Commits · 44c580131db370a46684c11ad52852434ad1e5b5 · Gaurav Kukreja / supercomputer_lab

26 Jun, 2012 1 commit
- using Strategy No 0 i.e. Parallel Minimax · 44c58013
  LRZ-SIM Funktionskennung authored Jun 25, 2012
```
 * Fix SIGSEGV on run
Signed-off-by: LRZ-SIM Funktionskennung <h039val@ice1-login.ice1.cos.lrz.de>
```
  44c58013
25 Jun, 2012 5 commits
- parallel-minimax Fix: Compiles. Does not run. · 51af7d43
  Gaurav Kukreja authored Jun 25, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  51af7d43
- Merge branch 'master' of https://code.google.com/p/supercomputer · c8cdde56
  Gaurav Kukreja authored Jun 25, 2012
  
  c8cdde56
- Merge branch 'master' of https://code.google.com/p/supercomputer · 680d90f7
  Gaurav Kukreja authored Jun 25, 2012
```
Conflicts:
	mpi-assign4/merged_code/heat
	mpi-assign4/merged_code/relax_jacobi.o
```
  680d90f7
- Removed Compiled Binaries · ee75f0ec
  Gaurav Kukreja authored Jun 25, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  ee75f0ec
- New folders for par and seq code Assign 5 · 76b05310
  Gaurav Kukreja authored Jun 25, 2012
```
New folders for Parallel-Minimax, and Sequential.
Parallel-Minimax compiles, but has bugs to be fixed.
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  76b05310
24 Jun, 2012 4 commits
- Updated presentation4 with correct seq values · f851a893
  Viktor Bogischef authored Jun 24, 2012
```
Also undid an unnecessary change for Jacobi. In the previos version a send- and receive buffer was used to communicate with the top and the bottom neighbor. This is not necessary and costs some performance.
```
  f851a893
- Merge branch 'master' of https://code.google.com/p/supercomputer · 634efad5
  Gaurav Kukreja authored Jun 24, 2012
  
  634efad5
- Sequential Minimax for Abalone · 41fe5b36
  Gaurav Kukreja authored Jun 24, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  41fe5b36
- Merged nonoptimised code and cleaned up folders · bd70a432
  Viktor Bogischef authored Jun 24, 2012
  
  bd70a432
22 Jun, 2012 3 commits
- Merge branch 'master' of https://code.google.com/p/supercomputer · 60edba14
  Viktor Bogischef authored Jun 22, 2012
  
  60edba14
- Merged all the code + minor bugfixes for jacobi · fec5043e
  Viktor Bogischef authored Jun 22, 2012
  
  fec5043e
- Changes to Presentation and code by Viktor · 94304c6d
  Gaurav Kukreja authored Jun 22, 2012
```
 * Changes provided by Viktor have been pushed in gauss1by4 and gauss4by1
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  94304c6d
21 Jun, 2012 2 commits
- Presentation for Assignment 4 · f8cf7da1
  Gaurav Kukreja authored Jun 21, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  f8cf7da1
- Added Jacobi and Gauss-nonopt · 7ff31f99
  Viktor Bogischef authored Jun 21, 2012
```
Signed-off-by: Viktor Bogischef <bogischv@informatik.tu-muenchen.de>
```
  7ff31f99
20 Jun, 2012 1 commit
- Vanilla Code for Assignment 5 · c721897a
  Gaurav Kukreja authored Jun 20, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  c721897a
17 Jun, 2012 2 commits

