[Mlir-commits] [mlir] [MLIR][Linalg][Docs] Add forms to Linalg	rationale docs (PR #156859)
    Andrzej WarzyĆski 
    llvmlistbot at llvm.org
       
    Mon Sep 15 06:56:30 PDT 2025
    
    
  
================
@@ -506,6 +506,58 @@ potential by introducing lower-level IR ops and *smaller* Linalg ops.
 This gradually reduces the potential, all the way to Loops + VectorOps
 and LLVMIR.
 
+### Interchangeability of Forms<a name="forms"></a>
+
+#### The Linalg Forms
+
+The core Linalg operation tree has three forms:
+* **Generic:** Represented by `linalg.generic` and can encode all perfectly-nested
+loop operations.
+* **Category:** Represented by `linalg.contract` and `linalg.elementwise`,
+which are special (einsum) forms of the `generic` operation. In the future, other
+category operations are planned (e.g.: `linalg.convolution` and `linalg.pooling`).
+* **Named:** All _named_ forms that can lower to either _category_ or
----------------
banach-space wrote:
I was also wondering whether "lower" is the right term here. Given that `linalg.generic` is at the top of the Linalg tree, one could see as going from `linalg.matmul` to `linalg.generic` as lifting. Perhaps "translate" or "rewrite" or "convert"?
https://github.com/llvm/llvm-project/pull/156859
    
    
More information about the Mlir-commits
mailing list