adds LoopRotate transformation

[lukastruemper]

No description provided.

[daisytuner]

Daisytuner Report - mlir_torch_models (chamomile)

@@                                   Benchmarks                                   @@
=====================================================================================
  Benchmark              Time        ΔTime       Thr         Energy      ΔEnergy     
=====================================================================================
# bn_conv_bn_relu_maxpool_torch18.72 s     +0.30%      N/A         3614.85 J   -0.95%      
# bn_conv_bn_relu_maxpool_run_none3.27 s      +0.51%      N/A         657.20 J    -0.45%      
# bn_conv_bn_relu_maxpool_run_sequential3.29 s      -0.21%      N/A         664.88 J    -1.25%      
# bn_conv_bn_relu_maxpool_run_openmp3.38 s      +3.17%      N/A         690.93 J    +2.81%      
# bn_conv_bn_relu_maxpool_run_cuda3.70 s      -0.28%      N/A         724.29 J    -1.27%      

[daisytuner]

Daisytuner Report - python_npbench (zinnia)

@@                                   Benchmarks                                   @@
=====================================================================================
  Benchmark              Time        ΔTime       Thr         Energy      ΔEnergy     
=====================================================================================
# adi_numpy              1.31 s      -2.26%      N/A         130.89 J    -1.94%      
# adi_omp                15.04 s     -1.42%      N/A         1490.30 J   -1.60%      
# adi_cuda               4.72 s      -0.24%      N/A         457.83 J    -0.27%      
# adi_seq_tuning         16.00 s     -0.20%      N/A         1512.42 J   -0.05%      
# atax_numpy             2.16 s      +0.20%      N/A         223.39 J    +0.31%      
# atax_omp               2.99 s      +0.56%      N/A         377.74 J    +0.67%      
# atax_cuda              4.11 s      +0.27%      N/A         423.28 J    +0.25%      
# atax_seq_tuning        3.72 s      -0.46%      N/A         376.86 J    -0.43%      
# gemm_numpy             1.21 s      -1.06%      N/A         192.57 J    -1.02%      
# gemm_omp               1.11 s      -0.14%      N/A         161.64 J    -0.01%      
# gemm_cuda              10.58 s     -0.43%      N/A         1006.31 J   -0.38%      
# gemm_seq_tuning        1.11 s      -0.00%      N/A         161.68 J    +0.15%      
# gesummv_numpy          1.75 s      +0.65%      N/A         249.81 J    +0.52%      
# gesummv_omp            1.95 s      -1.59%      N/A         305.92 J    -2.13%      
# gesummv_cuda           8.32 s      -0.38%      N/A         1000.51 J   -0.36%      
# gesummv_seq_tuning     6.67 s      +0.29%      N/A         814.08 J    +0.03%      
# gemver_numpy           1.08 s      +0.20%      N/A         166.61 J    +0.08%      
# gemver_omp             862.62 ms   +0.41%      N/A         112.68 J    +1.07%      
# gemver_cuda            3.86 s      +0.11%      N/A         386.34 J    -0.07%      
# gemver_seq_tuning      4.49 s      -0.13%      N/A         434.78 J    -0.09%      
# k2mm_numpy             1.20 s      +0.23%      N/A         195.93 J    +0.17%      
# k2mm_omp               3.56 s      +0.86%      N/A         663.85 J    +0.18%      
# k2mm_cuda              13.56 s     -0.44%      N/A         1286.59 J   -0.45%      
# k2mm_seq_tuning        3.64 s      +0.60%      N/A         469.79 J    +0.75%      
# k3mm_numpy             1.03 s      +0.34%      N/A         182.41 J    +0.03%      
# k3mm_omp               5.55 s      -0.33%      N/A         953.94 J    +0.22%      
# k3mm_cuda              19.80 s     +0.02%      N/A         1869.73 J   +0.16%      
# k3mm_seq_tuning        5.75 s      +0.79%      N/A         800.62 J    +1.15%      
# mvt_numpy              2.43 s      -0.04%      N/A         249.22 J    +0.02%      
# mvt_omp                2.74 s      -0.06%      N/A         284.40 J    +0.04%      
# mvt_cuda               3.34 s      -0.28%      N/A         341.16 J    -0.36%      
# mvt_seq_tuning         2.74 s      -0.13%      N/A         284.19 J    -0.13%      
# symm_numpy             784.41 ms   +0.59%      N/A         80.68 J     +0.49%      
# symm_omp               6.06 s      +0.09%      N/A         596.79 J    -0.15%      
# symm_seq_tuning        8.43 s      -0.26%      N/A         804.59 J    -0.20%      
# syr2k_numpy            887.89 ms   -0.21%      N/A         90.39 J     +0.06%      
# syr2k_omp              9.86 s      -0.40%      N/A         937.39 J    -0.35%      
# syr2k_cuda             1.62 s      -0.95%      N/A         168.76 J    -1.10%      
# syr2k_seq_tuning       9.82 s      -0.23%      N/A         934.56 J    -0.21%      
# syrk_numpy             770.40 ms   -0.45%      N/A         79.42 J     -0.34%      
# syrk_omp               5.96 s      +0.09%      N/A         574.08 J    +0.19%      
# syrk_cuda              1.51 s      +0.56%      N/A         158.11 J    +0.51%      
# syrk_seq_tuning        5.94 s      -0.36%      N/A         572.14 J    -0.23%      
# trmm_numpy             874.80 ms   -0.67%      N/A         89.18 J     -0.89%      
# trmm_omp               694.93 ms   -0.21%      N/A         89.66 J     +0.15%      
# trmm_seq_tuning        3.39 s      -0.78%      N/A         324.09 J    -0.68%      

