one supplied by Nvidia (problems with HTML formatting, see attachments at the botoom):</p>
<blockquote><p># Define installation location for CUDA and compilation flags compatible<br />
# with the CUDA include files.<br />
CUDAHOME    = /usr/local/cuda<br />
INCLUDEDIR  = -I$(CUDAHOME)/include<br />
INCLUDELIB  = -L$(CUDAHOME)/lib64 -lcufft -lcudart -Wl,-rpath,$(CUDAHOME)/lib64<br />
CFLAGS      = -fPIC -D_GNU_SOURCE -pthread -fexceptions -Wall<br />
COPTIMFLAGS = -O3 -funroll-loops -msse2</p>
<p># Define installation location for Octave<br />
MEX = /usr/bin/mkoctfile<br />
MEXEXT        = .mex<br />
MEXOPT   = &#8211;mex</p>
<p># nvmex is a modified mex script that knows how to handle CUDA .cu files.<br />
NVMEX = ./nvmkoctfile</p>
<p># List the mex files to be built.<br />
MEXFILES = fft2_cuda.mex       \<br />
fft2_cuda_sp_dp.mex \<br />
ifft2_cuda.mex      \<br />
Szeta.mex</p>
<p>all: $(MEXFILES)</p>
<p>clean:<br />
rm -f $(MEXFILES) *.linkinfo</p>
<p>#.SUFFIXES: .cu .o .mex</p>
<p>#.cu.o:<br />
#    nvcc -c COPTIMFLAGS=&#8217;$(COPTIMFLAGS)&#8217;  $&#60; $(INCLUDEDIR) $(INCLUDELIB)</p>
<p>#.mex:<br />
#    $(MKOCTFILE) CFLAGS=&#8217;$(CFLAGS)&#8217; COPTIMFLAGS=&#8217;$(COPTIMFLAGS)&#8217; $@ \<br />
#        $(INCLUDEDIR) $(INCLUDELIB)</p>
<p>.SUFFIXES: .cu .cu_o .mex</p>
<p>.c.mex:<br />
CFLAGS=&#8217;$(CFLAGS)&#8217; COPTIMFLAGS=&#8217;$(COPTIMFLAGS)&#8217; $(MEX) $(MEXOPT) $&#60; \<br />
$(INCLUDEDIR) $(INCLUDELIB)</p>
<p>.cu.mex:<br />
COPTIMFLAGS=&#8217;$(COPTIMFLAGS)&#8217; $(NVMEX) $(MEXOPT) $&#60; $(INCLUDEDIR) $(INCLUDELIB)</p></blockquote>
<p>3.Now perform these changes in the Octave suplied mkoctfile, and save it as nvmkoctfile in the source directory (or use the attached one, which is for Octave 3.0.1):</p>
<blockquote><p>mkoctfile always appends -g to<br />
: ${CFLAGS=&#8221;-g -O2&#8243;}<br />
: ${CPICFLAG=&#8221;-fPIC&#8221;}<br />
: ${CXX=&#8221;g++&#8221;}<br />
: ${CXXFLAGS=&#8221;-g -O2&#8243;}<br />
: ${CXXPICFLAG=&#8221;-fPIC&#8221;}</p>
<p>even if -g is not used.</p>
<p>Removed -g there and changed -O2 to -O3.</p>
<p>Also changed:</p>
<p>-W*)<br />
pass_on_options=&#8221;$pass_on_options $1&#8243;<br />
;;</p>
<p>to</p>
<p>-W*)<br />
pass_on_options=&#8221;$pass_on_options $1&#8243;<br />
;;<br />
-O*)<br />
pass_on_options=&#8221;$pass_on_options $1&#8243;<br />
;;<br />
-f*)<br />
pass_on_options=&#8221;$pass_on_options $1&#8243;<br />
;;</p></blockquote>
<p>Change the lines</p>
<blockquote><p>while [ $# -gt 0 ]; do<br />
file=<br />
case &#8220;$1&#8243; in<br />
*.c)<br />
file=$1<br />
cfiles=&#8221;$cfiles $file&#8221;<br />
;;</p></blockquote>
<p>to</p>
<blockquote><p>while [ $# -gt 0 ]; do<br />
file=<br />
case &#8220;$1&#8243; in<br />
*.c &#124; *.cu)<br />
file=$1<br />
cfiles=&#8221;$cfiles $file&#8221;<br />
;;</p></blockquote>
<p>Also</p>
<blockquote><p>: ${CC=&#8221;gcc&#8221;}<br />
: ${CFLAGS=&#8221;-O3&#8243;}<br />
: ${CPICFLAG=&#8221;-fPIC&#8221;}</p></blockquote>
<p>to</p>
<blockquote><p>: ${CC=&#8221;nvcc&#8221;}<br />
: ${CFLAGS=&#8217;-O3 -Xcompiler &#8220;-fPIC -D_GNU_SOURCE -pthread -fexceptions -m64&#8243; -Xcompiler &#8220;-O3 -funroll-loops -msse2 -DNDEBUG&#8221;&#8216;}<br />
: ${CPICFLAG=&#8221;"}</p></blockquote>
<p>4. Comment or remove in fft2_cuda_sp_dp.c the line</p>
<blockquote><p>#include &#8220;matrix.h&#8221;</p></blockquote>
<p>and make. If you have problems with &#8220;cannot be declared weak&#8221;, read http://forums.nvidia.com/index.php?showtopic=104673 .<br />
5. If something fails, remember to do a &#8220;make clean&#8221; AND delet all .o before compiling again.</p>
<p>6. Now you should have working .mex CUDA functions working in Octave. Use them as you please!</p>
<p>P.D. WordPress does not allow textfiles, but allow .doc. Remove the .doc or change it accordingly to the attached files!</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Guía de programación CUDA]]></title>
<link>http://cudagpu.wordpress.com/2009/11/12/guia-de-programacion-cuda/</link>
<pubDate>Thu, 12 Nov 2009 10:38:37 +0000</pubDate>
<dc:creator>cudagpu</dc:creator>
<guid>http://cudagpu.wordpress.com/2009/11/12/guia-de-programacion-cuda/</guid>
<description><![CDATA[Enlace a la guía de programación de CUDA, donde explican muy bien cómo afrontar este tipo de program]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Enlace a la guía de programación de CUDA, donde explican muy bien cómo afrontar este tipo de programación paralela, de obligada lectura a todo aquél que quiera programar en CUDA.</p>
<ul>
<li><a href="http://developer.download.nvidia.com/compute/cuda/2_3/toolkit/docs/NVIDIA_CUDA_Programming_Guide_2.3.pdf">Guía de programación CUDA 2.3</a></li>
<li><a href="http://developer.download.nvidia.com/compute/cuda/2_3/toolkit/docs/CUDA_Reference_Manual_2.3.pdf">Guía de referencia CUDA 2.3</a></li>
</ul>
<p><a href="http://www.nvidia.com/object/cuda_develop.html">Sitio original de nVidia</a>.</p>
<ul></ul>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Compilar en CUDA con Visual Studio]]></title>
<link>http://cudagpu.wordpress.com/2009/11/12/compilar-en-cuda-con-visual-studio/</link>
<pubDate>Thu, 12 Nov 2009 08:54:11 +0000</pubDate>
<dc:creator>cudagpu</dc:creator>
<guid>http://cudagpu.wordpress.com/2009/11/12/compilar-en-cuda-con-visual-studio/</guid>
<description><![CDATA[He decidido usar Visual Studio 2005 SP1, que es el que usamos en otras asignaturas, así no tengo que]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>He decidido usar Visual Studio 2005 SP1, que es el que usamos en otras asignaturas, así no tengo que preocuparme de conversiones y compatibilidad con 2008.</p>
<p><strong>Instalar SDK y configurar Visual Studio</strong></p>
<p>Lo primero es instalar la SDK de CUDA de la web de nVidia y también añadir los directorios de inclusión y bibliotecas. Debemos descargar <a href="http://www.nvidia.com/object/cuda_get.html">desde aquí</a> (pongo los enlaces para Windows XP 32bits):</p>
<ul>
<li><a href="http://www.nvidia.com/object/thankyou.html?url=/compute/cuda/2_3/drivers/cudadriver_2.3_winxp_32_190.38_general.exe">CUDA Driver</a></li>
<li><a href="http://www.nvidia.com/object/thankyou.html?url=/compute/cuda/2_3/toolkit/cudatoolkit_2.3_win_32.exe">CUDA Toolkit</a></li>
<li><a href="http://www.nvidia.com/object/thankyou.html?url=/compute/cuda/2_3/sdk/cudasdk_2.3_win_32.exe">CUDA SDK</a></li>
</ul>
<p>La compilación de ficheros CUDA con Visual Studio se puede hacer de dos maneras, con unas reglas de compilación personalizadas o con un plugin que ya existe.</p>
<p>Recomendable instalarlo todo en C:/CUDA y nos evitamos problemas. Lo siguiente incluir los directorios y bibliotecas. En Visual Studio: Herramientas -&#62; Opciones -&#62; Proyectos y soluciones. Una vez ahí, marcamos en el desplegable de la derecha: Archivos de inclusión y añadimos lo siguiente, supongo $(CUDA) como el directorio donde está instalado CUDA.</p>
<ul>
<li>$(CUDA)\include</li>
</ul>
<p>Ahora cambiamos el desplegable a: Archivos de biblioteca y añadimos:</p>
<ul>
<li>$(CUDA)\lib</li>
</ul>
<p>Esto nos asegura que todos los proyectos creados accederán a las cabeceras de CUDA y encontrarán sus librerías. A mí me ha quedado <a href="http://cudagpu.wordpress.com/files/2009/11/cuda-include.png">así</a> y <a href="http://cudagpu.wordpress.com/files/2009/11/cuda-libs.png">así</a>. Yo tengo añadido el glut de OpenGL para utilizarlo más adelante.</p>
<p><strong>Plugin CUDA VS Wizard</strong></p>
<p>Es una pequeña aplicación que añadirá un nuevo tipo de proyecto CUDA, muy útil y fácil de usar. Soporta aplicaciones para Win32 y Win64.</p>
<ul>
<li><a href="http://cudavswizard.sourceforge.net/">Web oficial</a></li>
<li>Descarga: <a href="http://sourceforge.net/projects/cudavswizard/files/cudavswizard/CUDA_VS_Wizard_2.0%20Release/CUDA_VS_Wizard_W32.2.0.zip/download">Win32</a> &#124; <a href="http://sourceforge.net/projects/cudavswizard/files/cudavswizard/CUDA_VS_Wizard_2.0%20Release/CUDA_VS_Wizard_W64.2.0.1.zip/download">Win64</a></li>
</ul>
<p>Yo sólo he podido probar la Win32 y puedo afirmar que funciona correctamente, sólo hay que hacer un nuevo proyecto y los ficheros con extensión .cu se compilarán con el compilador CUDA.</p>
<p><img class="aligncenter size-full wp-image-10" title="cudawizard" src="http://cudagpu.wordpress.com/files/2009/11/cudawizard.png" alt="cudawizard" width="666" height="306" /></p>
<p>Lo siguiente es poner algún ejemplo fácil como el siguiente:</p>
<pre class="brush: cpp; collapse: true; gutter: false; light: false; toolbar: true;">
#include &#60;cutil_inline.h&#62;
#include &#60;cutil_math.h&#62;

