vulkan: Implement topk_moe fused shader, ported from CUDA (#16641)

This is similar to the CUDA shader from #16130, but doesn't use shared memory
and handles different subgroup sizes.

ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp

@@ -920,6 +920,8 @@ void process_shaders() {
 
     string_to_spv("ssm_conv_f32", "ssm_conv.comp", {{"A_TYPE", "float"}});
 
+    string_to_spv("topk_moe_f32", "topk_moe.comp", {});
+
     for (auto &c : compiles) {
         c.wait();
     }