CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-15 11:17:31 +00:00
Code Issues Packages Projects Releases Wiki Activity

zdnn: refactor codebase + add docs (#16178)

Browse Source 
* zdnn: initial matmul refactor

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: rm static from funcs

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: update ggml-zdnn.h

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: change header files to hpp

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: switch to common.hpp

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: move mulmat forward around

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: rm inline from utils

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* ggml-zdnn: code cleanup

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

* docs: add zDNN docs

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>

---------

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-09-23 14:53:05 +08:00
committed by GitHub
parent 0bc7cc7154
commit 264f1b5187
11 changed files with 337 additions and 269 deletions
Download Patch File Download Diff File Expand all files Collapse all files

59
ggml/src/ggml-zdnn/common.hpp Normal file
View File

@@ -0,0 +1,59 @@
#ifndef GGML_ZDNN_COMMON_HPP
#define GGML_ZDNN_COMMON_HPP
#include "ggml.h"
#include "ggml-impl.h"
#include "zdnn.h"
#include <vector>
#include <memory>
#define GGML_ZDNN_NAME "zDNN"
#define GGML_ZDNN_VERSION ZDNN_VERNUM
#define ZDNN_CHECK(stmt) \
do { \
zdnn_status status = (stmt); \
GGML_ASSERT(status == ZDNN_OK); \
} while (0);
struct ggml_backend_zdnn_device_context {
int zdnn_device;
int zdnn_device_ref_count;
bool has_parmblkformat_0;
bool has_parmblkformat_1; // checks for z17
size_t max_size;
char name[128];
};
struct ggml_backend_zdnn_context {
int device;
ggml_cgraph * gf;
};
struct ggml_backend_zdnn_buffer {
void * data;
ggml_backend_zdnn_buffer * extra; // for bias, etc.
size_t size;
zdnn_tensor_desc pre_tfm_desc;
zdnn_tensor_desc tfm_desc;
zdnn_ztensor ztensor;
char name[GGML_MAX_NAME];
};
struct ggml_backend_zdnn_buffer_context {
void * all_data;
size_t all_size;
bool owned;
int n_buffers;
std::vector<std::unique_ptr<ggml_backend_zdnn_buffer>> buffers;
};
#endif // GGML_ZDNN_COMMON_HPP