#include &#60;stdio.h&#62;

__global__ void
kernelTrasMatrix(float * d_original, float * d_nueva, uint ancho, uint alto)
{
 uint viejo = blockIdx.x*ancho+threadIdx.x;
 uint nuevo = threadIdx.x*alto+blockIdx.x;

 d_nueva[nuevo]=d_original[viejo];
}

void cpuTrasMatrix(int i, int j, float * d_original, float * d_nueva, uint ancho, uint alto)
{
 uint viejo = i+j*ancho;
 uint nuevo = j+i*alto;

 d_nueva[nuevo]=d_original[viejo];
}

int main(){
 int ancho = 3, alto = 4;
 float original[]={1.0f,2.0f,3.0f,4.0f,5.0f,6.0f,7.0f,8.0f,9.0f,10.0f,11.0f,12.0f};

 cudaChannelFormatDesc channelDesc = cudaCreateChannelDesc&#60;float&#62;();
 float *d_original,*d_nueva;
 cudaMalloc(&#38;d_original,ancho*alto);
 cudaMalloc(&#38;d_nueva,ancho*alto);
 cudaMemcpy(d_original, original,sizeof(float)*ancho*alto,cudaMemcpyHostToDevice);

 int block_size = 3;
 int n_blocks = 4;
 kernelTrasMatrix&#60;&#60;&#60;n_blocks, block_size&#62;&#62;&#62;( d_original, d_nueva, ancho, alto );

 float *nueva = new float[ancho*alto];
 cudaMemcpy(nueva, d_nueva, sizeof(float)*alto*ancho, cudaMemcpyDeviceToHost);

 printf(&#34;-------- ORIGINAL ---------\n&#34;);
 for (int i=0; i&#60;alto; i++) {
 for(int j=0;j&#60;ancho;j++){
 printf(&#34;%3.1f    &#34;, original[j+i*ancho]);
 }
 printf(&#34;\n&#34;);
 }

 float *nuevaCPU = new float[ancho*alto];
 for(int i=0;i&#60;ancho;i++){
 for(int j=0;j&#60;alto;j++){
 cpuTrasMatrix(i,j,original,nuevaCPU,ancho,alto);
 }
 }

 printf(&#34;\n\n-------- TRASPUESTA GPU ---------\n&#34;);
 for (int i=0; i&#60;ancho; i++) {
 for(int j=0;j&#60;alto;j++){
 printf(&#34;%3.1f    &#34;, nueva[j+i*alto]);
 }
 printf(&#34;\n&#34;);
 }

 printf(&#34;\n\n-------- TRASPUESTA CPU ---------\n&#34;);
 for (int i=0; i&#60;ancho; i++) {
 for(int j=0;j&#60;alto;j++){
 printf(&#34;%3.1f    &#34;, nuevaCPU[j+i*alto]);
 }
 printf(&#34;\n&#34;);
}
 free(nueva);

 system(&#34;PAUSE&#34;);
}
</pre>
<p>El comportamiento es sencillo, hace la trasposición de una matriz en GPU y en CPU, lo hice rápidamente. Hay un ejemplo muy bueno de transposición de matrices en la SDK de CUDA (ya hablaré de ella) de nombre transpose, donde hacen dos tipos de transposiciones: una que es como la que hice yo y otra mucho más optimizada para leer y escribir en posiciones de memoria adyacentes, lo que acelera 10 veces el cálculo.</p>
<p><strong>Con una regla personalizada</strong></p>
<p>Si no nos gusta la idea de añadir el plugin, hay otra manera de compilar archivos en CUDA, esta vez tendremos que crear un proyecto de consola Win32 de C++ y escribir nuestro fichero CUDA (recordar extensión .cu).</p>
<p>Una vez hecho, entramos en las propiedades del fichero .cu y en General -&#62; Herramienta, marcamos Herramienta de generación personalizada y Aplicar.</p>
<p>Ahora entramos en Paso de generación personalizada y en Línea de comandos escribimos lo siguiente:</p>
<pre class="brush: cpp; toolbar: false; wrap-lines: false;">

