diff --git a/README.md b/README.md index 37e3f66..d6447f9 100644 --- a/README.md +++ b/README.md @@ -1,9 +1,18 @@ -# FASTEN: A Library of *Fast Segment* Operators +# Fasten: A Library of *Fast Segment* Operators ## Introduction -FASTEN is a library aimed at speeding up Heterogeneous Graph Neural Network (HGNN) workloads. -The current version of FASTEN focuses on improving segmented matrix multiplication, a critical operator in HGNNs. +Fasten is a library aimed at speeding up Heterogeneous Graph Neural Network (HGNN) workloads. +The current version of Fasten focuses on improving segmented matrix multiplication, a critical operator in HGNNs. Fasten implements a simple interface, making it easy to integrate with existing graph library PyG with minimal changes. +Fasten achieved an average speedup of 5.71$\times$ and 2.46$\times$ in operator-wise benchmarks compared to CUTLASS and cuBLAS, respectively + +### Fasten vs CUTLASS (no autotune) + +![figure9_fasten_vs_cutlass(pyg)](https://github.com/Deep-Learning-Profiling-Tools/fasten/assets/2306281/d88fab7c-a331-4978-9157-08e448afcce5) + +### Fasten vs cuBLAS + +![figure9_fasten_vs_cublas(torch)](https://github.com/Deep-Learning-Profiling-Tools/fasten/assets/2306281/4f8fbe5f-f8d4-45b2-9f92-ac3f7cb97c28) ## Installation @@ -24,7 +33,7 @@ pip install . ``` ## Examples -FASTEN's segment matrix multiplication operator has been integrated with various HGNN architecture such as *RGCN*, *HGT*, *RGAT* in PyG. +Fasten's segment matrix multiplication operator has been integrated with various HGNN architecture such as *RGCN*, *HGT*, *RGAT* in PyG. Examples on how to run the examples can be found below: ### GNN Examples