Add mirage, StructTensor, and SDQL

utensil · Oct 15, 2024 · 4ffbb24 · 4ffbb24
1 parent 999127f
commit 4ffbb24
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 4 deletions.
diff --git a/yard-rs/clifford-xp/README.md b/yard-rs/clifford-xp/README.md
@@ -63,6 +63,7 @@ For GPU backends, there are:
 - [mlx-rs](https://github.com/oxideai/mlx-rs)
 - [rai](https://github.com/cksac/rai) JAX-like
 - [krnl](https://github.com/charles-r-earp/krnl)
+- [luminal](https://github.com/jafioti/luminal) which uses composable compilers and uses egg
 - [cudarc](https://github.com/coreylowman/cudarc) for CUDA
 - [Ash](https://github.com/ash-rs/ash) for Vulkan
 - [gpu.cpp](https://github.com/AnswerDotAI/gpu.cpp) in C++
@@ -93,6 +94,10 @@ For symbolic backends, there are:
   - [GiNaC](https://www.ginac.de/)
   - [cadabra2](https://github.com/kpeeters/cadabra2)
 
+For program optimization, the approach of [mirage](https://github.com/mirage-project/mirage), [StructTensor](https://github.com/edin-dal/structtensor), and `STOREL`(*Optimizing Tensor Programs on Flexible Storage*) based on [SDQL](https://github.com/edin-dal/sdql) and egg should be considered.
+
+<!-- - [Enumo](https://github.com/uwplse/ruler), https://github.com/moves-rwth/caesar  which uses egg -->
+
 As for codegen, there are
 
 - [Gaalop](https://github.com/CallForSanity/Gaalop)

diff --git a/yard-rs/runpod-xp/justfile b/yard-rs/runpod-xp/justfile
@@ -14,12 +14,9 @@ kill:
 test-kill ID:
     RUNPOD_POD_ID={{ID}} python kill.py
 
-run CONFIG:
+run CONFIG="demo.yml":
     python run.py {{ justfile_directory() / "tasks" / CONFIG }}
 
-test-run:
-    just run demo.yml
-
 gpus:
     #!/usr/bin/env bash
     curl --request POST \