Regression/unoptimized_stream.cpp sporadically fails on AMD GPU

[bader]

Describe the bug
I noticed unexpected failure in the pre-commit validation for the documentation change - #5452. It's clear that documentation change can't cause the failure of this test.

FAIL: SYCL :: Regression/unoptimized_stream.cpp (644 of 738)
******************** TEST 'SYCL :: Regression/unoptimized_stream.cpp' FAILED ********************
Script:
--
: 'RUN: at line 1';    /__w/llvm/llvm/toolchain/bin/clang++     -Xsycl-target-backend=amdgcn-amd-amdhsa --offload-arch=gfx1031 -fsycl -fsycl-targets=amdgcn-amd-amdhsa /__w/llvm/llvm/llvm_test_suite/SYCL/Regression/unoptimized_stream.cpp -O0 -o /__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out
: 'RUN: at line 2';   env SYCL_DEVICE_FILTER=host  /__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out
: 'RUN: at line 3';   true /__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out
: 'RUN: at line 4';    env SYCL_DEVICE_FILTER=ext_oneapi_hip:gpu,host  /__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out
--
Exit Code: -11

Command Output (stdout):
--
$ ":" "RUN: at line 1"
note: command had no output on stdout or stderr
$ "/__w/llvm/llvm/toolchain/bin/clang++" "-Xsycl-target-backend=amdgcn-amd-amdhsa" "--offload-arch=gfx1031" "-fsycl" "-fsycl-targets=amdgcn-amd-amdhsa" "/__w/llvm/llvm/llvm_test_suite/SYCL/Regression/unoptimized_stream.cpp" "-O0" "-o" "/__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out"
# command stderr:
warning: linking module '/__w/llvm/llvm/toolchain/lib/clang/14.0.0/../../clc/remangled-l64-signed_char.libspirv-amdgcn--amdhsa.bc': Linking two modules of different target triples: '/__w/llvm/llvm/toolchain/lib/clang/14.0.0/../../clc/remangled-l64-signed_char.libspirv-amdgcn--amdhsa.bc' is 'amdgcn-unknown-amdhsa' whereas '/__w/llvm/llvm/llvm_test_suite/SYCL/Regression/unoptimized_stream.cpp' is 'amdgcn-amd-amdhsa'
 [-Wlinker-warnings]
1 warning generated.

$ ":" "RUN: at line 2"
note: command had no output on stdout or stderr
$ "env" "SYCL_DEVICE_FILTER=host" "/__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out"
note: command had no output on stdout or stderr
$ ":" "RUN: at line 3"
note: command had no output on stdout or stderr
$ "true" "/__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out"
note: command had no output on stdout or stderr
$ ":" "RUN: at line 4"
note: command had no output on stdout or stderr
$ "env" "SYCL_DEVICE_FILTER=ext_oneapi_hip:gpu,host" "/__w/llvm/llvm/build/SYCL/Regression/Output/unoptimized_stream.cpp.tmp.out"
note: command had no output on stdout or stderr
error: command failed with exit status: -11

--

********************

To Reproduce
That issue might be difficult to reproduce as it seems to be sporadic. I've seen this issue for the first time.
Probably the best way to reproduce it is to run the whole llvm-test-suite in a loop until it fails. :)
I'm leaving the record here hoping that someone someday will try to investigate this.

Environment (please complete the following information):

• OS: Linux

• Target device and vendor: AMD GPU

• DPC++ version: the compiler from [SYCL][DOC] Move discard queue events to supported #5452, but I suppose latest nightly should work as well.

• Dependencies version:

  Version  : HIP 40421.40
  Name     : AMD HIP BACKEND
  Vendor   : AMD Corporation
  Devices  : 1
      Device [#0]:
      Type       : gpu
      Version    : 0.0
      Name       : 
      Vendor     : AMD Corporation
      Driver     : HIP 40421.40