Running the new version of our "Sampling" program with the Assembly Language built into the program we were able to get an execution time of:
[jadach@aarchie spo600_20181_inline_assembler_lab]$ time ./vol_simd
Generating sample data.
Scaling samples.
Summing samples.
Result: -574
real 0m0.027s
user 0m0.018s
sys 0m0.009s
COMPARED TO our last iteration of :
[jadach@aarchie b_lab5]$ time ./vol2
Result: 416
real 0m3.229s
user 0m3.073s
sys 0m0.150s
========================
WHAT CHANGED ?
========================
Inside the new code we are using ASM to multiple multiple sets of data with a single instruction, so we are using SIMD to optimize the code.
The results above shows the difference in performance after the ASM code was added.
Comments