ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
  Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32
| model                          |       size |     params | backend    | ngl | n_ubatch | fa | mmap |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | ---: | --------------: | -------------------: |
| qwen3moe 235B.A22B Q3_K - Medium |  96.99 GiB |   235.09 B | ROCm       |  99 |     2048 |  1 |    0 | pp2048 @ d16384 |         73.22 ± 0.00 |
/opt/llama.cpp/ggml/src/ggml-cuda/ggml-cuda.cu:94: ROCm error
/usr/local/lib64/libggml-base.so.0(+0x35a5) [0x7f12702305a5]
/usr/local/lib64/libggml-base.so.0(ggml_print_backtrace+0x1eb) [0x7f127023096b]
/usr/local/lib64/libggml-base.so.0(ggml_abort+0x11f) [0x7f1270230aef]
/usr/local/lib64/libggml-hip.so.0(+0x2cb1972) [0x7f1272f9e972]
/usr/local/lib64/libggml-hip.so.0(+0x2cb6b0e) [0x7f1272fa3b0e]
/usr/local/lib64/libggml-base.so.0(ggml_backend_sched_synchronize+0x2e) [0x7f1270247e5e]
/usr/local/lib64/libllama.so.0(_ZN13llama_context11synchronizeEv+0x10) [0x7f127369aab0]
/usr/local/bin/llama-bench() [0x408c12]
/lib64/libc.so.6(+0x35b5) [0x7f126fbc65b5]
/lib64/libc.so.6(__libc_start_main+0x88) [0x7f126fbc6668]
/usr/local/bin/llama-bench() [0x409c25]
✖ ! [rocm-7alpha-rocwmma] Qwen3-235B-A22B-Instruct-2507-UD-Q3_K_XL-00001-of-00003__fa1 __longctx16384 failed (exit 0)