Faster topk implementation for large enough graphs

[mova]

Topk needs quite some time to construct the mask in my application.
I noticed I can avoid the loop, if all of the graphs in the batch are large than ratio.

[codecov]

Codecov Report

Merging #6123 (a45a193) into master (11c8cbd) will decrease coverage by 1.87%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #6123      +/-   ##
==========================================
- Coverage   86.21%   84.34%   -1.88%     
==========================================
  Files         363      363              
  Lines       20490    20488       -2     
==========================================
- Hits        17666    17280     -386     
- Misses       2824     3208     +384     

Impacted Files	Coverage Δ
torch_geometric/nn/pool/topk_pool.py	100.00% <100.00%> (ø)
torch_geometric/nn/models/dimenet_utils.py	0.00% <0.00%> (-75.52%)	⬇️
torch_geometric/nn/models/dimenet.py	14.90% <0.00%> (-52.76%)	⬇️
torch_geometric/profile/profile.py	36.27% <0.00%> (-26.48%)	⬇️
torch_geometric/nn/conv/utils/typing.py	81.25% <0.00%> (-17.50%)	⬇️
torch_geometric/nn/pool/asap.py	92.10% <0.00%> (-7.90%)	⬇️
torch_geometric/nn/inits.py	67.85% <0.00%> (-7.15%)	⬇️
torch_geometric/nn/dense/linear.py	87.40% <0.00%> (-5.93%)	⬇️
torch_geometric/transforms/add_self_loops.py	94.44% <0.00%> (-5.56%)	⬇️
torch_geometric/nn/models/attentive_fp.py	95.83% <0.00%> (-4.17%)	⬇️
... and 13 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[rusty1s]

This is great, thank you!