Bug: Vulkan backend fail to run basic test on adreno 690

[liangzelang]

What happened?

Compile llama.cpp with Vulkan backend, and push libggml.so/libllama.so to android device which gpu is qcom adreno 690.
But when excute the test got error like below:

./test-backend-ops perf -o GGML_ADD -b vulkan
ggml_vulkan: Found 1 Vulkan devices:
Vulkan0: Adreno (TM) 690 (Qualcomm Technologies Inc. Adreno Vulkan Driver) | uma: 1 | fp16: 1 | warp size: 64
libc++abi: terminating due to uncaught exception of type vk::UnknownError: vk::Device::createComputePipeline: ErrorUnknown
Aborted

Name and Version

./llama-cli --version
version: 3732 (8db003a)
built with Android (12285214, +pgo, +bolt, +lto, +mlgo, based on r522817b) clang version 18.0.2 (https://android.googlesource.com/toolchain/llvm-project d8003a456d14a3deb8054cdaa529ffbf02d9b262) for x86_64-unknown-linux-gnu

What operating system are you seeing the problem on?

No response

Relevant log output

./test-backend-ops perf -o GGML_ADD -b vulkan
ggml_vulkan: Found 1 Vulkan devices:
Vulkan0: Adreno (TM) 690 (Qualcomm Technologies Inc. Adreno Vulkan Driver) | uma: 1 | fp16: 1 | warp size: 64
libc++abi: terminating due to uncaught exception of type vk::UnknownError: vk::Device::createComputePipeline: ErrorUnknown
Aborted

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.