Team Slytherin: Difference between revisions

From Cpre584
Jump to navigation Jump to search
Chidvila (talk | contribs)
Bhavani (talk | contribs)
 
(44 intermediate revisions by 2 users not shown)
Line 1: Line 1:
=== Team Members ===
== Team Members ==
* Kevin Townsend
* Kevin Townsend
* Bhavani Satyanarayana Rao
* Bhavani Satyanarayana Rao
* Chidvila Gaddam
* Chidvila Gaddam


=== Assignment1 ===
== Assignment1 ==
#Review Docs
<hr width='100%'>
##Convey PDK Reference Manual
Custom Personality Development:
Analyze Application
Evaluate Hardware Options
Define Custom Instructions
Develop Software Model of Custom Personality
Modify Application to Use Coprocessor
Compile Application with Convey Compiler
Simulate Application with Convey Architecture Simulator
Develop FPGA Hardware
Simulate Hardware in Convey Simulation Environment
Integrate with Convey Hardware


##Convey Programmers Guide
=== Running the sample application with script files ===
##Convey Reference Manual


<div>
<ul><li>The runcp script located in the SampleAppVadd can be used to run the application in HW which sets the appropriate environment variables and runs the UserApp.exe </li>


#Understand how Vector Adder example works
    export CNY_PDK_PROJ = ~/pdk_sample/cae_pers_vadd
For the implementation of the vector adder example the make file from the SampleAppVadd is run to generate the UserApp execution file. The UserApp.c file is the software code for the addition implementation. It assumes a default vector size of 100, however can inline arguments for different sizes.
    ./runcp
It gets vector PDK signature and
##Allocates co processor memory for 2 arrays of numbers.
##Polpulates arrays with 64 integers.
##Call the co processor rooutine to sum the numbers.
It also compares the actual output with the co processor output and prints message based on true or false.


The UserApp.c calls for the cpTestEx1 and cpTestEx2 routines which are implemented in cpTest1.s
<li>The run script located in the SampleAppVadd can be used to run the application in SW</li>
The makefile compiles the C code(UserApp.c) and the assembly code(cpTest1.s) using the GNU compiler to generate the UserApp.o with both x86-64 and Convey coprocessor instructions.
<li>The scripts above can be used to run the application on Simulator using the "-vsim" option which sets the environment variable CNY_CAE_EMULATOR to ./run_simulation</li>


The CaeSimPers acts as an emulator for the output.
  ./runcp -vsim
      or
  ./run -vsim
</ul>
</div>


#Implement Vector Adder example  ( SW, ModelSim, HW)
The implementation of the example in Software(host processor), Modelsim(Simulator) or the hardware(Co processor) can be realized by setting the appropriate environmental values for CNY_SIM_THREAD and CNY_CAE_EMULATOR.


{| class="wikitable"
<table border='1px'>
|-
<tr><th>Environment Variables</th><th>Software</th><th>ModelSim</th><th>Hardware</th></tr>
!
<tr><td>CNY_SIM_THREAD</td><td>libcpSimLib2.so</td><td>libcpSimLib2.so</td><td>''unset''</td></tr>
! Software(Host Processor)
<tr><td>CNY_CAE_EMULATOR</td><td><location for CaeSimPers></td><td>./run_simulation</td><td>''unset''</td></tr>
! Modelsim(Simulator)
</table>
! Hardware(Co Processor)
|-
| CNY_SIM_THREAD
| libcpSimLib2.so
| libcpSimLib2.so
|
|-
| CNY_CAE_EMULATOR
|
| path for CaeSimPers
|
|}


Alternatively, running the application using run, runcp, run simulator from the SampleAppVadd also helps realize the same.
 
 
=== Changes made to run the application on ModelSim ===
<ul>
<li>Setup as explained in here [[http://wikis.ece.iastate.edu/cpre584/index.php/Convey_environment_setup]]</li>
<li>Set the CNY_PDK_PROJ environment variable to the location of cae_pers_vadd</li>
 
      export CNY_PDK_PROJ = ~/pdk_sample/cae_pers_vadd
 
<li>Copied Jones' my_makefile from the temp dir and replaced the makefile in CY_PDK_PROJ/testbench</li>
<li>Add line to the makefile in testbench dir</li>
    CNY_PDK_MODELSIM_USER_SIM_OPTIONS= -i -do "run -all"
<li>Modify the values of LIB_DIR and LDFLAGS from 32-bit to 64-bit as given below</li>
    LIB_DIR = lib64
    LDFLAGS = -m64
</ul>
 
