hugomardbrink/resource-optimised-DCT
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add outline for software optimisation

Browse source
This commit is contained in:
Hugo Mårdbrink 2024-03-30 23:12:02 +01:00
parent c493857e6c
commit 29b5365918
1 changed files with 12 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

15
Design_and_analysis.md
View file

@ -67,12 +67,15 @@ For the initial and naive software implementation, some hardware configurations
- L1 instruction cache size: 16kB - L1 instruction cache size: 16kB
- L1 data cache size: 64kB - L1 data cache size: 64kB
- L2 cache size: 256kB - L2 cache size: 256kB
- Vector length: 256 - Vector length: 256 bits
- Element length: 64 - Element length: 64 bits
- Number of threads: 1 - Number of cores & threads: 1
- L1 cache associativity: 2 - L1 cache associativity: 2
- L2 cache associativity: 8 - L2 cache associativity: 8
These are by no means optimal configurations, but to find the optimal, we need to first optimise the software implementation.
Making these configurations good enough as a baseline.
### Constants and definitions ### Constants and definitions
Throughout the code, several constants and definitions are defined for ease to try different configurations. These are defined in the following way: Throughout the code, several constants and definitions are defined for ease to try different configurations. These are defined in the following way:
- `DCT_SIZE` is the size of the DCT block - `DCT_SIZE` is the size of the DCT block
@ -135,5 +138,11 @@ void dct_2d(element_t** matrix_in, element_t** matrix_out) {
} }
``` ```
### Software optimisations
## Compile time constants
## Flattening arrays and loops
## Vectorisation
## Changing data types
## Compiler optimisations