llama: update vendored code to commit 46e3556 (#8308)

llama/ggml-cuda/template-instances/mmq-instance-q3_k.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q4_0.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q4_1.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q4_k.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q5_0.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q5_1.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q5_k.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q6_k.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/template-instances/mmq-instance-q8_0.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/tsembd.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/tsembd.cuh

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/unary.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/unary.cuh

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/upscale.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/upscale.cuh

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/vecdotq.cuh

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *

llama/ggml-cuda/vendors/cuda.h

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *
@@ -29,6 +29,7 @@
 #include <cuda_runtime.h>
 #include <cuda.h>
 #include <cublas_v2.h>
+#include <cuda_bf16.h>
 #include <cuda_fp16.h>
 
 #if CUDART_VERSION < 11020

llama/ggml-cuda/vendors/hip.h

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *
@@ -29,6 +29,7 @@
 #include <hip/hip_runtime.h>
 #include <hipblas/hipblas.h>
 #include <hip/hip_fp16.h>
+#include <hip/hip_bfloat16.h>
 #ifdef __HIP_PLATFORM_AMD__
 // for rocblas_initialize()
 #include "rocblas/rocblas.h"
@@ -147,6 +148,8 @@
     #define __has_builtin(x) 0
 #endif
 
+typedef hip_bfloat16 nv_bfloat16;
+
 typedef int8_t int8x4_t __attribute__((ext_vector_type(4)));
 typedef uint8_t uint8x4_t __attribute__((ext_vector_type(4)));
 static __device__ __forceinline__ int __vsubss4(const int a, const int b) {

llama/ggml-cuda/vendors/musa.h

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *
@@ -29,6 +29,7 @@
 #include <musa_runtime.h>
 #include <musa.h>
 #include <mublas.h>
+#include <musa_bf16.h>
 #include <musa_fp16.h>
 #define CUBLAS_COMPUTE_16F CUDA_R_16F
 #define CUBLAS_COMPUTE_32F CUDA_R_32F
@@ -158,3 +159,5 @@
 #define cudaKernelNodeParams musaKernelNodeParams
 #define cudaStreamCaptureModeRelaxed musaStreamCaptureModeRelaxed
 #define cudaStreamEndCapture musaStreamEndCapture
+
+typedef mt_bfloat16 nv_bfloat16;

llama/ggml-cuda/wkv6.cu

 /**
- * llama.cpp - commit ba1cb19cdd0d92e012e0f6e009e0620f854b6afd - do not edit this file
+ * llama.cpp - commit 46e3556e01b824e52395fb050b29804b6cff2a7c - do not edit this file
  *
  * MIT License
  *