vulkan: Catch pipeline creation failure and print an error message (llama/11436)

* vulkan: Catch pipeline creation failure and print an error message

Also, fix some warnings from my on-demand compile change.

* vulkan: fix pipeline creation logging

ggml/src/ggml-vulkan/ggml-vulkan.cpp

@@ -774,12 +774,12 @@ static uint32_t compile_count = 0;
 static std::mutex compile_count_mutex;
 static std::condition_variable compile_count_cond;
 
-static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipeline, const std::string name, size_t spv_size, const void* spv_data, const std::string entrypoint,
-                                         uint32_t parameter_count, uint32_t push_constant_size, std::array<uint32_t, 3> wg_denoms, std::vector<uint32_t> specialization_constants,
-                                         uint32_t align, bool disable_robustness, bool require_full_subgroups, uint32_t required_subgroup_size) {
-    VK_LOG_DEBUG("ggml_vk_create_pipeline(" << device->name << ", " << name << ", " << entrypoint << ", " << parameter_count << ", " << push_constant_size <<
-                 ", (" << wg_denoms[0] << "," << wg_denoms[1] << "," << wg_denoms[2] << "), specialization_constants, " << align <<
-                 ", " << disable_robustness << ", " << require_full_subgroups << ", " << required_subgroup_size << ")");
+static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipeline, size_t spv_size, const void* spv_data, const std::string entrypoint,
+                                         uint32_t parameter_count, std::array<uint32_t, 3> wg_denoms, std::vector<uint32_t> specialization_constants,
+                                         bool disable_robustness, bool require_full_subgroups, uint32_t required_subgroup_size) {
+    VK_LOG_DEBUG("ggml_vk_create_pipeline(" << device->name << ", " << pipeline->name << ", " << entrypoint << ", " << parameter_count <<
+                 ", (" << wg_denoms[0] << "," << wg_denoms[1] << "," << wg_denoms[2] << "), specialization_constants, " <<
+                 disable_robustness << ", " << require_full_subgroups << ", " << required_subgroup_size << ")");
     GGML_ASSERT(parameter_count > 0);
     GGML_ASSERT(wg_denoms[0] > 0 && wg_denoms[1] > 0 && wg_denoms[2] > 0); // NOLINT
 
@@ -864,7 +864,13 @@ static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipelin
         compute_pipeline_create_info.setPNext(&rci);
     }
 
-    pipeline->pipeline = device->device.createComputePipeline(VK_NULL_HANDLE, compute_pipeline_create_info).value;
+    try {
+        pipeline->pipeline = device->device.createComputePipeline(VK_NULL_HANDLE, compute_pipeline_create_info).value;
+    } catch (const vk::SystemError& e) {
+        std::cerr << "ggml_vulkan: Compute pipeline creation failed for " << pipeline->name << std::endl;
+        std::cerr << "ggml_vulkan: " << e.what() << std::endl;
+        throw e;
+    }
     pipeline->compiled = true;
 
     {
@@ -1560,8 +1566,8 @@ static void ggml_vk_load_shaders(vk_device& device) {
             }
             compile_count++;
         }
-        compiles.push_back(std::async(ggml_vk_create_pipeline_func, std::ref(device), std::ref(pipeline), name, spv_size, spv_data, entrypoint,
-                                      parameter_count, push_constant_size, wg_denoms, specialization_constants, align, disable_robustness, require_full_subgroups, required_subgroup_size));
+        compiles.push_back(std::async(ggml_vk_create_pipeline_func, std::ref(device), std::ref(pipeline), spv_size, spv_data, entrypoint,
+                                      parameter_count, wg_denoms, specialization_constants, disable_robustness, require_full_subgroups, required_subgroup_size));
     };
 
 #if defined(VK_NV_cooperative_matrix2) && defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT)