[Issue]: hipMemSetGetAccess failing in hip-tests/catch/unit/virtualMemoryManagement

[tjtanaa]

Problem Description

Description

hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess test is failing.

The ROCm and HIP version is based on this docker image rocm/dev-ubuntu-22.04:6.3.1-complete.

root:/app/hip-tests/build/catch_tests/unit/virtualMemoryManagement# ./VirtualMemoryManagementTest 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
VirtualMemoryManagementTest is a Catch v2.13.4 host application.
Run with -? for options

-------------------------------------------------------------------------------
Unit_hipMemCreate_MapNonContiguousChunks
-------------------------------------------------------------------------------
/app/hip-tests/catch/unit/virtualMemoryManagement/hipMemCreate.cc:272
...............................................................................

/app/hip-tests/catch/unit/virtualMemoryManagement/hipMemCreate.cc:323: FAILED:
  REQUIRE( true == std::equal(B_h.begin(), B_h.end(), C_h.data()) )
with expansion:
  true == false

Memory access fault by GPU node-9 (Agent handle: 0x2373990) on address 0x7f88a6dd5000. Reason: Unknown.
GPU core dump created: gpucore.27623
-------------------------------------------------------------------------------
Unit_hipMemSetAccess_Vmm2UnifiedMemCpy
-------------------------------------------------------------------------------
/app/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc:552
...............................................................................

/app/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc:552: FAILED:
  {Unknown expression after the reported line}
due to a fatal error condition:
  SIGABRT - Abort (abnormal termination) signal

===============================================================================
test cases:    24 |    22 passed | 2 failed
assertions: 73852 | 73850 passed | 2 failed

Aborted

Operating System

Ubuntu 22.04.4 LTS (Jammy Jellyfish)

CPU

AMD EPYC 9654 96-Core Processor

GPU

AMD Instinct MI300X

ROCm Version

ROCm 6.3.1

ROCm Component

HIP

Steps to Reproduce

Steps:

1. Launch Docker

#!/bin/bash
docker run -it \
   --network=host \
   --group-add=video \
   --ipc=host \
   --cap-add=SYS_PTRACE \
   --security-opt seccomp=unconfined \
   --device /dev/kfd \
   --device /dev/dri \
   rocm/dev-ubuntu-22.04:6.3.1-complete \
   bash

2. Compile ROCM/hip-tests

export ROCM_BRANCH=rocm-6.3.x
git clone -b "$ROCM_BRANCH" https://github.com/ROCm/hip-tests.git
export HIPTESTS_DIR="$(readlink -f hip-tests)"
cd "$HIPTESTS_DIR"
mkdir -p build; cd build
cmake ../catch -DHIP_PLATFORM=amd -DHIP_PATH=$CLR_DIR/build/install
make build_tests
cd catch_tests/unit/virtualMemoryManagement/
./VirtualMemoryManagementTest 

(Optional for Linux users) Output of /opt/rocm/bin/rocminfo --support

No response

Additional Information

No response

[ppanchad-amd]

Hi @tjtanaa. Internal ticket has been created to investigate this issue. Thanks!