ggml/hip: fix APU compatibility - soft error handling for hipMemAdviseSetCoarseGrain...

* ggml/hip: fix APU compatibility - soft error handling for hipMemAdviseSetCoarseGrain

On AMD APU/iGPU devices (unified memory architecture), hipMemAdviseSetCoarseGrain
returns hipErrorInvalidValue because the hint is not applicable to UMA systems.
The previous CUDA_CHECK() call treated this as a fatal error, causing crashes on
APU systems such as AMD Strix Halo (gfx1151).

Fix: treat hipMemAdviseSetCoarseGrain as an optional performance hint - call it
without error checking and clear any resulting error with hipGetLastError().

Also add pre-allocation debug logging (GGML_LOG_DEBUG) to help diagnose memory
issues on APU systems, and store totalGlobalMem in device info.

Context: AMD APUs on Windows are affected by a ROCm runtime bug that limits
hipMallocManaged to ~64GB regardless of available system RAM. A fix has been
submitted upstream: https://github.com/ROCm/rocm-systems/pull/4077

Co-Authored-By: Claude Sonnet 4.6 <redacted>
* ggml/hip: remove unrelated changes, keep only hipMemAdviseSetCoarseGrain fix

---------

Co-authored-by: moonshadow-25 <redacted>
Co-authored-by: Claude Sonnet 4.6 <redacted>

diff --git a/ggml/src/ggml-cuda/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu
index ce7a80acde8ff9ac70bcf05fb4825a216df49075..3886290c5ff9c8fa9ee9a61f8006fb58f692062f 100644 (file)
--- a/ggml/src/ggml-cuda/ggml-cuda.cu
+++ b/ggml/src/ggml-cuda/ggml-cuda.cu
@@ -124,7 +124,10 @@ static cudaError_t ggml_cuda_device_malloc(void ** ptr, size_t size, int device)
         err = cudaMallocManaged(ptr, size);
 #if defined(GGML_USE_HIP)
         if (err == hipSuccess) {
-            CUDA_CHECK(cudaMemAdvise(*ptr, size, hipMemAdviseSetCoarseGrain, device));
+            // hipMemAdviseSetCoarseGrain is an optional performance hint;
+            // ignore errors (e.g. hipErrorInvalidValue on some APU/iGPU configs).
+            cudaMemAdvise(*ptr, size, hipMemAdviseSetCoarseGrain, device);
+            (void)hipGetLastError(); // clear any error
         }
 
         // fall back to cudaMalloc if not supported (e.g. on Windows)