&#34;$(CUDA_BIN_PATH)\nvcc.exe&#34; -ccbin &#34;$(VCInstallDir)bin&#34; -c -D_DEBUG -DWIN32 -D_CONSOLE -D_MBCS -Xcompiler /EHsc,/W3,/nologo,/Wp64,/Od,/Zi,/MTd -I&#34;$(CUDA_INC_PATH)&#34; -I./ -o $(ConfigurationName)\example1.obj example1.cu
</pre>
<p>CUIDADO: al copiar las comillas puede que os esté copiando comillas no estándar, sustituidlas.</p>
<p>CUIDADO 2: esto supone que el fichero se llama <em>example1.cu</em>, importante cambiarlo si vuestro fichero CUDA tiene un nombre diferente.</p>
<p>Lo siguiente es ir a la parte de Resultados y poner:</p>
<pre class="brush: cpp; toolbar: false; wrap-lines: false;">$(ConfigurationName)\example1.obj</pre>
<p>Aquí lo mismo, example1.obj es el resultado de la compilación del paso anterior, así que hay que cambiarlo a lo que se tenga en ese momento.</p>
<p><strong>Añadir bibliotecas adicionales</strong></p>
<p>Una vez ya tenemos los pasos de compilación indicados, bien sea por el asistente o a mano, vamos a las propiedades del proyecto -&#62; Vinculador -&#62; Entrada -&#62; Dependencias adicionales y añadimos:</p>
<ul>
<li>cudart.lib cutil32D.lib</li>
</ul>
<p>En el caso de querer compilar ejemplos que usen glut o glew necesitaremos añadir otros libs ahí, pero ya depende de cada proyecto.</p>
<p>Y si no me dejo nada esto es todo lo que hace falta (¡que no es poco!) para compilar tu primer programa en CUDA.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Best of SEMA: Muscle Cars]]></title>
<link>http://blog.cardomain.com/2009/11/11/best-of-sema-muscle-cars/</link>
<pubDate>Wed, 11 Nov 2009 16:47:12 +0000</pubDate>
<dc:creator>Rob Einaudi</dc:creator>
<guid>http://blog.cardomain.com/2009/11/11/best-of-sema-muscle-cars/</guid>
<description><![CDATA[So it&#8217;s time to kick of our Best of SEMA voting, and I thought I&#8217;d start with muscle. Th]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>So it&#8217;s time to kick of our Best of SEMA voting, and I thought I&#8217;d start with muscle. There were a ton of cool muscle cars at SEMA this year (half of them were Camaros), so it was difficult to narrow it down to just five cars. But here goes. For your consideration, the <a href="http://blog.cardomain.com/2009/11/02/trick-pony-fords-completely-carbon" target="_blank">Mustang RTR-C,</a> the <a href="http://blog.cardomain.com/2009/11/10/neils-camaro/" target="_blank">Tjin Edition Camaro</a>, the <a href="http://blog.cardomain.com/2009/11/05/super-cuda-can-i-keep-it-pretty-pl" target="_blank">Super Cuda</a>, the <a href="http://blog.cardomain.com/2009/11/06/sema-best-of-show-ring-brothers-64" target="_blank">Ring Brothers Fairlane</a> and the <a href="http://blog.cardomain.com/2009/11/04/tom-nelson-camaro-at-sema/" target="_blank">Tom Nelson Camaro</a>. Vote for your favorite, and be sure to come back over the next two weeks to vote on vehicles in each of ten categories. We&#8217;ll announce the winners shorly after Thanksgiving.</p>
<p>!!!<!--Slide.com error: provide id, w, h--></p>
<a name="pd_a_2236039"></a><div class="PDS_Poll" id="PDI_container2236039" style="display:inline-block;"></div><script type="text/javascript" language="javascript" charset="utf-8" src="http://static.polldaddy.com/p/2236039.js"></script>
		<noscript>
		<a href="http://answers.polldaddy.com/poll/2236039/">View This Poll</a><br/><span style="font-size:10px;"><a href="http://www.polldaddy.com">survey software</a></span>
		</noscript>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Hemi Cuda - Plymouth - musclemania - 2008]]></title>
<link>http://amocolecoes.wordpress.com/2009/11/11/hemi-cuda-plymouth-musclemania-2008/</link>
<pubDate>Wed, 11 Nov 2009 15:10:43 +0000</pubDate>
<dc:creator>Gustavo Bomfim</dc:creator>
<guid>http://amocolecoes.wordpress.com/2009/11/11/hemi-cuda-plymouth-musclemania-2008/</guid>
<description><![CDATA[]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><img class="alignnone size-medium wp-image-61" title="PB060156" src="http://amocolecoes.wordpress.com/files/2009/11/pb060156.jpg?w=300" alt="PB060156" width="300" height="225" /></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Tutorial de CUDA]]></title>
<link>http://cudagpu.wordpress.com/2009/11/11/tutorial-de-cuda/</link>
<pubDate>Wed, 11 Nov 2009 13:58:05 +0000</pubDate>
<dc:creator>cudagpu</dc:creator>
<guid>http://cudagpu.wordpress.com/2009/11/11/tutorial-de-cuda/</guid>
<description><![CDATA[Este es un tutorial cortesía de Parallel Panorama y Dr. Dobb&#8217;s, muy útil para empezar: Empezan]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Este es un tutorial cortesía de <a href="http://llpanorama.wordpress.com/cuda-tutorial/" target="_blank">Parallel Panorama</a> y <a href="http://www.ddj.com/">Dr. Dobb&#8217;s</a>, muy útil para empezar:</p>
<ul>
<li><a href="http://llpanorama.wordpress.com/2008/04/24/getting-started-with-cuda/">Empezando con CUDA</a></li>
<li><a href="http://llpanorama.wordpress.com/2008/05/21/my-first-cuda-program/">Mi primer programa en CUDA</a></li>
<li><a href="http://llpanorama.wordpress.com/2008/06/11/threads-and-blocks-and-grids-oh-my/">Hilos, bloques y rejillas</a></li>
<li><a href="http://llpanorama.wordpress.com/2009/08/07/updating-to-cuda-2-3/">CUDA 2.3</a></li>
<li>Tutorial CUDA:
<ul>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/207200659">Parte 1: Introducción</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/207402986">Parte 2: Mi primer Kernel</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/207603131">Parte 3: Manejo de errores</a></li>
<li><a href="http://www.ddj.com/architect/208401741">Parte 4: La memoria compartida</a> (1)</li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/208801731">Parte 5: La memoria compartida</a> (2)</li>
<li><a href="http://www.ddj.com/architect/209601096">Parte 6: La memoria global</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/210102115">Parte 7: La siguiente generación de hardware CUDA</a></li>
<li><a href="http://www.ddj.com/architect/210602684">Parte 8: Usando bibliotecas con CUDA</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/211800683">Parte 9: Extendiendo lenguajes de alto nivel con CUDA</a></li>
<li><a href="http://www.ddj.com/architect/212903437">Parte 10: La biblioteca CUDPP</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/215900921">Parte 11: Revisitando los espacios de memoria CUDA</a></li>
<li><a href="http://www.ddj.com/architect/217500110">Parte 12: CUDA 2.2 cambia el paradigma de datos</a></li>
<li><a href="http://www.ddj.com/cpp/218100902?pgno=1">Parte 13: La memoria de texturas en CUDA</a></li>
<li><a href="http://www.ddj.com/hpc-high-performance-computing/220601124">Parte 14: Depuración en CUDA con CUDA-GDB</a></li>
</ul>
</li>
</ul>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Swedenborg - świat ducha i koniec świata]]></title>
<link>http://moon5.wordpress.com/2009/11/11/swedenborg-swiat-ducha-i-koniec-swiata/</link>
<pubDate>Wed, 11 Nov 2009 03:22:38 +0000</pubDate>
<dc:creator>Moon</dc:creator>
<guid>http://moon5.wordpress.com/2009/11/11/swedenborg-swiat-ducha-i-koniec-swiata/</guid>
<description><![CDATA[Według Emanuela Swedenborga (1688-1772), szwedzkiego matematyka, fizyka, filozofa i mistyka, w 1770 ]]></description>
