Difference: JaysJournal (9 vs. 10)

Revision 102010-05-15 - jayzhang

Line: 1 to 1

META TOPICPARENT	name="NGSAlignerProject"

05/06/10

I'm starting this journal 3 days late (whoops). Here's a brief overview of what happened during my first three days:

Line: 82 to 82

(Much later) Add in my improvements to the original local aligner.
(Still much later) Polish up the different aligners and have CMake compile either the vectorized or non-vectorized versions of the aligners based on CPU support.

Added:

>
>

05/14/10

I finished the vectorized banded implementation with all test cases passing! I really thought it would take more time and I was surprised when all the test cases passed. The implementation works only with square matrices, and uses 16x 8-bit vectors, so it should be quite a speed-up from the regular banded implementation. The strange thing with 8-bit vectors is that the instruction set is missing a few things compared to 16-bit vectors. For example, there's no insert or extract function, so I ended up casting the __m128i 's into an int8_t array and inserting/extracting manually; I'm not really sure if that's a good idea, though. Also, there was no max function for signed integers, so I had to write my own for that as well.

For Monday:

Edit the CMakeList.txt files so that it compiles vector/non-vector versions based whether SSE2 is available or not.
Polish up all the code I've written and merge.
Add in my improvements to the original local aligner.

View topic | History: r73 < r72 < r71 < r70 | More topic actions...