tree 6d4e660fffff28a6dead3f33169e529b354a0e40
parent 242e524fbff4f50ea6f4af4b99dfbf70b8948079
author Zhewen Yu <zhewenyu@amd.com> 1761125064 +0100
committer GitHub <noreply@github.com> 1761125064 +0100
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJo+KLICRC1aQ7uu5UhlAAAOFwQAGvLz/0+0M+IMF26LGuzNAlt
 rODalKQYcbukyd/FP2lNTyoFHo7FZKsLsRv5hX/sMeWlqyq/7BI+3VJcTKXtHVVS
 3QGgk/zivpUBlCHwzcd2i/YglGgyBLAhOgSznUToNbXeVRFEBomyif7Qt5VQPjjA
 NZG2QY/09WX361o48O9crejkyXdJFBbwLeSUhLvXwp3YgKxfowM0zQ63DavhTsuS
 i5ltpGMerx9wUd93f2KrF5UvtQjhGcdHGTm6VHwdsYWdfNg7e0rv9F/cXR5QcNNX
 +/Q40UD+yxw/wfDisylztOg1p0+tDYW+OI0m0Wffed8ggpQHz5zJdJzxzgY0lmU8
 A3Umqbo7ex13RvipL0wIPH0veyq8E+lMWt1JqEC/IF2Vfn7nQfnQ14LAUf0FRJfO
 o0JEaoNbbpjNB1LSWAyNI0dFYfKvpzvdWlTbx/GVx0JwqvIDYadG8KgFVWj79uib
 /ZykwpTfExoV2Wz1AnmWQCuT6M0laOGMyZWZ0RRR0YSiWq0VaeBKm5EPKXdzD2dk
 oTgST2wSknGliDwUT4A4TqN8RNfD1ZjgTxkEHITtQheepBtlCe+fYiTanfHYHIdY
 /7q1z/gwap9WPelpxXoBBKgbiHke/0mKFEzj7rJqWubypbyq2S0QFOWOM7uLcVAH
 cu8/B+Qy2ZH2Kc6b3XQo
 =McWi
 -----END PGP SIGNATURE-----
 

[LLVMCPU] Refactor multi lowering config propagation and setting (#22126)

The previous propagation logic was fragile because it assumed all ops in
a dispatch had an identity mapping between their parallel dimensions. It
also relied on indexing with `maxLoopNums`, which breaks for cases like
mmt4d.

This change replaces that approach with `IterationDimTracker +
totalLoopNum` for more robust propagation.

Additionally, a new helper class `MultiLoweringConfigGenerator` is
introduced to streamline the configuration process.

Closes: #21696

---------

Signed-off-by: Yu-Zhewen <zhewenyu@amd.com>