<content:encoded><![CDATA[Według Emanuela Swedenborga (1688-1772), szwedzkiego matematyka, fizyka, filozofa i mistyka, w 1770 ]]></content:encoded>
</item>
<item>
<title><![CDATA[Open CL.  Dead man walking?]]></title>
<link>http://rndness.com/2009/11/10/open-cl-dead-man-walking/</link>
<pubDate>Tue, 10 Nov 2009 21:36:32 +0000</pubDate>
<dc:creator>stuckat1</dc:creator>
<guid>http://rndness.com/2009/11/10/open-cl-dead-man-walking/</guid>
<description><![CDATA[As a longtime CUDA experimenter, I have always wondered how OpenCL compares to CUDA. HPC Wire has pu]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><img class="aligncenter size-full wp-image-75" title="OpenCL" src="http://rndness.wordpress.com/files/2009/11/opencl_logo_rgb.png" alt="OpenCL" width="150" height="150"></p>
<p>As a longtime CUDA experimenter, I have always wondered how OpenCL compares to CUDA.  HPC Wire has published an interesting article on OpenCL called <a href="http://www.hpcwire.com/features/Compilers-and-More-OpenCL-Promises-and-Potential-58625442.html?viewAll=y"> &#8220;Compilers and More: OpenCL Promises and Potential&#8221;</a> that answers this question.<br />
<!--more--><br />
Compared to CUDA, OpenCL focuses on a low-level programming model.  This is unfortunate because its is the higher level programming model that has made CUDA so popular.   Since CUDA has more than two years head start on OpenCL I don&#8217;t see NVIDIA promoting OpenCL more than CUDA.  Apple, the biggest supporter, can really push OpenCL as much as it wants, like including it in Snow Leopard, but I don&#8217;t think they have the muscle to drive this new programming model, especially if their biggest partner is ATI.</p>
<p>With the way AMD is doing financially today, I can&#8217;t imagine ATI or AMD putting a dent in either CPU or GPU markets.  It is an Intel and NVIDIA world.  Hell, with all the recent speculation, its more likely that<a href="http://www.pcmag.com/article2/0,2817,2355167,00.asp"> NVIDIA enters the x86 market </a>with their own CPU, maybe they could buy the once great AMD &#8230; and kill off AMD. <img src='http://s.wordpress.com/wp-includes/images/smilies/icon_wink.gif' alt=';)' class='wp-smiley' /> </p>
<p>The real question is how the development world will change when Intel Larabee arrives.  Larabee being the uberly massive multicore upcoming processor.  Obviously, Intel has been pushing hard the newly released <a href="http://software.intel.com/en-us/intel-parallel-studio-home/">Parallel Studio</a> so they realize how important the software tool-chain is.  Will this be the new development standard?</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[NVIDIA: Fermi GT300 adiado até 2010]]></title>
<link>http://nvidiathepoweroffuture.wordpress.com/2009/11/10/nvidia-fermi-gt300-adiado-ate-2010/</link>
<pubDate>Tue, 10 Nov 2009 21:30:40 +0000</pubDate>
<dc:creator>ppinheiro</dc:creator>
<guid>http://nvidiathepoweroffuture.wordpress.com/2009/11/10/nvidia-fermi-gt300-adiado-ate-2010/</guid>
<description><![CDATA[    Segundo relatórios da Digitimes o lançamento do novo GPU da Nvidia, o GT300, foi adiado para o a]]></description>
<content:encoded><![CDATA[    Segundo relatórios da Digitimes o lançamento do novo GPU da Nvidia, o GT300, foi adiado para o a]]></content:encoded>
</item>
<item>
<title><![CDATA[Miłość czyni cuda]]></title>
<link>http://pascalalter.wordpress.com/2009/11/07/milosc-czyni-cuda/</link>
<pubDate>Sat, 07 Nov 2009 20:42:54 +0000</pubDate>
<dc:creator>pascalalter</dc:creator>
<guid>http://pascalalter.wordpress.com/2009/11/07/milosc-czyni-cuda/</guid>
<description><![CDATA[sobota, 7 listopada 2009 Sie haben eine Karte erhalten Hallo Notatnik Pascala, Pascal Alter hat auf ]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><h2>sobota, 7 listopada 2009</h2>
<p><a name="8064660938345073562"></a></p>
<h3><a href="http://pascalalter.blogspot.com/2009/11/sie-haben-eine-karte-erhalten.html">Sie haben eine Karte erhalten</a></h3>
<p>Hallo Notatnik Pascala,</p>
<p>Pascal Alter hat auf der Website <a href="http://donaukurier.de/">donaukurier.de</a> eine elektronische Grußkarte für Sie entworfen.<br />
Um sich die Karte anzusehen, brauchen Sie nur auf die unten angegebene URL zu klicken oder<br />
rufen Sie einfach die folgende URL mit Ihrem Internet-Browser auf:</p>
<p><a href="http://www.donaukurier.de/_/ecards.html?PUC=9e34fede5f0b38e7">http://www.donaukurier.de/_/ecards.html?PUC=9e34fede5f0b38e7</a></p>
<p>ADRESSE: <a href="http://www.donaukurier.de/_/ecards.html">http://www.donaukurier.de/_/ecards.html</a><br />
PICKUP-CODE: 9e34fede5f0b38e7</p>
<p>Ihre Grußkarte wird 30 Tage gespeichert, und danach automatisch<br />
gelöscht (gültig bis 07.12.2009).<br />
Deshalb sollten Sie Ihre Karte möglichst schnell ausdrucken oder auf andere<br />
Art sichern.</p>
<p>Mit freundlichen Grüßen<br />
Ihr Team von <a href="http://donaukurier.de/">donaukurier.de</a></p>
<p>Tel.     +49 841 885560-0<br />
Fax      +49 841 885560-11<br />
E-Mail   <a href="mailto:online@donaukurier.de">online@donaukurier.de</a><br />
Internet <a href="http://www.donaukurier.de/">www.donaukurier.de</a></p>
<p>DONAUKURIER Verlagsgesellschaft mbH &#38; Co. KG<br />
Stauffenbergstr. 2a, D-85051 Ingolstadt<br />
USt-IDNr. DE 813250256, Steuernummer 124/156/00516<br />
Reg.-Gericht Ingolstadt HRA 1402 P.H.G. DONAUKURIER Verlag<br />
Verwaltungsgesellschaft mbH, Reg.Gericht Ingolstadt HRB 3036<br />
Geschäftsführung: Elin Reissmüller, Wolfgang Lichtenegger, Georg Schäff</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Rant on CUDA, FPGA and Cell]]></title>
<link>http://rndness.com/2009/11/06/rant-on-cuda-fpga-and-cell/</link>
<pubDate>Fri, 06 Nov 2009 22:05:01 +0000</pubDate>
<dc:creator>stuckat1</dc:creator>
<guid>http://rndness.com/2009/11/06/rant-on-cuda-fpga-and-cell/</guid>
<description><![CDATA[I wrote this really long rant on the Quant Net site about different emerging technologies used in co]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><img src="http://rndness.wordpress.com/files/2009/11/rant_small.jpg" alt="Rant" title="Rant" width="278" height="320" class="aligncenter size-full wp-image-43" /></p>
<p>I wrote this really <a href="http://www.quantnet.com/forum/showthread.php?t=2810"> long rant</a> on the <a href="http://www.quantnet.com">Quant Net</a> site about different emerging technologies used in computational finance.  It focuses on the differences between developing on Nvidia video cards, Field Programmable Gate Arrays (FPGAs) and IBM&#8217;s Cell processor.  Please take a look.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[CUDA + VS 2008 = Sweetness]]></title>
<link>http://rndness.com/2009/11/06/cuda-vs-2008-sweetness/</link>
<pubDate>Fri, 06 Nov 2009 19:49:07 +0000</pubDate>
<dc:creator>stuckat1</dc:creator>
<guid>http://rndness.com/2009/11/06/cuda-vs-2008-sweetness/</guid>
