[Example] Add More Hopper Matmul Examples

[WilliamZhang20]

Added v4/v5 of hopper_matmul, with the following results on H200 NVL:

Benchmark results (m=8192, n=8192, k=8192):
 version  latency (ms)     tflops  % of cublas
tilus_v0      3.842016 286.180909    41.596911
tilus_v1      2.169152 506.885464    73.676716
tilus_v2      2.113136 520.322213    75.629772
tilus_v3      1.922096 572.037845    83.146732
tilus_v4      1.953824 562.748557    81.796517
tilus_v5      1.740832 631.601216    91.804375
  cublas      1.605648 684.777511   100.000000

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[yaoyaoding]

/ok to test c07413f