[[Image:Modelsim.PNG|500px|ModelSim Snapshot]]
 
=== Extra Credit: Exploring gdb ===
 
<ul><li>We invoked gdb using the gdb <programname> command or by using the run/runcp script files in with "-gdb" option</li>
   
    gdb UserApp.exe
      or
    ./run -gdb
   
 
    [[Image:gdb1.png|500px|GDB Snapshot]]
 
<li>One can find different command classes on gdb by typing help.To get details on a particular command class/command, type help on gdb followed by the name of the command class/command </li>
   
    [[Image:gdb2.png|500px|GDB help]]
<li>We explored some commands like run, start (to begin execution), stepi (for single instruction stepping into a file), kill (kill a running program) and quit on gdb </li>
   
    [[Image:gdb3.png|500px|GDB]]   
 
</ul>
<hr width='100%'>
== Assignment2 ==
<hr width='100%'>
=== About Sobel Algorithm ===
Sobel Algorithm is used for edge detection in image processing.The link[http://dasl.mem.drexel.edu/alumni/bGreen/www.pages.drexel.edu/_weg22/edge.html] explains
<ul><li>Why edge detection is important?</li>
<li>How sobel algorithm uses the gradient method (jump in the pixel values along edges in the images compared to the pixels around it) to detect edges in images?</li></ul>
==== The Algorithm ====
<ul><li>Use Sobel masks to find the gradient of a pixel w.r.t its 8 immediate neighboring pixel values.
The Sobel masks are GX={(-1,0,+1),(-2,0,+2),(-1,0,+1)) which gives the gradient in the x direction
and Gy={(+1,+2,+1),(0,0,0),(-1,-2,-1)} which gives the gradient in the x direction.</li>
<li>Find the approximate magnitude of the gradient using the formula</li>
    |G|=|Gx|+|Gy|
<li>Compare |G| against the threshold values (0 and 255 in our case).</li>
    If (|G| > 255) then the pixel value is 0 in the output image.
    If (|G| < 0) then the pixel value is 255 in the output image.
<li>Repeat the above steps for all the pixels in the image</li>
</ul>
 
=== C code for Sobel Algorithm ===
 
Source: Edge Detection Tutorial-Author: Bill Green (2002)[http://dasl.mem.drexel.edu/alumni/bGreen/www.pages.drexel.edu/_weg22/edge.html]
 
  for(Y=0; Y<=(originalImage.rows-1); Y++)  {
        for(X=0; X<=(originalImage.cols-1); X++)  {
            sumX = 0;
            sumY = 0;
            /* image boundaries */
            if(Y==0 || Y==originalImage.rows-1)
                  SUM = 0;
            else if(X==0 || X==originalImage.cols-1)
                  SUM = 0;
            /* Convolution starts here */
            else  {
              /*-------X GRADIENT APPROXIMATION------*/
              for(I=-1; I<=1; I++)  {
                  for(J=-1; J<=1; J++)  {
                      sumX = sumX + (int)( (*(originalImage.data + X + I + (Y + J)*originalImage.cols)) * GX[I+1][J+1]);
                  }
              }
              if(sumX>255)  sumX=255;
              if(sumX<0)    sumX=0;
              /*-------Y GRADIENT APPROXIMATION-------*/
              for(I=-1; I<=1; I++)  {
                  for(J=-1; J<=1; J++)  {
                      sumY = sumY + (int)( (*(originalImage.data + X + I + (Y + J)*originalImage.cols)) * GY[I+1][J+1]);
                  }
              }
              if(sumY>255)  sumY=255;
              if(sumY<0)    sumY=0;
              SUM = abs(sumX) + abs(sumY); /*---GRADIENT MAGNITUDE APPROXIMATION (Myler p.218)----*/
            }
            *(edgeImage.data + X + Y*originalImage.cols) = 255 - (unsigned char)(SUM);  /* make edges black and background white */
            fwrite( (edgeImage.data + X + Y*originalImage.cols), sizeof(char), 1, bmpOutput);
 