<description><![CDATA[It seems that every week I get some lame email from Nvidia announcing updates to CUDA. Since the rel]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><span style='text-align:center; display: block;'><object width='425' height='350'><param name='movie' value='http://www.youtube.com/v/FLQuqXhlx40&#038;rel=1&#038;fs=1&#038;showsearch=0&#038;hd=0' /><param name='allowfullscreen' value='true' /><param name='wmode' value='transparent' /><embed src='http://www.youtube.com/v/FLQuqXhlx40&#038;rel=1&#038;fs=1&#038;showsearch=0&#038;hd=0' type='application/x-shockwave-flash' allowfullscreen='true' width='425' height='350' wmode='transparent'></embed></object></span></p>
<p>It seems that every week I get some lame email from Nvidia announcing updates to CUDA.  Since the release which introduced compatibility with Windows XP x64 support, none of the updates have been that interesting &#8230; until today.  The CUDA 3.0 Beta release contains a slew of additions, like C++ Class Inheritance and Templates and debugging in emulation-mode.  Historically, code executed under CUDA had to be straight C: no longer.  THe more interesting part of the release is the introduction of <a href="http://developer.nvidia.com/object/nexus.html">Nexus</a>.<br />
<!--more--><br />
Nexus permits debugging, profiling and anlaysis of GPU code programmed using CUDA and Open CL all within Visual Studio 2008.  No longer do you have to guess why your code runs differently between emulation and real silicon.  Also, its all runs within the confines of a single program namely, Visual Studio 2008.   </p>
<p>For those who like read release notes, here is an exert:</p>
<p>This information is titled: CUDA Toolkit 3.0 Beta</p>
<p>This information is for: General</p>
<p>The information description is:<br />
The CUDA Toolkit 3.0 Beta is now available to GPU Computing registered<br />
developers.</p>
<p>Highlights for this release include:</p>
<p>* CUDA Driver / Runtime Buffer Interoperability, which allows<br />
  applications using the CUDA Driver API to also use libraries<br />
  implemented using the CUDA C Runtime.</p>
<p>* A new, separate version of the CUDA C Runtime (CUDART) for debugging<br />
  in emulation-mode.</p>
<p>* C++ Class Inheritance and Template Inheritance support for increased<br />
  programmer productivity</p>
<p>* A new unified interoperability API for Direct3D and OpenGL, with<br />
  support for:<br />
  * OpenGL texture interop<br />
  * Direct3D 11 interop support</p>
<p>* cuda-gdb hardware debugging support for applications that use the CUDA<br />
  Driver API</p>
<p>* New CUDA Memory Checker reports misalignment and out of bounds errors,<br />
  available as a debugging mode within cuda-gdb and also as a<br />
  stand-alone utility.</p>
<p>* CUDA Toolkit libraries are now versioned, enabling applications to<br />
  require a specific version, support multiple versions explicitly, etc.</p>
<p>* CUDA C/C++ kernels are now compiled to standard ELF format</p>
<p>* Support for all the OpenCL features in the latest R195.39 beta driver:<br />
  * Double Precision<br />
  * OpenGL Interoperability, for interactive high performance<br />
    visualization<br />
  * Query for Compute Capability, so you can target optimizations for<br />
    GPU architectures (cl_nv_device_attribute_query)<br />
  * Ability to control compiler optimization settings, etc. via support<br />
    for NVIDIA Compiler Flags (cl_nv_compiler_options)<br />
  * OpenCL Images support, for better/faster image filtering<br />
  * 32-bit Atomics for fast, convenient data manipulation<br />
  * Byte Addressable Stores, for faster video/image processing and<br />
    compression algorithms<br />
  * Support for the latest OpenCL spec revision 48 and latest official<br />
    Khronos OpenCL headers as of 11/1/2009</p>
<p>* Early support for the Fermi architecture, including:<br />
  * Native 64-bit GPU support<br />
  * Multiple Copy Engine support<br />
  * ECC reporting<br />
  * Concurrent Kernel Execution<br />
  * Fermi HW debugging support in cuda-gdb</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[CUDA Complex Library]]></title>
<link>http://boredmathematician.wordpress.com/2009/11/06/cuda-complex-library/</link>
<pubDate>Fri, 06 Nov 2009 15:57:02 +0000</pubDate>
<dc:creator>Tina</dc:creator>
<guid>http://boredmathematician.wordpress.com/2009/11/06/cuda-complex-library/</guid>
<description><![CDATA[Well, at least life isn’t as boring as it used to be. I’ve been devoting most of my time figuring ou]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Well, at least life isn’t as boring as it used to be. I’ve been devoting most of my time figuring out how to implement the algorithm I’m studying using CUDA. And more than half of the time, I’m just figuring out how to use it. Suddenly, I miss the things I take for granted when I am programming in Scilab or Matlab. Gah.</p>
<p>Recently a new annoyance came up. I thought that using complex numbers is as easy as how it is done in C. Apparently, that is not the case. The inherent complex library in CUDA is not as extensive as the one in C, so my code is a mess. Now I’m having trouble figuring out pointers to the complex-valued arrays in my problem. The “C-approach” is not working.</p>
<p>Things would have been a little less stressful if I don’t have a countdown. I only have at most 31 days to finish this code, together with 2 other projects that are entirely unrelated to this problem. I am so screwed.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[1971 Plymouth Hemi Cuda Engine Classic 1280X960]]></title>
<link>http://carphotos1.wordpress.com/2009/11/04/1971-plymouth-hemi-cuda-engine-classic-1280x960/</link>
<pubDate>Wed, 04 Nov 2009 00:56:06 +0000</pubDate>
<dc:creator>carphotos1</dc:creator>
<guid>http://carphotos1.wordpress.com/2009/11/04/1971-plymouth-hemi-cuda-engine-classic-1280x960/</guid>
<description><![CDATA[1971 plymouth hemi cuda engine Classic 1280&#215;960 wallpaper picture of 1971 plymouth hemi cuda en]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><strong>1971 plymouth hemi cuda engine Classic 1280&#215;960 wallpaper</strong><a title=" 1971 plymouth hemi cuda engine Classic 1280x960" href="http://www.carpictures1.com/index.php/1971_plymouth_hemi_cuda_engine_Classic_1280x960.jpg"><img class="aligncenter" title="1971 plymouth hemi cuda engine Classic 1280x960" src="http://www.carpictures1.com/var/resizes/1971_plymouth_hemi_cuda_engine_Classic_1280x960.jpg" alt="1971 plymouth hemi cuda engine Classic 1280x960 wallpaper" width="640" height="480" /></a><strong> picture of 1971 plymouth hemi cuda engine Classic 1280&#215;960</strong><br />to view full size click1971 Plymouth Hemi Cuda Engine Classic 1280X960 wallpaper <strong>( 1971 Plymouth Hemi Cuda Engine Classic 1280X960 wallpaper )</strong></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[tech news]]></title>
<link>http://xollothnews.wordpress.com/2009/11/02/tech-news-17/</link>
<pubDate>Mon, 02 Nov 2009 09:42:14 +0000</pubDate>
<dc:creator>xollothnews</dc:creator>
<guid>http://xollothnews.wordpress.com/2009/11/02/tech-news-17/</guid>
<description><![CDATA[Μιλήστε online χωρίς χρεώσεις v4.ethnos.gr &#8211; Μιλήστε online χωρίς χρεώσεις &#8211; e-life , τε]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/online.html">Μιλήστε online χωρίς χρεώσεις</a></h3>
<div class="post-body entry-content">
<div><a href="http://www.ethnos.gr/article.asp?catid=11429&#38;subid=2&#38;pubid=7690870">v4.ethnos.gr &#8211; Μιλήστε online χωρίς χρεώσεις &#8211; e-life , τεχνολογια</a></p>
