vulkan: Add fusion support for RMS_NORM+MUL (ggml-org#14366)

* vulkan: Add fusion support for RMS_NORM+MUL

- Add a use_count to ggml_tensor, so we can detect if an output is used more than once.
- Change the ggml-vulkan rms_norm shader to optionally multiply by another tensor.
- Add detection logic and basic fusion logic in ggml-vulkan.
- Add some testing support for fusion. Rather than computing one node at a time, allow
for computing the whole graph and just testing one node's results. Add rms_norm_mul tests
and enable a llama test.

* extract some common fusion logic

* fix -Winconsistent-missing-override

* move ggml_can_fuse to a common function

* build fix

* C and C++ versions of can_fuse

* move use count to the graph to avoid data races and double increments when used in multiple threads

* use hash table lookup to find node index

* change use_counts to be indexed by hash table slot

* minimize hash lookups

style fixes

* last node doesn't need single use.
fix type.
handle mul operands being swapped.

* remove redundant parameter

---------

Co-authored-by: slaren <slarengh@gmail.com>

Author: 2 people

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

@@ -991,6 +991,10 @@ struct ggml_backend_vk_context {
 
     vk_command_pool compute_cmd_pool;
     vk_command_pool transfer_cmd_pool;
+
+    // number of additional consecutive nodes that are being fused with the
+    // node currently being processed
+    uint32_t num_additional_fused_ops {};
 };
 
 static void * const vk_ptr_base = (void *)(uintptr_t) 0x1000;  // NOLINT

tests/test-backend-ops.cpp

@@ -4896,6 +4896,8 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
 
     test_cases.emplace_back(new test_opt_step_adamw(GGML_TYPE_F32, {10, 5, 4, 3}));
 
+    test_cases.emplace_back(new test_llama(2, true));
+    // these tests are disabled to save execution time, but they can be handy for debugging
 #if 0
     // these tests are disabled to save execution time, sbut they can be handy for debugging
     test_cases.emplace_back(new test_llama(2, true));