ekalda commented on PR #16862:
URL: https://github.com/apache/tvm/pull/16862#issuecomment-2057241247
Thanks @lhutton1 :)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To
ekalda merged PR #16862:
URL: https://github.com/apache/tvm/pull/16862
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or
lhutton1 commented on code in PR #16862:
URL: https://github.com/apache/tvm/pull/16862#discussion_r1562343236
##
tests/python/arith/test_arith_simplify.py:
##
@@ -53,6 +56,33 @@ def test_simplify_symbolic_comparison():
assert ana.can_prove((n + 31) // 32 * 32 >= i0 * 32 + i
lhutton1 commented on code in PR #16862:
URL: https://github.com/apache/tvm/pull/16862#discussion_r1562342420
##
src/arith/scalable_expression.cc:
##
@@ -50,5 +66,20 @@ std::optional ExtractVscaleFactor(const PrimExpr&
lanes) {
}
}
+bool CanProveVscaleExpressionFromKnownV
ekalda commented on code in PR #16862:
URL: https://github.com/apache/tvm/pull/16862#discussion_r1560717574
##
tests/python/arith/test_arith_simplify.py:
##
@@ -53,6 +56,33 @@ def test_simplify_symbolic_comparison():
assert ana.can_prove((n + 31) // 32 * 32 >= i0 * 32 + i1,
lhutton1 commented on PR #16862:
URL: https://github.com/apache/tvm/pull/16862#issuecomment-2047480696
cc @ekalda @Lunderberg @cbalint13 @Anndrey24 @tqchen
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above
lhutton1 opened a new pull request, #16862:
URL: https://github.com/apache/tvm/pull/16862
This commit adds support for splitting via the compile-time unknown constant
`vscale`. Two main changes are introduced; they are described below.
The split scheduling primitive has a new paramete