[daisytuner]

Daisytuner Report - mlir_torch_layers (chamomile)

@@                                   Benchmarks                                   @@
=====================================================================================
  Benchmark              Time        ΔTime       Thr         Energy      ΔEnergy     
=====================================================================================
# batchnorm_torch        18.98 s     -0.14%      N/A         3720.84 J   +4.40%      
# batchnorm_run_none     6.19 s      -2.39%      N/A         1204.99 J   +2.41%      
# batchnorm_run_sequential6.49 s      -0.53%      N/A         1264.07 J   +4.24%      
# batchnorm_run_openmp   5.76 s      +0.70%      N/A         1357.99 J   +4.65%      
# batchnorm_run_cuda     8.07 s      -0.54%      N/A         1583.17 J   +4.89%      
# conv2d_torch           18.57 s     +0.15%      N/A         3648.18 J   +5.01%      
# conv2d_run_openmp      5.03 s      +0.57%      N/A         1228.26 J   +5.08%      
# conv2d_run_cuda        6.83 s      -0.36%      N/A         1339.46 J   +4.85%      
# linear_torch           6.09 s      -0.77%      N/A         1463.54 J   +1.47%      
# linear_run_none        11.69 s     +0.29%      N/A         3191.00 J   +2.78%      
# linear_run_sequential  9.92 s      -0.07%      N/A         2762.25 J   +3.43%      
# linear_run_openmp      9.71 s      -0.29%      N/A         2859.92 J   +2.90%      
# linear_run_cuda        9.21 s      -0.01%      N/A         1802.41 J   +5.88%      
# matmul_torch           6.08 s      -2.94%      N/A         1466.98 J   -0.01%      
# matmul_run_none        11.58 s     +0.15%      N/A         3149.69 J   +2.52%      
# matmul_run_sequential  9.96 s      -0.03%      N/A         2777.63 J   +3.99%      
# matmul_run_openmp      9.76 s      +0.05%      N/A         2877.49 J   +3.27%      
# matmul_run_cuda        9.03 s      -0.38%      N/A         1765.23 J   +4.85%      
# pooling_torch          25.62 s     -0.77%      N/A         5109.91 J   +3.71%      
# pooling_run_none       25.08 s     -1.39%      N/A         4812.43 J   +3.24%      
# pooling_run_sequential 25.73 s     -0.47%      N/A         4931.15 J   +4.12%      
# pooling_run_openmp     17.15 s     -1.41%      N/A         3660.41 J   +3.07%      
# pooling_run_cuda       31.18 s     -0.58%      N/A         6071.61 J   +4.46%      
# relu_torch             18.98 s     -0.19%      N/A         3717.68 J   +4.30%      
# relu_run_none          5.24 s      +0.08%      N/A         1021.21 J   +4.40%      
# relu_run_sequential    6.33 s      -0.63%      N/A         1232.34 J   +4.05%      
# relu_run_openmp        5.81 s      +1.58%      N/A         1360.20 J   +5.55%      
# relu_run_cuda          8.29 s      -0.91%      N/A         1624.15 J   +4.29%