Various tests getting PI_ERROR_OUT_OF_RESOURCES on HIP

### Describe the bug

```
FAIL: SYCL :: Basic/stream/stream.cpp (330 of 2085)
******************** TEST 'SYCL :: Basic/stream/stream.cpp' FAILED ********************
Exit Code: -6
Command Output (stdout):
--
# 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/sycl/test-e2e/Basic/stream/stream.cpp -fsycl-device-code-split=per_kernel -o /__w/llvm/llvm/build-e2e/Basic/stream/Output/stream.cpp.tmp.out
# executed command: /__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/sycl/test-e2e/Basic/stream/stream.cpp -fsycl-device-code-split=per_kernel -o /__w/llvm/llvm/build-e2e/Basic/stream/Output/stream.cpp.tmp.out
# note: command had no output on stdout or stderr
# RUN: at line 2
env ONEAPI_DEVICE_SELECTOR=hip:gpu /__w/llvm/llvm/build-e2e/Basic/stream/Output/stream.cpp.tmp.out | /__w/llvm/llvm/toolchain/bin/FileCheck /__w/llvm/llvm/llvm/sycl/test-e2e/Basic/stream/stream.cpp
# executed command: env ONEAPI_DEVICE_SELECTOR=hip:gpu /__w/llvm/llvm/build-e2e/Basic/stream/Output/stream.cpp.tmp.out
# .---command stderr------------
# | terminate called after throwing an instance of 'sycl::_V1::runtime_error'
# | what(): Native API failed. Native API returns: -5 (PI_ERROR_OUT_OF_RESOURCES) -5 (PI_ERROR_OUT_OF_RESOURCES)
# `-----------------------------
# error: command failed with exit status: -6
# executed command: /__w/llvm/llvm/toolchain/bin/FileCheck /__w/llvm/llvm/llvm/sycl/test-e2e/Basic/stream/stream.cpp
# note: command had no output on stdout or stderr
--
```

```
AIL: SYCL :: Regression/commandlist/gpu.cpp (1728 of 2085)
******************** TEST 'SYCL :: Regression/commandlist/gpu.cpp' FAILED ********************
Exit Code: -6

Command Output (stdout):
--
# RUN: at line 5
/__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/sycl/test-e2e/Regression/commandlist/Inputs/FindPrimesSYCL.cpp /__w/llvm/llvm/llvm/sycl/test-e2e/Regression/commandlist/Inputs/main.cpp -o /__w/llvm/llvm/build-e2e/Regression/commandlist/Output/gpu.cpp.tmp.out -lpthread
# executed command: /__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/sycl/test-e2e/Regression/commandlist/Inputs/FindPrimesSYCL.cpp /__w/llvm/llvm/llvm/sycl/test-e2e/Regression/commandlist/Inputs/main.cpp -o /__w/llvm/llvm/build-e2e/Regression/commandlist/Output/gpu.cpp.tmp.out -lpthread

...

.---command stderr------------
# | terminate called after throwing an instance of 'sycl::_V1::runtime_error'
# | what(): Native API failed. Native API returns: -5 (PI_ERROR_OUT_OF_RESOURCES) -5 (PI_ERROR_OUT_OF_RESOURCES)
# `-----------------------------
# error: command failed with exit status: -6

```

https://github.com/intel/llvm/actions/runs/9670250347/job/26678954766?pr=14259

https://github.com/intel/llvm/actions/runs/9670250347/job/26710866376?pr=14259

### To reproduce

_No response_

### Environment

_No response_

### Additional context

_No response_

Start with llvm/sycl/test-e2e/Basic/stream/stream.cpp and llvm/sycl/test-e2e/Regression/commandlist/gpu.cpp, then review the linked CI jobs for the HIP configuration and failure context. Re-run these tests with the gfx1031 target if the environment is available; done means the resource errors are understood and the affected tests pass reliably.

Written by the indexing model from the issue text.