=== Profiling C code Using gprof ===
 
Compile the c code with -pg option
  gcc -pg -lm edgesob.c
Run the a.out file
  ./a.out lena.bmp
Execute the gprof command, provide it with the executable file name and pipe the output to a text file.
Note the use of -z option.Unless the `-z' option is given, functions with no apparent time spent in them,
and no apparent calls to them, are not mentioned.
source:The GNU Profiler-Jay Fenlason and Richard Stallman[http://www.cs.utah.edu/dept/old/texinfo/as/gprof_toc.html]
  gprof -b -z a.out >result.txt
The output of gprof:
The link [http://www.cs.utah.edu/dept/old/texinfo/as/gprof.html#SEC5] explains on how to understand a Flat Profile.
====Flat profile====
 
  Each sample counts as 0.01 seconds.
    %  cumulative  self              self    total
  time  seconds  seconds    calls  Ts/call  Ts/call  name
  100.31      0.03    0.03                            main
  0.00      0.03    0.00        4    0.00    0.00  getImageInfo
  0.00      0.03    0.00        1    0.00    0.00  copyColorTable
  0.00      0.03    0.00        1    0.00    0.00  copyImageInfo
  0.00      0.03    0.00                            __do_global_ctors_aux
  0.00      0.03    0.00                            __do_global_dtors_aux
  0.00      0.03    0.00                            __gmon_start__
  0.00      0.03    0.00                            __libc_csu_fini
  0.00      0.03    0.00                            __libc_csu_init
  0.00      0.03    0.00                            _fini
  0.00      0.03    0.00                            _init
  0.00      0.03    0.00                            _start
  0.00      0.03    0.00                            atexit
  0.00      0.03    0.00                            call_gmon_start
  0.00      0.03    0.00                            data_start
  0.00      0.03    0.00                            frame_dummy
 
====Call graph====
The link [http://www.cs.utah.edu/dept/old/texinfo/as/gprof.html#SEC6] explains on how to read the call graph.
  granularity: each sample hit covers 2 byte(s) for 33.23% of 0.03 seconds
  index % time    self  children    called    name
                                                <spontaneous>
  [1]    100.0    0.03    0.00                main [1]
                  0.00    0.00      4/4          getImageInfo [2]
                  0.00    0.00      1/1          copyImageInfo [4]
                  0.00    0.00      1/1          copyColorTable [3]
  -----------------------------------------------
                  0.00    0.00      4/4          main [1]
  [2]      0.0    0.00    0.00      4        getImageInfo [2]
  -----------------------------------------------
                  0.00    0.00      1/1          main [1]
  [3]      0.0    0.00    0.00      1        copyColorTable [3]
  -----------------------------------------------
                  0.00    0.00      1/1          main [1]
  [4]      0.0    0.00    0.00      1        copyImageInfo [4]
  -----------------------------------------------
                                                  <spontaneous>
  [7]      0.0    0.00    0.00                atexit [7]
  -----------------------------------------------
                                                  <spontaneous>
  [8]      0.0    0.00    0.00                call_gmon_start [8]
  -----------------------------------------------
                                                  <spontaneous>
  [9]      0.0    0.00    0.00                data_start [9]
  -----------------------------------------------
                                                  <spontaneous>
  [10]    0.0    0.00    0.00                frame_dummy [10]
  -----------------------------------------------
                                                  <spontaneous>
  [11]    0.0    0.00    0.00                __do_global_ctors_aux [11]
  -----------------------------------------------
                                                  <spontaneous>
  [12]    0.0    0.00    0.00                __do_global_dtors_aux [12]
  -----------------------------------------------
                                                  <spontaneous>
  [13]    0.0    0.00    0.00                __gmon_start__ [13]
  -----------------------------------------------
                                                  <spontaneous>
  [14]    0.0    0.00    0.00                __libc_csu_fini [14]
  -----------------------------------------------
                                                  <spontaneous>
  [15]    0.0    0.00    0.00                __libc_csu_init [15]
  -----------------------------------------------
                                                  <spontaneous>
  [16]    0.0    0.00    0.00                _fini [16]
  -----------------------------------------------
                                                  <spontaneous>
  [17]    0.0    0.00    0.00                _init [17]
  -----------------------------------------------
                                                  <spontaneous>
  [18]    0.0    0.00    0.00                _start [18]
  -----------------------------------------------
 
=== Software emulation of HW ===
 
Application file : [[Team Slytherin - Software emulation source code]]

