Add H100SXM benchmark

Author: tornikeo

README.md

@@ -49,11 +49,13 @@ For details, visit the [preprint](https://www.biorxiv.org/content/biorxiv/early/
 ## Hardware
 
 Any GPU [supported](https://numba.pydata.org/numba-doc/dev/cuda/overview.html#requirements) by Numba can be used. We tested a number of GPUs:
-- GTX 1050 Ti, used on a local laptop
-- RTX 2070, used on a local machine
-- T4 GPU, offered for free on Colab
-- RTX 4090 GPU, offered on vast.ai
-- Any A100 GPU, offered on vast.ai
+
+- GTX 1050 Ti
+- RTX 2070
+- T4 GPU (free on [Google Colab](https://colab.research.google.com/))
+- RTX 4090 (on vast.ai)
+- A100 80GB (on vast.ai)
+- H100SXM 80GB (on vast.ai)
 
 Performance (comparisons/s) is proportional to the peak GPU bandwidth (GB/s).