simplifies GPU transformations

[lukastruemper]

No description provided.

[daisytuner]

Daisytuner Report - mlir_torch_models (chamomile)

@@                                   Benchmarks                                   @@
=====================================================================================
  Benchmark              Time        ΔTime       Thr         Energy      ΔEnergy     
=====================================================================================
# resnet18_torch         19.90 s     -0.06%      N/A         3998.00 J   +1.49%      
# resnet18_docc_none     17.92 s     -0.36%      N/A         4750.79 J   -0.12%      
# resnet18_docc_sequential17.59 s     +0.02%      N/A         4657.03 J   +0.55%      
# resnet18_docc_openmp   23.57 s     +0.11%      N/A         6946.06 J   +0.51%      
# resnet18_docc_cuda     5.50 s      +1.99%      N/A         1071.28 J   +1.60%      

[daisytuner]

Daisytuner Report - python_npbench (zinnia)

@@                                   Benchmarks                                   @@
=====================================================================================
  Benchmark              Time        ΔTime       Thr         Energy      ΔEnergy     
=====================================================================================
# adi_numpy              1.31 s      +0.73%      N/A         137.78 J    +2.87%      
# adi_omp                2.74 s      +2.98%      N/A         356.63 J    +4.33%      
# adi_cuda               3.74 s      +1.74%      N/A         387.13 J    +3.88%      
# adi_seq_tuning         2.59 s      +1.67%      N/A         251.88 J    +3.57%      
# atax_numpy             2.14 s      -0.38%      N/A         233.10 J    +1.88%      
# atax_omp               3.13 s      +1.21%      N/A         413.02 J    +3.43%      
# atax_cuda              4.28 s      +0.03%      N/A         460.96 J    +2.04%      
# atax_seq_tuning        2.86 s      -1.05%      N/A         296.94 J    +1.84%      
# gemm_numpy             1.24 s      +2.51%      N/A         203.43 J    +3.42%      
# gemm_omp               1.17 s      +1.13%      N/A         173.69 J    +2.35%      
# gemm_cuda              10.73 s     +0.14%      N/A         1077.75 J   +2.59%      
# gemm_seq_tuning        1.16 s      +0.79%      N/A         171.70 J    +1.36%      
# gesummv_numpy          1.76 s      +0.28%      N/A         259.76 J    +1.37%      
# gesummv_omp            2.18 s      +1.78%      N/A         360.34 J    +3.10%      
# gesummv_cuda           5.48 s      +3.32%      N/A         757.17 J    +2.20%      
# gesummv_seq_tuning     4.77 s      +2.19%      N/A         625.78 J    +1.63%      
# gemver_numpy           1.08 s      -0.24%      N/A         171.25 J    +0.33%      
# gemver_omp             942.56 ms   -1.15%      N/A         126.05 J    -3.04%      
# gemver_cuda            2.55 s      +0.76%      N/A         280.79 J    +2.65%      
# gemver_seq_tuning      1.62 s      -4.20%      N/A         149.14 J    +2.16%      
# k2mm_numpy             1.20 s      -0.13%      N/A         201.42 J    +0.77%      
# k2mm_omp               3.60 s      +0.46%      N/A         678.65 J    -0.42%      
# k2mm_cuda              12.78 s     +0.27%      N/A         1279.84 J   +2.76%      
# k2mm_seq_tuning        3.04 s      +1.77%      N/A         416.56 J    +1.96%      
# k3mm_numpy             1.03 s      +0.59%      N/A         186.33 J    +0.65%      
# k3mm_omp               5.61 s      -0.11%      N/A         989.21 J    +2.27%      
# k3mm_cuda              18.50 s     +0.34%      N/A         1842.84 J   +2.88%      
# k3mm_seq_tuning        4.96 s      +0.27%      N/A         705.32 J    +0.02%      
# mvt_numpy              2.41 s      -1.11%      N/A         258.72 J    +0.98%      
# mvt_omp                2.77 s      -0.37%      N/A         302.82 J    +1.92%      
# mvt_cuda               3.44 s      -0.17%      N/A         369.30 J    +2.23%      
# mvt_seq_tuning         2.77 s      -0.54%      N/A         301.89 J    +1.71%      
# symm_numpy             784.91 ms   +0.92%      N/A         85.00 J     +3.34%      
# symm_omp               1.07 s      +6.80%      N/A         132.96 J    +8.67%      
# symm_seq_tuning        1.81 s      -4.28%      N/A         159.69 J    +1.47%      
# syr2k_numpy            885.80 ms   +1.83%      N/A         94.60 J     +3.90%      
# syr2k_omp              983.41 ms   +3.20%      N/A         113.84 J    +4.53%      
# syr2k_cuda             1.64 s      +3.57%      N/A         179.35 J    +5.20%      
# syr2k_seq_tuning       980.63 ms   +3.51%      N/A         113.41 J    +4.59%      
# syrk_numpy             778.02 ms   +1.76%      N/A         84.16 J     +3.80%      
# syrk_omp               810.44 ms   +2.75%      N/A         96.62 J     +3.70%      
# syrk_cuda              1.48 s      +2.40%      N/A         163.01 J    +4.21%      
# syrk_seq_tuning        812.60 ms   +4.41%      N/A         96.79 J     +5.13%      
# trmm_numpy             879.60 ms   +1.05%      N/A         94.44 J     +3.39%      
# trmm_omp               812.54 ms   +8.19%      N/A         105.65 J    +9.64%      
+ trmm_seq_tuning        1.65 s      -10.49%     N/A         141.76 J    -1.89%