From 58dfb5c6fcbcc75b4b311e2294b1f891bc56d397 Mon Sep 17 00:00:00 2001 From: Oleksandr Pavlyk Date: Mon, 13 May 2024 12:56:26 -0700 Subject: [PATCH] Fix for crash reported in gh-1654 This also fixes a crash in dpt.sort on CUDA device for sorting of 256 elements of floating point numbers. --- dpctl/tensor/libtensor/include/kernels/sorting/sort.hpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dpctl/tensor/libtensor/include/kernels/sorting/sort.hpp b/dpctl/tensor/libtensor/include/kernels/sorting/sort.hpp index df69f95b8d..9a86781faa 100644 --- a/dpctl/tensor/libtensor/include/kernels/sorting/sort.hpp +++ b/dpctl/tensor/libtensor/include/kernels/sorting/sort.hpp @@ -67,7 +67,7 @@ void merge_impl(const std::size_t offset, const std::size_t local_size_2 = local_end_2 - local_start_2; const auto r_item_1 = in_acc[end_1 - 1]; - const auto l_item_2 = in_acc[start_2]; + const auto l_item_2 = (start_2 < end_2) ? in_acc[start_2] : r_item_1; // Copy if the sequences are sorted with respect to each other or merge // otherwise