<blockquote><p><a href="http://www.ethnos.gr/article.asp?catid=11429&#38;subid=2&#38;pubid=7690870#"><img src="http://engine.feed.gr/?catid=11429&#38;imgid=6691305&#38;srcid=420&#38;type=11495" border="0" alt="521 εκατομμύρια είναι ο συνολικός αριθμός των εγγεγραμμένων  χρηστών της εταιρείας Skype. " /></a>ηλεφωνικές κλήσεις δωρεάν, βιντεοκλήσεις δωρεάν κι άμεση αποστολή μηνυμάτων (καθώς και ηχομηνυμάτων)&#8230; επίσης δωρεάν. Εντυπωσιακά προνομιακό, το εν λόγω πακέτο υπηρεσιών προσφέρεται στους απανταχού κυβερνοναύτες μέσω εφαρμογών ιντερνετικής τηλεφωνίας (των επονομαζόμενων εφαρμογών «VoIP»). Λόγος για προγράμματα λογισμικού τα οποία επιτρέπουν τη χωρίς χρέωση πραγματοποίηση φωνητικών, τηλεφωνικού τύπου κλήσεων μέσω Διαδικτύου, από υπολογιστή σε υπολογιστή (με την προϋπόθεση ότι και τα δύο PCs είναι εξοπλισμένα είτε με μικρόφωνο και ακουστικά είτε με κάποια VoIP τηλεφωνική συσκευή),<a href="http://www.ethnos.gr/article.asp?catid=11429&#38;subid=2&#38;pubid=7690870">[next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
<div class="post hentry"><a name="1152807097066345288"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/robot-rovio-webcam.html">Robot Rovio με Webcam</a></h3>
<div class="post-body entry-content">
<div><a href="http://www.smartgadget.gr/content/robot-rovio-me-webcam">Robot Rovio με Webcam &#124; SmartGadget.gr</a></p>
<blockquote><p><a class="lightbox-processed" rel="lightbox[][rovio.jpg]" href="http://www.smartgadget.gr/sites/default/files/images/rovio.jpg"><img class="image image-preview " title="rovio.jpg" src="http://www.smartgadget.gr/sites/default/files/images/rovio.jpg" alt="rovio.jpg" width="320" height="320" /></a>Το απόλυτο εργαλείο παρακολούθησης! Το ρομποτάκι Rovio είναι ένα ρομποτάκι-ασφαλείας με κάμερα, το οποίο χειρίζεστε εξ αποστάσεως μέσω web browser ή μέσω κινητού με σύνδεση στο διαδίκτυο.<a href="http://www.smartgadget.gr/content/robot-rovio-me-webcam">[next</a>]
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
</div>
<div class="post hentry"><a name="8292663693020231592"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/blog-post_8091.html">Αναβαθμίστε τη σύνδεσή σας χωρίς να αγοράζετε παραπανήσιες χωρητικότητες</a></h3>
<div class="post-body entry-content">
<div><a href="http://www.mobile-news.gr/gr/articles/view/4758/anavathmiste_th_sundesh_sa_a_chwri_a_na_agorazete_parapanhsie_a_chwrhtikuthte_a/">MobileNews / Αναβαθμίστε τη σύνδεσή σας χωρίς να αγοράζετε παραπανήσιες χωρητικότητες / ΕΛΛΑΔΑ / News</a></p>
<blockquote><p><img src="http://www.mobile-news.gr/web/files/article_images/2009/11/4758/thumb_350x250_7525.jpg" alt="Αναβαθμίστε τη σύνδεσή σας χωρίς να αγοράζετε παραπανήσιες χωρητικότητες" width="350" height="250" />Το βασικό πρωτόκολλο μετάδοσης στις τηλεπικοινωνίες, το Ethernet, χάρη στην OTEGLOBE προσφέρεται με δύο λύσεις, την Ethernet VPN και την Ethernet Transport, με ταχύτητες από 1Mbps έως και 10Gbps για κάθε πάροχο και εταιρεία.[<a href="http://www.mobile-news.gr/gr/articles/view/4758/anavathmiste_th_sundesh_sa_a_chwri_a_na_agorazete_parapanhsie_a_chwrhtikuthte_a/">next]</a>
<p>&#160;</p>
</blockquote>
</div>
</div>
<p>&#160;</p>
</div>
<div class="post hentry"><a name="1689126167761776894"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/beta-drivers-geforce-19539-nvidia.html">Beta drivers GeForce 195.39 από την NVidia</a></h3>
<div class="post-body entry-content">
<div><a href="http://away.gr/2009/11/02/geforce-19539-beta-drivers-from-nvidia/">Beta drivers GeForce 195.39 από την NVidia &#124; Away.gr, ελληνικό online media για την τεχνολογία, τα games και την επιχειρηματικότητα</a></p>
<blockquote><p><a href="http://images.google.com/imgres?imgurl=http://nunetherlands.files.wordpress.com/2009/08/nvidia-400px.jpg&#38;imgrefurl=http://nunetherlands.wordpress.com/&#38;usg=__W2rGvKU_P6xzR61zFWaOfLWifPg=&#38;h=432&#38;w=450&#38;sz=20&#38;hl=en&#38;start=4&#38;tbnid=HKA5yhWZAZowaM:&#38;tbnh=122&#38;tbnw=127&#38;prev=/images%3Fq%3DNVidia%26ndsp%3D18%26hl%3Den%26lr%3D%26sa%3DN%26start%3D1"><img style="border:1px solid;" src="http://t1.gstatic.com/images?q=tbn%3AHKA5yhWZAZowaM%3Ahttp%3A%2F%2Fnunetherlands.files.wordpress.com%2F2009%2F08%2Fnvidia-400px.jpg&#038;w=127&#038;h=122" alt="" width="127" height="122" /></a>Η NVidia κυκλοφόρησε τους νέους beta drivers 195.39, οι οποίοι είναι οι πρώτοι που προσφέρουν υποστήριξη στο <a class="zem_slink" title="OpenCL" rel="homepage" href="http://www.khronos.org/opencl/">OpenCL</a> 1.0 specification. Ακόμη η συγκεκριμένη version διορθώνει 200 γνωστά bugs που υπήρχαν, προσφέρει υποστήριξη για το <a class="zem_slink" title="CUDA" rel="homepage" href="http://www.nvidia.com/object/cuda_home.html">CUDA</a> Toolkit 3.0 (το οποίο θα κυκλοφορήσει μέσα στις επόμενες βδομάδες<a href="http://away.gr/2009/11/02/geforce-19539-beta-drivers-from-nvidia/">)[next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
</div>
<div class="post hentry"><a name="8045293802019195261"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/olympus-pen.html">Αντίστροφη μέτρηση για τη νέα Olympus Pen;</a></h3>
<div class="post-body entry-content">
<div><a href="http://t3mag.gr/news/2118">Αντίστροφη μέτρηση για τη νέα Olympus Pen; &#124; T3 Magazine</a></p>
<blockquote><p><a class="thickbox" title="olympuse-p2-leaklg1.jpg" rel="sitenews" href="http://t3mag.gr/files/mainimages/olympuse-p2-leaklg1.jpg"><img src="http://t3mag.gr/files/imagecache/main_image_sitenews/files/mainimages/olympuse-p2-leaklg1.jpg" alt="olympuse-p2-leaklg1.jpg" /></a>Πριν από λίγες ημέρες είδαν το φως τις δημοσιότητας κάποιες πληροφορίες που έφεραν την Olympus έτοιμη να παρουσιάσει τη νέα έκδοση της επιτυχημένης Pen E-P1 που έκανε ντεμπούτο μόλις τον περασμένο Ιούνιο. Οι φήμες φουντώνουν μέρα με τη μέρα και δεν είναι λίγοι αυτοί που εμφανίζονται πεπεισμένοι ότι η εταιρεία θα κάνει τις επίσημες ανακοινώσεις σε ελάχιστες ημέρες.[<a href="http://t3mag.gr/news/2118">next</a>]
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
</div>
<div class="post hentry"><a name="929551973236188477"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/google-voice.html">Σύντομα το Google Voice και εκτός Αμερικής;</a></h3>
<div class="post-body entry-content">
<div><a href="http://www.digitalnews.gr/google-voice-global">Σύντομα το Google Voice και εκτός Αμερικής; &#124; DigitalNews</a></p>
