ggml-hexagon: sync from project kantv(fix a long-term issue which introduced in kantv-ai/kantv#281)

Author: zhouwg

CMakeLists.txt

@@ -11,8 +11,7 @@ if(CMAKE_SYSTEM_NAME STREQUAL "Android")
     if(DEFINED HTP_ARCH_VERSION)
         if (${HTP_ARCH_VERSION} STREQUAL "v75" OR ${HTP_ARCH_VERSION} STREQUAL "v79")
             #works fine on Snapdragon 8Gen3&8Elite with 1.5x - 3x performance gains with the default ggml backend
-            #set(OPT_FLAG " -O3 -march=armv8.7-a -mcpu=cortex-x1 -mtune=cortex-x1 -flto -D_GNU_SOURCE -fvectorize -ffp-model=fast -fno-finite-math-only")
-            set(OPT_FLAG " -O3 -march=armv8.7-a -flto -D_GNU_SOURCE -fvectorize -ffp-model=fast -fno-finite-math-only")
+            set(OPT_FLAG " -O3 -march=armv8.7-a -mcpu=cortex-x1 -mtune=cortex-x1 -ffp-model=fast -fno-finite-math-only")
             message("OPT_FLAG:${OPT_FLAG}")
             set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS}   -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
             set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")

ggml/src/ggml-hexagon/ggml-hexagon.cpp

@@ -6134,13 +6134,18 @@ static void ggml_backend_hexagon_device_get_memory(ggml_backend_dev_t dev, size_
     } else if (HEXAGON_BACKEND_QNNNPU == ctx->device) {
         size_t rpc_ion_memsize = 0;
         size_t rpc_ion_usage   = 0;
-        if (HWACCEL_CDSP != g_hexagon_appcfg.hwaccel_approach) {
-            rpc_ion_memsize = ctx->instance->get_rpcmem_capacity();
-            rpc_ion_usage   = ctx->instance->get_rpcmem_usage();
-        } else {
-            rpc_ion_memsize = ctx->rpc_mempool_capacity;
-            rpc_ion_usage   = ctx->rpc_mempool_usage;
-        }
+        GGML_ASSERT(nullptr != ctx->instance);
+        rpc_ion_memsize = ctx->instance->get_rpcmem_capacity();
+        rpc_ion_usage   = ctx->instance->get_rpcmem_usage();
+        *total = rpc_ion_memsize;
+        *free = (rpc_ion_memsize - rpc_ion_usage);
+        GGMLHEXAGON_LOG_DEBUG("rpc memsize %d MiB", rpc_ion_memsize / SIZE_IN_MB);
+        GGMLHEXAGON_LOG_DEBUG("rpc usage %d MiB\n\n", rpc_ion_usage / SIZE_IN_MB);
+    } else if (HEXAGON_BACKEND_CDSP == ctx->device) {
+        size_t rpc_ion_memsize = 0;
+        size_t rpc_ion_usage   = 0;
+        rpc_ion_memsize = ctx->rpc_mempool_capacity;
+        rpc_ion_usage   = ctx->rpc_mempool_usage;
         *total = rpc_ion_memsize;
         *free = (rpc_ion_memsize - rpc_ion_usage);
         GGMLHEXAGON_LOG_DEBUG("rpc memsize %d MiB", rpc_ion_memsize / SIZE_IN_MB);