Synergy-Plus configuration

I often use my laptop together with a main desktop.
To save time i share the keyboard and the mouse utilizing a small utility called Synergy
that was introduced to me by my colleague Miki Tebeka

Assuming two computers: one MacOS - macbook pro and one PC Windows 7 and
I want to use the mouse and keyboard of the Windows machine:

The configuration on windows to be a server is quite tricky:
Options t; Add the network name of both the computer:
FMILO-DESKTOP
FMILO-MACOS
and then add
 0-100 % of left of FMILO-DESKTOP to 0-100% of FMILO-MACOS
 0-100 % of right of FMILO-MACOS to 0-100% of FMILO-DESKTOP

Here 0 - 100 % means how much of the side as a line should be sensitive to the mouse crossing.

To debug the session is quite useful to look at the log console while in Test Mode: Just right click (on Windows) in the dock icon and set show log Messages.

Hope this is helpful to someone else.

Prefix Sum on CUDA

Is not as trivial as it seems. I just finished to implement one and I can tell what you need
read the scan Gpu Gems 3 Article in particular chapter **39.3.1 Stream Compaction**.

To implement your own start from the LargeArrayScan example in the SDK, that will give you just the prescan. Assuming you have the selection array in device memory (an array of 1 and 0 meaning 1- select 0 - discard), dev_selection_array a dev_elements_array elements to be selected a dev_prescan_array and a dev_result_array all of size N then you do

   prescan(dev_prescan_array,dev_selection_array, N);
   scatter(dev_result_array, dev_prescan_array,
            dev_selection_array, dev_elements_array, N);

where the scatter is

    __global__ void scatter_kernel( T*dev_result_array,
                      const T* dev_prescan_array,
                      const T* dev_selection_array,
                      const T* dev_elements_array, std::size_t size){

 unsigned int idx = blockIdx.x * blockDim.x + threadIdx.x;
 if (idx >= size) return;
 if (dev_selection_array[idx] == 1){
  dev_result_array[dev_prescan_array[idx]] = dev_elements_array[idx];
 }
   }

for other nice application of the prescan see the paper Ble93

Otherwise you can use the CUDPP library where such primitives exists and are highly optimized even more that the sdk one.

New Blog Visual scheme

Reading the latimes today I found this new cool feature for customizing blogger appearance.

This is a fist update version, when I have more time I will play more with. Well done google!

C++ Tricky

Nothing better than a bit of C++ for breakfast:

Reading A StackOverflow post I hacked a bit with an Object called SaveRestore.
The trick is to know that when the object gets out of scope { }, is destroyed.

You need to maintain a reference to the object and its value.

Many Core and Reconfigurable SuperComputing Conference

I will speak at the Many Core and Reconfigurable SuperComputing Conference in Rome the 22-24 March 2010
with a talk entitled
A GPU Based Architecture for Distributing Dictionary Attacks to OpenPGP Secret Keyrings
To see all the other exciting talks: http://www.mrsc2010.eu/Programme