Performance Update (2025.04.22) (#71)

* Fix benchmark script

* Performance optimization for compute-bound cases

* Add new testcase (s_k = 16384)

* Update README.md

* Update comment

* Update README.md

* Add the deep-dive blog

* Add background color for MLA Kernel Sched.drawio.svg

* Use relative path for the schedule image

* Move flash_mla.h to kernels/params.h

csrc/kernels/mla_combine.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include "params.h"
+
+template<typename ElementT>
+void run_flash_mla_combine_kernel(Flash_fwd_mla_params &params, cudaStream_t stream);