<blockquote><p><a href="http://images.google.com/imgres?imgurl=http://blogs.voices.com/thebiz/google-voice-activated-search.jpg&#38;imgrefurl=http://blogs.voices.com/thebiz/2007/12/say_what_google_building_voice_activated_search_engine.html&#38;usg=__me2VCcmFwD2dFBOh33fjzOLXP8w=&#38;h=300&#38;w=233&#38;sz=27&#38;hl=en&#38;start=4&#38;tbnid=DBGJk0h3DETAsM:&#38;tbnh=116&#38;tbnw=90&#38;prev=/images%3Fq%3DGoogle%2BVoice%26ndsp%3D18%26hl%3Den%26lr%3D%26sa%3DN%26start%3D1"><img style="border:1px solid;" src="http://t1.gstatic.com/images?q=tbn:DBGJk0h3DETAsM:http://blogs.voices.com/thebiz/google-voice-activated-search.jpg" alt="" width="90" height="116" /></a>Σύμφωνα με πληροφορίες που διέρρευσαν από λάθος (?) η υπηρεσία τηλεφωνίας <a class="zem_slink" title="Google" rel="homepage" href="http://google.com">Google</a> <a href="http://images.google.com/imgres?imgurl=http://www.matthewrex.com/wp/wp-content/uploads/2009/07/GoogleVoicePreview.jpg&#38;imgrefurl=http://www.matthewrex.com/2009/07/13/icon-google-voice/&#38;usg=__1G70DyQhhlrFU0tv6ueDvRpKmdI=&#38;h=360&#38;w=500&#38;sz=151&#38;hl=en&#38;start=14&#38;tbnid=uvC5UJyB029NTM:&#38;tbnh=94&#38;tbnw=130&#38;prev=/images%3Fq%3DGoogle%2BVoice%26ndsp%3D18%26hl%3Den%26lr%3D%26sa%3DN%26start%3D1"><img style="border:1px solid;" src="http://t1.gstatic.com/images?q=tbn:uvC5UJyB029NTM:http://www.matthewrex.com/wp/wp-content/uploads/2009/07/GoogleVoicePreview.jpg" alt="" width="130" height="94" /></a>Voice της Google πρόκειται να γίνει διαθέσιμη σύντομα και εκτός Αμερικής. Η Google υποστηρίζει ότι ήδη υπάρχουν 1.419 εκατομμύρια χρήστες εγγεγραμμένοι ενώ 570.000 χρήστες χρησιμοποιούν την υπηρεσία κάθε μέρα.[<a href="http://www.digitalnews.gr/google-voice-global">next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
<div class="post-footer">
<div class="post-footer-line post-footer-line-1"><span class="post-author vcard"> <span class="fn"> </span> </span> <span class="post-timestamp"> <a class="timestamp-link" title="permanent link" rel="bookmark" href="http://xolloth.blogspot.com/2009/11/google-voice.html"><abbr class="published" title="2009-11-02T11:23:00+02:00"></abbr></a></span><span class="post-comment-link"><br />
</span><span class="post-icons"><span class="item-action"> </span></span><span class="post-backlinks post-comment-link"> </span></div>
<div class="post-footer-line post-footer-line-2"><span class="post-labels"> </span></div>
<p>&#160;</p>
</div>
<p>&#160;</p>
</div>
<div class="post hentry"><a name="6939388011278771730"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/blog-post_4060.html">Παρέμβαση της Κομισιόν για την προστασία του ανταγωνισμού στην ιταλική τηλεπικοινωνιακή αγορά</a></h3>
<div class="post-body entry-content">
<div><a href="http://tech.pathfinder.gr/tech/lawnet/792415.html">Παρέμβαση της Κομισιόν για την προστασία του ανταγωνισμού στην ιταλική τηλεπικοινωνιακή αγορά &#8211; Pathfinder Tech</a></p>
<blockquote><p><img src="http://pf.pstatic.gr/CMAN/i/46/I6/792415-180407050.jpg" alt="" width="128" height="70" />Με επιστολή της προς την ιταλική ρυθμιστική αρχή τηλεπικοινωνιών (AGCOM), η Ευρωπαϊκή Επιτροπή υπογραμμίζει την ανάγκη διασφάλισης του υγιούς, αποτελεσματικού ανταγωνισμού στην τηλεπικοινωνιακή αγορά μέσα από τις συμφωνίες που προωθεί η αρχή με την Telecom Italia, τη μεγαλύτερη τηλεπικοινωνιακή εταιρία της γειτονικής χώρας<a href="http://tech.pathfinder.gr/tech/lawnet/792415.html">.[next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
<div class="post-footer">
<div class="post-footer-line post-footer-line-1"><span class="post-author vcard"> <span class="fn"> </span> </span> <span class="post-timestamp"> <a class="timestamp-link" title="permanent link" rel="bookmark" href="http://xolloth.blogspot.com/2009/11/blog-post_4060.html"><abbr class="published" title="2009-11-02T11:22:00+02:00"></abbr></a></span><span class="post-comment-link"><br />
</span><span class="post-icons"><span class="item-action"> </span></span><span class="post-backlinks post-comment-link"> </span></div>
<div class="post-footer-line post-footer-line-2"><span class="post-labels"> </span></div>
<p>&#160;</p>
</div>
<p>&#160;</p>
</div>
<div class="post hentry"><a name="252642569323247675"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/microsoft-yahoo.html">Καθυστερεί η εφαρμογή της συμφωνίας Microsoft-Yahoo!</a></h3>
<div class="post-body entry-content">
<div><a href="http://e-pcmag.gr/news/61405">Καθυστερεί η εφαρμογή της συμφωνίας Microsoft-Yahoo! &#124; PC Magazine</a></p>
<blockquote><p><a class="lightbox-processed" rel="lightshow[field_first_image][]" href="http://e-pcmag.gr/files/microsoft_yahoo_0.jpg"><img src="http://e-pcmag.gr/files/imagecache/news_fullnode/microsoft_yahoo_0.jpg" alt="" width="250" height="188" /></a>Έχουν περάσει αρκετές εβδομάδες από τότε που η Microsoft και η Yahoo! ανακοίνωσαν τη συνεργασία τους όσον αφορά το χώρο της διαδικτυακής αναζήτησης, ωστόσο φαίνεται ότι υπάρχουν ακόμα κάποια εμπόδια που καθυστερούν την υλοποίηση της απόφασης. Αρκεί μια επίσκεψη στη μηχανή αναζήτησης της Yahoo! για να διαπιστώσει κανείς ότι δεν έχει αλλάξει τίποτα ακόμα.<a href="http://e-pcmag.gr/news/61405">[next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
<p>&#160;</p>
</div>
<div class="post hentry"><a name="8540599652379361963"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/blog-post_8763.html">Η ψηφιοποίηση των βιβλίων διχάζει</a></h3>
<div class="post-body entry-content">
<div><a href="http://www.pcw.gr/Article/Latest/digital_books_google_publishing_consortium/179-4700.html">Η ψηφιοποίηση των βιβλίων διχάζει &#8211; ψηφιακά βιβία, Google, digital books, εκδοτικοί οίκοι &#8211; PCW</a></p>
<blockquote><p><a href="http://images.google.com/imgres?imgurl=http://www.microlix.gr/blog/wp-content/uploads/2009/01/googlemicrolixgr.jpg&#38;imgrefurl=http://www.thasosnet.gr/Greg/2009/09/07/&#38;usg=__4k7YHHQh9l9y1GmoLTFJd4WPRNw=&#38;h=250&#38;w=320&#38;sz=25&#38;hl=en&#38;start=8&#38;tbnid=4mmXr5CJRFSgGM:&#38;tbnh=92&#38;tbnw=118&#38;prev=/images%3Fq%3D%25CF%2588%25CE%25B7%25CF%2586%25CE%25B9%25CE%25BF%25CF%2580%25CE%25BF%25CE%25AF%25CE%25B7%25CF%2583%25CE%25B7%2B%25CF%2584%25CF%2589%25CE%25BD%2B%25CE%25B2%25CE%25B9%25CE%25B2%25CE%25BB%25CE%25AF%25CF%2589%25CE%25BD%26ndsp%3D18%26hl%3Den%26lr%3D%26sa%3DN%26start%3D1"><img style="border:1px solid;" src="http://t3.gstatic.com/images?q=tbn:4mmXr5CJRFSgGM:http://www.microlix.gr/blog/wp-content/uploads/2009/01/googlemicrolixgr.jpg" alt="" width="118" height="92" /></a>Η μεγάλη προσπάθεια που έχει ξεκινήσει η Google με κόστος περίπου 100 εκατομμύρια ευρώ, για να φτιάξει ένα τεράστιο αποθετήριο ψηφιοποιημένων βιβλίων (http://books.google.com) έχει προκαλέσει αντικρουόμενες αντιδράσεις και έντονη πολιτική δραστηριότητα. Πρόσφατα η Google ήρθε σε συνεννόηση με ένα consortium μεγάλων Αμερικάνων εκδοτών στους οποίους προσέφερε ένα ποσοστό με αντάλλαγμα την άδεια χρήσης τίτλων κατοχυρωμένων με πνευματικά δικαιώματα. Η συμφωνία αυτή είναι αμοιβαία επικερδής, αλλά από την άλλη πλευρά εντείνει τις αντιδράσεις των πολιτικών, εκδοτών και ακαδημαϊκών, οι οποίοι δεν θέλουν να αφήσουν την Google να μονοπωλήσει το χώρο του ψηφιακού βιβλίου<a href="http://www.pcw.gr/Article/Latest/digital_books_google_publishing_consortium/179-4700.html">[next]</a>
<p>&#160;</p>
</blockquote>
</div>
<p>&#160;</p>
</div>
<div class="post-footer">
<div class="post-footer-line post-footer-line-1"><span class="post-author vcard"> <span class="fn"> </span> </span> <span class="post-timestamp"> <a class="timestamp-link" title="permanent link" rel="bookmark" href="http://xolloth.blogspot.com/2009/11/blog-post_8763.html"><abbr class="published" title="2009-11-02T11:19:00+02:00"></abbr></a></span><span class="post-comment-link"><br />
</span><span class="post-icons"><span class="item-action"> </span></span><span class="post-backlinks post-comment-link"> </span></div>
<div class="post-footer-line post-footer-line-2"><span class="post-labels"> </span></div>
<p>&#160;</p>
</div>
<p>&#160;</p>
</div>
