fix: use lapack instead of lapacke

feat: get_dimension_size batch interface

feat: implement jitcall batching with shlo_generic_batch_op_interface

refactor: reuse batching interface for LU factorization

fix: remove old changes

feat: mark memory effects

fix: update to new API

fix: use correct return

fix: rework svd lowering to use lapack and workspace buffers

fix: use func return

fix: leading dimension

fix: use backend config

fix: remove unused vars

fix: tpu lowering

test: update

Author: avik-pal

src/enzyme_ad/jax/Dialect/EnzymeXLAOps.td

@@ -397,7 +397,7 @@ def CommRegionOp : EnzymeXLA_Op<"comm_region", [
 
 // Linear Algebra Ops
 
-def SymmOp: EnzymeXLA_Op<"lapack.symm", [Pure, SameOperandsAndResultElementType]> {
+def SymmOp: EnzymeXLA_Op<"blas.symm", [Pure, SameOperandsAndResultElementType]> {
   let summary = "Multiplication involving a symmetric matrix";
 
   let description = [{