ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

musa: fix all warnings, re-enable -DLLAMA_FATAL_WARNINGS=ON in ci and update doc (#12611)

Browse source 
* musa: fix all warnings

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* musa: enable -DLLAMA_FATAL_WARNINGS=ON in run.sh

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* musa: update ci doc (install ccache)

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* fix Windows build issue

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* Address review comments

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* Address review comments

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

---------

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>
This commit is contained in:
R0CKSTAR 2025-03-30 16:59:38 +08:00 • committed by GitHub
parent d3f1f0acfb
commit 492d7f1ff7
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
20 changed files with 191 additions and 77 deletions
Download patch file Download diff file Expand all files Collapse all files

18
ggml/src/ggml-common.h
View file

@ -158,6 +158,12 @@ typedef sycl::half2 ggml_half2;
#endif // GGML_COMMON_DECL_CUDA || GGML_COMMON_DECL_HIP
#ifdef _MSC_VER
#define GGML_EXTENSION
#else // _MSC_VER
#define GGML_EXTENSION __extension__
#endif // _MSC_VER
#define QK4_0 32
typedef struct {
ggml_half d; // delta
@ -167,7 +173,7 @@ static_assert(sizeof(block_q4_0) == sizeof(ggml_half) + QK4_0 / 2, "wrong q4_0 b
#define QK4_1 32
typedef struct {
union {
GGML_EXTENSION union {
struct {
ggml_half d; // delta
ggml_half m; // min
@ -188,7 +194,7 @@ static_assert(sizeof(block_q5_0) == sizeof(ggml_half) + sizeof(uint32_t) + QK5_0
#define QK5_1 32
typedef struct {
union {
GGML_EXTENSION union {
struct {
ggml_half d; // delta
ggml_half m; // min
@ -209,7 +215,7 @@ static_assert(sizeof(block_q8_0) == sizeof(ggml_half) + QK8_0, "wrong q8_0 block
#define QK8_1 32
typedef struct {
union {
GGML_EXTENSION union {
struct {
ggml_half d; // delta
ggml_half s; // d * sum(qs[i])
@ -250,7 +256,7 @@ static_assert(sizeof(block_tq2_0) == sizeof(ggml_half) + QK_K / 4, "wrong tq2_0
typedef struct {
uint8_t scales[QK_K/16]; // scales and mins, quantized with 4 bits
uint8_t qs[QK_K/4]; // quants
union {
GGML_EXTENSION union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins
@ -277,7 +283,7 @@ static_assert(sizeof(block_q3_K) == sizeof(ggml_half) + QK_K / 4 + QK_K / 8 + 12
// weight is represented as x = a * q + b
// Effectively 4.5 bits per weight
typedef struct {
union {
GGML_EXTENSION union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins
@ -294,7 +300,7 @@ static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_half) + K_SCALE_SIZE + QK_K/2,
// weight is represented as x = a * q + b
// Effectively 5.5 bits per weight
typedef struct {
union {
GGML_EXTENSION union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins