add linalg and linalg_xdsl matmul f32

[superlopuh]

No description provided.

[superlopuh]

oh yeah I added linalg here as well since it just works

[compor]

linalg == "MLIR" flow? right?

[superlopuh]

note that this isn't the input to "ours" as we cannot yet legalize it for f32 in xDSL

[github-actions]

kernels,baseline,linalg,linalg_xdsl,snitch_stream,snrt
conv2d_d1_s1_3x3 4x4xf64,667,858,308,,
ddot 128xf64,956,965,,213,577
dense 8x8xf64,3206,3530,,2741,2723
fill 4x4xf64,50,50,64,,
matmul 4x16x8xf64,2495,2694,708,,
matmul_transb 4x16x16xf32,3386,5038,,845,849
matmul_transb 4x16x16xf64,,5142,1295,,
pooling_nchw_max_d1_s2_3x3 4x4xf64,584,484,275,,
pooling_nchw_sum_d1_s2_3x3 4x4xf64,902,832,271,,
relu 4x4xf64,142,125,72,,
relu 4x8xf32,297,210,,67,85
saxpy 64xf32,634,634,,,140
sum 4x4xf64,129,162,87,,
sum 4x8xf32,238,247,87,,72
sum 8x8xf16,821,,65,,

[github-actions]

kernels,baseline,linalg,linalg_xdsl,snitch_stream,snrt
conv2d_d1_s1_3x3 4x4xf64,0.22,0.17,0.58,,
ddot 128xf64,0.13,0.13,,0.64,0.22
dense 8x8xf64,0.20,0.18,,0.26,0.26
fill 4x4xf64,0.02,0.02,0.28,,
matmul 4x16x8xf64,0.21,0.19,0.82,,
matmul_transb 4x16x16xf32,0.21,0.20,,0.76,0.79
matmul_transb 4x16x16xf64,,0.20,0.89,,
pooling_nchw_max_d1_s2_3x3 4x4xf64,0.39,0.30,0.65,,
pooling_nchw_sum_d1_s2_3x3 4x4xf64,0.22,0.17,0.66,,
relu 4x4xf64,0.13,0.14,0.25,,
relu 4x8xf32,0.33,0.16,,0.28,0.22
saxpy 64xf32,0.10,0.10,,,0.46
sum 4x4xf64,0.13,0.10,0.20,,
sum 4x8xf32,0.14,0.13,0.20,,0.24
sum 8x8xf16,0.31,,0.26,,

[github-actions]

kernels,baseline,linalg,linalg_xdsl,snitch_stream,snrt
conv2d_d1_s1_3x3 4x4xf64,667,858,308,,
ddot 128xf64,956,965,,213,577
dense 8x8xf64,3206,3530,,2741,2723
fill 4x4xf64,50,50,64,,
matmul 4x16x8xf64,2495,2694,708,,
matmul_transb 4x16x16xf32,3386,5038,,845,849
matmul_transb 4x16x16xf64,,5142,1295,,
pooling_nchw_max_d1_s2_3x3 4x4xf64,584,484,275,,
pooling_nchw_sum_d1_s2_3x3 4x4xf64,902,832,271,,
relu 4x4xf64,142,125,72,,
relu 4x8xf32,297,210,,67,85
saxpy 64xf32,634,634,,,140
sum 4x4xf64,129,162,87,,
sum 4x8xf32,238,247,87,,72
sum 8x8xf16,821,,65,,

[github-actions]

kernels,baseline,linalg,linalg_xdsl,snitch_stream,snrt
conv2d_d1_s1_3x3 4x4xf64,0.22,0.17,0.58,,
ddot 128xf64,0.13,0.13,,0.64,0.22
dense 8x8xf64,0.20,0.18,,0.26,0.26
fill 4x4xf64,0.02,0.02,0.28,,
matmul 4x16x8xf64,0.21,0.19,0.82,,
matmul_transb 4x16x16xf32,0.21,0.20,,0.76,0.79
matmul_transb 4x16x16xf64,,0.20,0.89,,
pooling_nchw_max_d1_s2_3x3 4x4xf64,0.39,0.30,0.65,,
pooling_nchw_sum_d1_s2_3x3 4x4xf64,0.22,0.17,0.66,,
relu 4x4xf64,0.13,0.14,0.25,,
relu 4x8xf32,0.33,0.16,,0.28,0.22
saxpy 64xf32,0.10,0.10,,,0.46
sum 4x4xf64,0.13,0.10,0.20,,
sum 4x8xf32,0.14,0.13,0.20,,0.24
sum 8x8xf16,0.31,,0.26,,