Mps matrix multiplication kernel

[ivarflakstad]

Output from running cargo run --release --example mps-matrix-multiplication --features=mps

Correctness:
[=================================================] ✅

Performance:
Generating input matrices: (f32 4096x4096 and f16 4096x4096)
Running with transpose left: false, transpose right: false, alpha: 1, beta: 0
[=================================================] ✅
Avg GFLOPS: 2714.017354808345
Total time: 2.534734497s
Avg time: 50.694689ms

Running with transpose left: true, transpose right: false, alpha: 1, beta: 0
[=================================================] ✅
Avg GFLOPS: 2544.2273256244875
Total time: 2.702094294s
Avg time: 54.041885ms

Running with transpose left: false, transpose right: true, alpha: 1, beta: 0
[=================================================] ✅
Avg GFLOPS: 2696.778482443825
Total time: 2.550065042s
Avg time: 51.0013ms

Running with transpose left: false, transpose right: false, alpha: 0.5, beta: 0
[=================================================] ✅
Avg GFLOPS: 2668.0019304379844
Total time: 2.577410417s
Avg time: 51.548208ms

Running with transpose left: false, transpose right: false, alpha: 1, beta: 0.5
[=================================================] ✅
Avg GFLOPS: 2667.1376704607605
Total time: 2.578379166s
Avg time: 51.567583ms