This is an automated email from the ASF dual-hosted git repository.

tlopex pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm-tirx.git


The following commit(s) were added to refs/heads/main by this push:
     new 37e6ef030 [DOCS] Illustrate TIRx authoring layers with an HTML diagram 
(#196)
37e6ef030 is described below

commit 37e6ef030b87995be00520cc126fc0f5a6f65a99
Author: Shushi Hong <[email protected]>
AuthorDate: Wed Nov 4 20:30:13 2026 -0500

    [DOCS] Illustrate TIRx authoring layers with an HTML diagram (#196)
    
    This PR replaces the authoring-layer table and ASCII flowchart in
    Programming Guides with an embedded HTML/CSS diagram. It shows core
    language constructs throughout one TIRx program, tile expansion into
    core IR and backend calls, and a thin direct backend interface.
---
 docs/programming_guides/index.rst | 46 +++++++++++----------------------------
 1 file changed, 13 insertions(+), 33 deletions(-)

diff --git a/docs/programming_guides/index.rst 
b/docs/programming_guides/index.rst
index c2f050dd7..eb5dc9b29 100644
--- a/docs/programming_guides/index.rst
+++ b/docs/programming_guides/index.rst
@@ -75,39 +75,19 @@ TIRx programs use one TVMScript namespace alias:
 
    from tvm_tirx.script import tirx as T
 
-The namespace exposes three authoring layers. They share the same TIRx
-program, but enter the compiler at different levels:
-
-.. list-table::
-   :header-rows: 1
-   :widths: 25 30 45
-
-   * - Authoring layer
-     - Syntax
-     - Representation and lowering
-   * - Core language
-     - ``T.*``
-     - Creates TIRx statements, expressions, tensors, and control flow.
-   * - Backend operations
-     - ``T.cuda.*`` and ``T.ptx.*``
-     - Creates backend calls directly; these calls bypass tile operation
-       dispatch.
-   * - Tile operations
-     - ``T.cuda.tile.*``
-     - Creates a tile call that is replaced by a target-specific
-       implementation during tile operation dispatch.
-
-``tvm_tirx.script.tirx`` is the construction syntax for the TIRx IR, not a
-separate IR. Tile operation dispatch and layout and scope lowering replace
-high-level TIRx nodes with lower-level TIRx statements and calls:
-
-.. code-block:: text
-
-   T.* ───────────────────────────▶ TIRx statements and expressions ──┐
-   T.cuda.* / T.ptx.* ────────────▶ backend calls ────────────────────┤
-   T.cuda.tile.* ─▶ tile call ─▶ tile operation dispatch ─────────────┤
-                                                                      ▼
-                                        layout and scope lowering ─▶ codegen
+The core language supplies loops, branches, scalar expressions, and tensor
+accesses throughout a kernel. CUDA tile operations add an optional expansion
+layer over these constructs and native backend operations. All of them
+coexist in the same TIRx IR:
+
+.. raw:: html
+
+   <iframe data-tirx-authoring-layers
+           src="https://tvm.apache.org/images/docs/tirx/authoring_layers.html";
+           title="TIRx authoring layers: core language, tile operations, and 
backend calls"
+           style="display: block; width: 100%; height: 480px; border: 0;"
+           loading="lazy"></iframe>
+   <script defer 
src="https://tvm.apache.org/images/docs/tirx/authoring_layers_embed.js";></script>
 
 The model rests on three constructs:
 

Reply via email to