<p><a name="6804889316995184252"></a></p>
<h3 class="post-title entry-title"><a href="http://xolloth.blogspot.com/2009/11/blog-post_7081.html">Αυξάνονται οι συνδέσεις στο Διαδίκτυο</a></h3>
<div><a href="http://tech.pathfinder.gr/tech/792391.html">Αυξάνονται οι συνδέσεις στο Διαδίκτυο σε μικρές επιχειρήσεις και νοικοκυριά &#8211; Pathfinder Tech</a></p>
<blockquote><p><img title="Το 27% δεν έχει τις ικανότητες και τις δεξιότητες για να χειριστεί το Διαδίκτυο" src="http://pf.pstatic.gr/CMAN/i/45/I6/792391-180405702.jpg?r=1257148863" alt="Το 27% δεν έχει τις ικανότητες και τις δεξιότητες για να χειριστεί το Διαδίκτυο" width="358" height="238" />Αθήνα: Εντυπωσιακή είναι η αύξηση στις συνδέσεις στο Διαδίκτυο τόσο στις μικρές επιχειρήσεις όσο και στα νοικοκυριά από το 2005 μέχρι το 2008 βάσει έρευνας του Παρατηρητηρίου για την Κοινωνία της Πληροφορίας. Όπως αναφέρουν «Τα Νέα», ενώ το 2005 το 75% των μικρών επιχειρήσεων είχαν πρόσβαση στο Διαδίκτυο, τρία χρόνια μετά το 2008 το ποσοστό αυτό ανέβηκε στο 83%.[<a href="http://tech.pathfinder.gr/tech/792391.html">next]</a>
<p>&#160;</p>
</blockquote>
</div>
<div class="zemanta-pixie" style="margin-top:10px;height:15px;"><a class="zemanta-pixie-a" title="Reblog this post [with Zemanta]" href="http://reblog.zemanta.com/zemified/7cf9b37c-cfa3-8aad-9a4c-918bcde6eb0c/"><img class="zemanta-pixie-img" style="border:medium none;float:right;" src="http://img.zemanta.com/reblog_e.png?x-id=7cf9b37c-cfa3-8aad-9a4c-918bcde6eb0c" alt="Reblog this post [with Zemanta]" /></a></div>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Number crunching on Nvidia Fermi]]></title>
<link>http://bison.wordpress.com/2009/10/31/number-crunching-on-nvidia-fermi/</link>
<pubDate>Sat, 31 Oct 2009 15:20:34 +0000</pubDate>
<dc:creator>bison</dc:creator>
<guid>http://bison.wordpress.com/2009/10/31/number-crunching-on-nvidia-fermi/</guid>
<description><![CDATA[After the Nvidia GPU conference I&#8217;ve been reading up on the Fermi architecture. Now for gaming]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>After the <a href="http://www.nvidia.com/object/gpu_technology_conference.html">Nvidia GPU conference</a> I&#8217;ve been reading up on the <a href="http://www.nvidia.com/object/fermi_architecture.html">Fermi</a> architecture. Now for gaming, the card is cool, but you quickly realize that gaming is not its targeted audience. It just happens to be that the work I do has a lot to do with images and doing stuff in parallel. The thing with images are that they can usually be very of very high resolution and the output might consist of voxels which also might be of high resolution. So far bandwidth has been a big bottleneck. Even though the PCI-express is incredibly fast, the calculations might actually be faster than the transfer. Therefore if one applies <a href="http://en.wikipedia.org/wiki/SIMD">SIMD</a> instructions and take advantage of the considerably larger memories on <a href="http://en.wikipedia.org/wiki/64-bit">64-bit</a> machines thinks are done faster since you don&#8217;t need to shuffle stuff around.</p>
<p>This has been, in my own humble opinion, a big issue for the professional market. Of course at the GPU conference many companies shared their thoughts on the subject and showed that although this is a limitation, it is possible to work around. The bad news of course is that it requires one to have good knowledge of how <a href="http://www.nvidia.com/object/cuda_home.html">CUDA</a> utilies the raw power of graphics card. Such knowledge takes around a year or two and hasn&#8217;t been much of an option for companies without proper R&#38;D.</p>
<p>Thankfully Nvidia acknowledges the issue and therefore they implemented 64-bit memory addressing capabilities on the Fermi card (actually the first gen is going to use like 40 bits) which in turn allows for those 8 GB graphics card. For the professional market this is the right way to go, and programmers will always be happy with more memory (if a bit more lazy).</p>
<p>The next thing that blew me out of the water is the raw computation power of the Fermi architecture. Now, for single precision calculations graphic cards have been viable for quite some time. But when doing stuff in a more scientific way you always want double precision. <a href="http://en.wikipedia.org/wiki/Cell_(microprocessor)">The Cell processor</a> has so far been a real killer at this, and if the memory wasn&#8217;t so lousy on the Playstation 3 could possibly become the weapon of choice (4 SIMD registers per SPU if I recall correctly, proved 10 times faster number crunching than Intel processor). Now the Fermi has 512 cores, which I think would translate into registers? Anyway, with 16 Streaming Multiprocessors each with 32 single precision calculations per cycle the Fermi card shows incredible potential. Especially as doing double precision is not suppose to be slower (giving 16 calculations per cycle). Now this I won&#8217;t know before testing, if given the opportunity of course. Still the sum becomes 256 double precision calculations per cycle compared to 2 per core on a normal processor. Even with new quad cores with utilizing hyper threading and having 2-4 times the clock rate the Fermi should still outperform considerably.</p>
<p>I can only assume the <a href="http://en.wikipedia.org/wiki/Larrabee_(GPU)">Larrabee</a> will blow Fermi out the water (because it has to), and there fore the parallel computing power on the horizon must be crazy. Exciting times for the performance geek. Also, <a href="http://developer.nvidia.com/object/nexus.html">Nexus</a> allows for proper debugging of CUDA (FINALLY!) which will allow companies to capitalize on this sick performance more quickly.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Spacerując po ....]]></title>
<link>http://wmuza.wordpress.com/2009/10/30/spacerujac-po/</link>
<pubDate>Fri, 30 Oct 2009 11:44:07 +0000</pubDate>
<dc:creator>Wojciech Muża</dc:creator>
<guid>http://wmuza.wordpress.com/2009/10/30/spacerujac-po/</guid>
<description><![CDATA[Prawie jak Batman &#8230;.:) bo tak młodzi Wejherowianie nazywają pomnik Jakuba Wejhera stojący w ce]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Prawie jak Batman &#8230;.:) bo tak młodzi Wejherowianie nazywają pomnik Jakuba Wejhera stojący w centralnym miejscu naszego rynku (o którym kilka ciekawostek napiszę później). Założyciel Wejherowa jako młodzieniec był pokojowcem Królewicza Władysława, gdy dorósł stał się żołnierzem walczacym o wolność Najjaśniejszej, pełnił funkcję wojewody malborskiego oraz generała ziem pruskich w wojnie ze szwedami. Warto zgłębiać historie bo kto wie (z przymrużeniem oka oczywiście) może nasz Jakub Wejher był pierwowzorem amerykańskiego batmana <img src='http://s.wordpress.com/wp-includes/images/smilies/icon_wink.gif' alt=';)' class='wp-smiley' />    </p>
<p><a href="http://wmuza.files.wordpress.com/2009/10/p_1600_1200_b21a0b02-f6e8-47c7-8ecc-23ff061b043a.jpeg"><img src="http://wmuza.files.wordpress.com/2009/10/p_1600_1200_b21a0b02-f6e8-47c7-8ecc-23ff061b043a.jpeg?w=225&#038;h=300" alt="" width="225" height="300" class="alignnone size-full wp-image-364" /></a></p>
</div>]]></content:encoded>
</item>

</channel>
</rss>
