metal: implement OUR_PROD for F32

[HankBO]

Overview

Support OUT_PROD op partially for metal backend: <F32, F32>.

Add test cases for OUT_PROD performance testing.

ref #14909. close my previous PR #23724 and open a new one to focus on f32.

Additional information

ran testing command with build/bin/test-backend-ops test -o OUT_PROD -b MTL0 and all test cases passed on my m4 air.

Testing 3 devices

Backend 1/3: MTL0
  Device description: Apple M4
  Device memory: 18186 MB (18185 MB free)

  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[1,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=1,bs=[3,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[1,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=1,k=16,bs=[3,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[1,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=1,bs=[3,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[1,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,1],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,1],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,1],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,1],nr=[2,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,3],nr=[1,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,3],nr=[1,2],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,3],nr=[2,1],trans_b=0): �[1;32mOK�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=256,n=16,k=16,bs=[3,3],nr=[2,2],trans_b=0): �[1;32mOK�[0m

did comparison with my perf testing case ./bin/test-backend-ops perf -o OUT_PROD -b MTL0, build/bin/test-backend-ops perf -o OUT_PROD -b CPU

Testing 3 devices

Backend 1/3: MTL0
  Device description: Apple M4
  Device memory: 18186 MB (18185 MB free)

  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=32,bs=[1,1],nr=[1,1],trans_b=0):                7654 runs -   194.66 us/run -     8768 kB/run - �[1;34m  42.96 GB/s�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=64,bs=[1,1],nr=[1,1],trans_b=0):                3592 runs -   326.16 us/run -     9344 kB/run - �[1;34m  27.32 GB/s�[0m
  Backend MTL0: �[1;32mOK�[0m
Backend 2/3: BLAS
  Device description: Accelerate
  Device memory: 0 MB (0 MB free)

  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=32,bs=[1,1],nr=[1,1],trans_b=0):                3827 runs -   314.65 us/run -     8768 kB/run - �[1;34m  26.57 GB/s�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=64,bs=[1,1],nr=[1,1],trans_b=0):                3592 runs -   398.90 us/run -     9344 kB/run - �[1;34m  22.34 GB/s�[0m
  Backend BLAS: �[1;32mOK�[0m
Backend 3/3: CPU
  Skipping CPU backend
3/3 backends passed
�[1;32mOK�[0m

Testing 3 devices

Backend 1/3: MTL0
  Skipping
Backend 2/3: BLAS
  Skipping
Backend 3/3: CPU
  Device description: Apple M4
  Device memory: 24576 MB (24576 MB free)

  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=32,bs=[1,1],nr=[1,1],trans_b=0):                 957 runs -  3108.74 us/run -     8768 kB/run - �[1;34m   2.69 GB/s�[0m
  OUT_PROD(type_a=f32,type_b=f32,m=4096,n=512,k=64,bs=[1,1],nr=[1,1],trans_b=0):                 898 runs -  6023.31 us/run -     9344 kB/run - �[1;34m   1.48 GB/s�[0m
  Backend CPU: �[1;32mOK�[0m
3/3 backends passed
�[1;32mOK�[0m

Requirements

• I have read and agree with the contributing guidelines
• AI usage disclosure: Yes, I used AI to help me learn the codebase, and refine my coding. I do the design and implementation by myself.