[Midend] Enhancements and Optimizations for batch matmul and convolution [Examples] Added MLIRLinalg Examples for Various Optimization Options.

[6somehow]

[Midend] Enhancements and Optimizations

• BatchMatMul:

  • Updated unroll, tile, and vectorization strategies for BatchMatMul operations.
  • Enhanced SCF (Structured Control Flow) version for BatchMatMul to support further optimization and work.

• Conv2D NHWC-FHWC:

  • Improved vectorization and tiling techniques for Conv2D operations in NHWC-FHWC layout.
  • These updates provide a vectorization option and a tile+vectorization option for buddy-opt.

• Depthwise Conv2D NHWC-HWC:

  • Introduced vectorization and tiling strategies for Depthwise Conv2D in NHWC-HWC format, ensuring more efficient execution.

[Examples] Added MLIRLinalg Examples for Various Optimization Options

New MLIRLinalg examples were added to demonstrate the usage of new buddy-opt various optimization options , including:

1. conv-nhwc-fhwc-optimize: Conv2D NHWC-FHWC layout vectorization optimization.
2. conv-nhwc-fhwc-tile-optimize: Conv2D NHWC-FHWC vectorization with tiling optimizations.
3. depthwise-conv-nhwc-hwc-optimize: Depthwise Conv2D NHWC-HWC
   vectorization layout optimization.
4. batchmatmul-tile-optimize: tiling and unroll optimization for BatchMatMul.
5. batchmatmul-scf-optimize: using SCF optimization for BatchMatMul.

Example MLIR Files:

• batchmatmul
• conv2d_nhwc_fhwc
• depthwise_conv_2d_nhwc_hwc

These updates collectively improve the performance of matrix and convolution operations in MLIR, providing optimized patterns for common workloads like BatchMatMul and convolution layers.

[xlinsist]

…ion [Examples] Added MLIRLinalg Examples for Various Optimization Options. fixed thirdparty.

You can copy the description from the former PR(#384) into this one.

[xlinsist]

Do our optimized methods need to add into both MatMulOptimization and BatchMatMulOptimization?

[xlinsist]

LGTM, thanks!