Gaurav Kukreja authored Jun 17, 2012

 * Gauss 2D Non-blocking works for 2x2, 1x4, 4x1
 * Gauss special code for 1x4
 * Gauss special code for 4x1
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

ae696c66

Sequential Optimization of Gauss · 099699b9
Gaurav Kukreja authored Jun 17, 2012
```
 * residual interleaving
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
099699b9

06 Jun, 2012 2 commits
- Matrix Borders Initialization for Assign 4 · c083da69
  Gaurav Kukreja authored Jun 06, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  c083da69
- Assignment 4 Vanilla Code in assign4 folder · a5651abc
  Gaurav Kukreja authored Jun 06, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  a5651abc
23 May, 2012 1 commit

Assign2 Heat manual Parallelization · 378b8d21

Gaurav Kukreja authored May 23, 2012

 * 16 threads, change in Makefile
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

378b8d21

19 May, 2012 1 commit
- Minor Bugs in script and bandwidth code resolved · 383950c6
  Gaurav Kukreja authored May 19, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  383950c6
18 May, 2012 3 commits
- Batch Scripts · 1b859709
  Gaurav Kukreja authored May 18, 2012
```
 * assignment 3 scripts are super awesome! - Gaurav
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  1b859709
- Removed debug message from bandwidth · 490b19b0
  Gaurav Kukreja authored May 18, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  490b19b0
- Solved minor bug in ping-pong.c · d2aa6af4
  Gaurav Kukreja authored May 18, 2012
```
 * Gaurav
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
  d2aa6af4
16 May, 2012 3 commits

Reduction Code updated · 3e829627

Gaurav Kukreja authored May 16, 2012

 * checks if the number of threads is a power of 2 - Viktor
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

3e829627

Bandwidth Measurement · cf295448

Gaurav Kukreja authored May 16, 2012

 * ping-pong-bandwidth.c - Gaurav
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

cf295448

ping-pong.c for Latency · ababf07b

Gaurav Kukreja authored May 16, 2012

Working Code - Umbreen
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

ababf07b

14 May, 2012 1 commit

Assignment 3 - MPI · fd2df844

Gaurav Kukreja authored May 14, 2012

 * ping-pong for latency code done by Umbreen - buggy
 * reduction done by Viktor
 * result.c done by Gaurav
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

fd2df844

09 May, 2012 4 commits

Added test1.dat · 7aeae8b9

Gaurav Kukreja authored May 09, 2012

 - resolution 1024 to 4096 with step size 1024
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

7aeae8b9

Autoparallel, num-threads=8 · 797c08cd
Gaurav Kukreja authored May 09, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
797c08cd

Autoparallelization · 11603e91

Gaurav Kukreja authored May 09, 2012

 - Added Compiler options in Makefile
 - Added hint in relax_jacobi.c\
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

11603e91

Created a copy of sequential code in head-parallel · b7b0f5df
Gaurav Kukreja authored May 09, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
b7b0f5df

02 May, 2012 7 commits

Offset Calculation · a2c0ce6e
Gaurav Kukreja authored May 02, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
a2c0ce6e
Interleaving of row processing · eb2eef7a
Gaurav Kukreja authored May 02, 2012
```
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
eb2eef7a

Loop Blocking · 3006f1f7

Gaurav Kukreja authored May 02, 2012

To improve Cache Hit Ratio for larger resolutions
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

3006f1f7

Cutting the residual calculation · 40d2c9da

Gaurav Kukreja authored May 02, 2012

Instead of calculating residual for the next iteration, we calculate residual
iteration. This is an acceptable deviation, and a significant optimization.
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

40d2c9da

Interleaved Residaul Calculation · 51562a59
Gaurav Kukreja authored May 02, 2012
```
Improves Cache Hit Ratio
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>
```
51562a59

Swapped Pointers for u and uhelp after iteration · 476a1e0f

Gaurav Kukreja authored May 02, 2012

Copying of matrix uhelp to u each time is avoided, by swapping pointers.
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

476a1e0f

Changed the access pattern into matrix · dea4cf6b

Gaurav Kukreja authored May 02, 2012

Exchanged the nested loops in relax_jacobi and residual_jacobi to ensure
linear memory access.
Signed-off-by: Gaurav Kukreja <mailme.gaurav@gmail.com>

dea4cf6b