Latest revision as of 21:26, 22 February 2012

Team Members

  • Kevin Townsend
  • Bhavani Satyanarayana Rao
  • Chidvila Gaddam

Assignment1


Running the sample application with script files

  • The runcp script located in the SampleAppVadd can be used to run the application in HW which sets the appropriate environment variables and runs the UserApp.exe
  • export CNY_PDK_PROJ = ~/pdk_sample/cae_pers_vadd ./runcp
  • The run script located in the SampleAppVadd can be used to run the application in SW
  • The scripts above can be used to run the application on Simulator using the "-vsim" option which sets the environment variable CNY_CAE_EMULATOR to ./run_simulation
  • ./runcp -vsim or ./run -vsim


Environment VariablesSoftwareModelSimHardware
CNY_SIM_THREADlibcpSimLib2.solibcpSimLib2.sounset
CNY_CAE_EMULATOR<location for CaeSimPers>./run_simulationunset


Changes made to run the application on ModelSim

  • Setup as explained in here [[1]]
  • Set the CNY_PDK_PROJ environment variable to the location of cae_pers_vadd
  • export CNY_PDK_PROJ = ~/pdk_sample/cae_pers_vadd
  • Copied Jones' my_makefile from the temp dir and replaced the makefile in CY_PDK_PROJ/testbench
  • Add line to the makefile in testbench dir
  • CNY_PDK_MODELSIM_USER_SIM_OPTIONS= -i -do "run -all"
  • Modify the values of LIB_DIR and LDFLAGS from 32-bit to 64-bit as given below
  • LIB_DIR = lib64 LDFLAGS = -m64

Error creating thumbnail: File missing

Extra Credit: Exploring gdb

  • We invoked gdb using the gdb <programname> command or by using the run/runcp script files in with "-gdb" option
  • gdb UserApp.exe or ./run -gdb Error creating thumbnail: File missing
  • One can find different command classes on gdb by typing help.To get details on a particular command class/command, type help on gdb followed by the name of the command class/command
  • Error creating thumbnail: File missing
  • We explored some commands like run, start (to begin execution), stepi (for single instruction stepping into a file), kill (kill a running program) and quit on gdb
  • Error creating thumbnail: File missing

Assignment2


About Sobel Algorithm

Sobel Algorithm is used for edge detection in image processing.The link[2] explains

  • Why edge detection is important?
  • How sobel algorithm uses the gradient method (jump in the pixel values along edges in the images compared to the pixels around it) to detect edges in images?

The Algorithm

  • Use Sobel masks to find the gradient of a pixel w.r.t its 8 immediate neighboring pixel values. The Sobel masks are GX={(-1,0,+1),(-2,0,+2),(-1,0,+1)) which gives the gradient in the x direction and Gy={(+1,+2,+1),(0,0,0),(-1,-2,-1)} which gives the gradient in the x direction.
  • Find the approximate magnitude of the gradient using the formula
  • |G|=|Gx|+|Gy|
  • Compare |G| against the threshold values (0 and 255 in our case).
  • If (|G| > 255) then the pixel value is 0 in the output image. If (|G| < 0) then the pixel value is 255 in the output image.
  • Repeat the above steps for all the pixels in the image

C code for Sobel Algorithm

Source: Edge Detection Tutorial-Author: Bill Green (2002)[3]

 for(Y=0; Y<=(originalImage.rows-1); Y++)  {
       for(X=0; X<=(originalImage.cols-1); X++)  {
            sumX = 0;
            sumY = 0;
            /* image boundaries */
            if(Y==0 || Y==originalImage.rows-1)
                 SUM = 0;
            else if(X==0 || X==originalImage.cols-1)
                 SUM = 0;
            /* Convolution starts here */
            else   {
              /*-------X GRADIENT APPROXIMATION------*/
              for(I=-1; I<=1; I++)  {
                  for(J=-1; J<=1; J++)  {
                     sumX = sumX + (int)( (*(originalImage.data + X + I + (Y + J)*originalImage.cols)) * GX[I+1][J+1]);
                  }
              }
              if(sumX>255)  sumX=255;
              if(sumX<0)    sumX=0;
              /*-------Y GRADIENT APPROXIMATION-------*/
              for(I=-1; I<=1; I++)  {
                  for(J=-1; J<=1; J++)  {
                      sumY = sumY + (int)( (*(originalImage.data + X + I + (Y + J)*originalImage.cols)) * GY[I+1][J+1]);
                  }
              }
              if(sumY>255)   sumY=255;
              if(sumY<0)     sumY=0;
              SUM = abs(sumX) + abs(sumY); /*---GRADIENT MAGNITUDE APPROXIMATION (Myler p.218)----*/
            }
            *(edgeImage.data + X + Y*originalImage.cols) = 255 - (unsigned char)(SUM);  /* make edges black and background white */
            fwrite( (edgeImage.data + X + Y*originalImage.cols), sizeof(char), 1, bmpOutput);

