Skip to content

Distribute and run LLMs with a single file.

vulkan: clearer message when --gpu vulkan is explicitly unavailable (#1006)

Davide Eynard9ff770c7172f
diff --git a/llamafile/vulkan.c b/llamafile/vulkan.c
index 55cd27bb..63f6ff05 100644
--- a/llamafile/vulkan.c
+++ b/llamafile/vulkan.c
@@ -100,8 +100,17 @@ static void ImportVulkan(void) {
int count = gpu_call_device_count(g_vulkan.get_device_count);
llamafile_info("vulkan", "found %d GPU device(s)", count);
} else if (FLAG_gpu == LLAMAFILE_GPU_VULKAN) {
- fprintf(stderr, "fatal error: support for --gpu vulkan was explicitly requested, "
- "but it wasn't available\n");
+ // Vulkan was explicitly requested but the probe found no usable device
+ // (or the driver faulted during probe and was skipped to keep the
+ // process alive -- see gpu_backend.c). Fail loudly rather than
+ // silently downgrade, but tell the user how to proceed.
+ fprintf(stderr,
+ "fatal error: --gpu vulkan was explicitly requested but Vulkan is not "
+ "usable on this system\n"
+ " - no Vulkan-capable device was detected, or the driver "
+ "failed/crashed during probe and was skipped for safety\n"
+ " - retry with --gpu auto to use the best available backend "
+ "(falls back to CPU), or --gpu disabled to force CPU\n");
exit(1);
}
}