Fix for GPU CI (#559)

Use level-zero runtime for test

Co-authored-by: [email protected] <Nishant Patel>

Author: nbpatel

test/PlaidML/OpTest.ExplicitPaddingNegInf.mlir

@@ -5,7 +5,7 @@
 // RUN: %gpu_skip || %python_executable %imex_runner -i %s --pass-pipeline-file=%p/linalg-to-llvm.pp \
 // RUN:                                       --runner mlir-cpu-runner -e main \
 // RUN:                                       --entry-point-result=void \
-// RUN:                                       --shared-libs=%mlir_runner_utils,%sycl_runtime | FileCheck %s
+// RUN:                                       --shared-libs=%mlir_runner_utils,%levelzero_runtime | FileCheck %s
 #map0 = affine_map<(d0, d1) -> (d0, d1)>
 #map1 = affine_map<(d0, d1) -> (d0 + 2, d1 + 1)>
 module @explicit_padding {