Profiling C code Using gprof

Compile the c code with -pg option

 gcc -pg -lm edgesob.c

Run the a.out file

 ./a.out lena.bmp

Execute the gprof command, provide it with the executable file name and pipe the output to a text file. Note the use of -z option.Unless the `-z' option is given, functions with no apparent time spent in them, and no apparent calls to them, are not mentioned. source:The GNU Profiler-Jay Fenlason and Richard Stallman[4]

 gprof -b -z a.out >result.txt

The output of gprof: The link [5] explains on how to understand a Flat Profile.

Flat profile

 Each sample counts as 0.01 seconds.
   %   cumulative   self              self     total
  time   seconds   seconds    calls  Ts/call  Ts/call  name
 100.31      0.03     0.03                             main
 0.00      0.03     0.00        4     0.00     0.00  getImageInfo
 0.00      0.03     0.00        1     0.00     0.00  copyColorTable
 0.00      0.03     0.00        1     0.00     0.00  copyImageInfo
 0.00      0.03     0.00                             __do_global_ctors_aux
 0.00      0.03     0.00                             __do_global_dtors_aux
 0.00      0.03     0.00                             __gmon_start__
 0.00      0.03     0.00                             __libc_csu_fini
 0.00      0.03     0.00                             __libc_csu_init
 0.00      0.03     0.00                             _fini
 0.00      0.03     0.00                             _init
 0.00      0.03     0.00                             _start
 0.00      0.03     0.00                             atexit
 0.00      0.03     0.00                             call_gmon_start
 0.00      0.03     0.00                             data_start
 0.00      0.03     0.00                             frame_dummy

Call graph

The link [6] explains on how to read the call graph.

 granularity: each sample hit covers 2 byte(s) for 33.23% of 0.03 seconds
 index % time    self  children    called     name
                                                <spontaneous>
 [1]    100.0    0.03    0.00                 main [1]
                 0.00    0.00       4/4           getImageInfo [2]
                 0.00    0.00       1/1           copyImageInfo [4]
                 0.00    0.00       1/1           copyColorTable [3]
 -----------------------------------------------
                 0.00    0.00       4/4           main [1]
 [2]      0.0    0.00    0.00       4         getImageInfo [2]
 -----------------------------------------------
                 0.00    0.00       1/1           main [1]
 [3]      0.0    0.00    0.00       1         copyColorTable [3]
 -----------------------------------------------
                 0.00    0.00       1/1           main [1]
 [4]      0.0    0.00    0.00       1         copyImageInfo [4]
 -----------------------------------------------
                                                  <spontaneous>
 [7]      0.0    0.00    0.00                 atexit [7]
 -----------------------------------------------
                                                  <spontaneous>
 [8]      0.0    0.00    0.00                 call_gmon_start [8]
 -----------------------------------------------
                                                  <spontaneous>
 [9]      0.0    0.00    0.00                 data_start [9]
 -----------------------------------------------
                                                  <spontaneous>
 [10]     0.0    0.00    0.00                 frame_dummy [10]
 -----------------------------------------------
                                                  <spontaneous>
 [11]     0.0    0.00    0.00                 __do_global_ctors_aux [11]
 -----------------------------------------------
                                                  <spontaneous>
 [12]     0.0    0.00    0.00                 __do_global_dtors_aux [12]
 -----------------------------------------------
                                                  <spontaneous>
 [13]     0.0    0.00    0.00                 __gmon_start__ [13]
 -----------------------------------------------
                                                  <spontaneous>
 [14]     0.0    0.00    0.00                 __libc_csu_fini [14]
 -----------------------------------------------
                                                  <spontaneous>
 [15]     0.0    0.00    0.00                 __libc_csu_init [15]
 -----------------------------------------------
                                                  <spontaneous>
 [16]     0.0    0.00    0.00                 _fini [16]
 -----------------------------------------------
                                                  <spontaneous>
 [17]     0.0    0.00    0.00                 _init [17]
 -----------------------------------------------
                                                  <spontaneous>
 [18]     0.0    0.00    0.00                 _start [18]
 -----------------------------------------------

Software emulation of HW

Application file : Team Slytherin - Software emulation source code