vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 09:34:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.github/workflows/build.yml
#	.github/workflows/close-issue.yml
#	.github/workflows/nix-ci-aarch64.yml
#	.github/workflows/nix-ci.yml
#	README.md
#	ci/run.sh
#	examples/server/README.md
#	ggml/src/ggml-cuda.cu
#	ggml/src/ggml-metal.m
#	scripts/sync-ggml.last
#	tests/test-backend-ops.cpp
This commit is contained in:
Concedo 2024-10-05 22:24:08 +08:00
commit da6cf261a8
28 changed files with 725 additions and 421 deletions
Download patch file Download diff file Expand all files Collapse all files

102
ggml/src/ggml-backend.cpp
View file

@ -1,5 +1,13 @@
// Note: porting this file to C++ is a work in progress
#ifdef _WIN32
#define WIN32_LEAN_AND_MEAN
#ifndef NOMINMAX
# define NOMINMAX
#endif
#include <windows.h>
#endif
#include "ggml-backend-impl.h"
#include "ggml-alloc.h"
#include "ggml-impl.h"
@ -10,9 +18,15 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <string>
#include <vector>
#ifdef __APPLE__
#include <sys/types.h>
#include <sys/sysctl.h>
#endif
// backend buffer type
const char * ggml_backend_buft_name(ggml_backend_buffer_type_t buft) {
@ -505,12 +519,6 @@ void * ggml_backend_reg_get_proc_address(ggml_backend_reg_t reg, const char * na
return reg->iface.get_proc_address(reg, name);
}
void ggml_backend_reg_set_log_callback(ggml_backend_reg_t reg, ggml_log_callback log_callback, void * user_data) {
if (reg->iface.set_log_callback) {
reg->iface.set_log_callback(reg, log_callback, user_data);
}
}
// Backend registry
#ifdef GGML_USE_CUDA
@ -614,13 +622,6 @@ ggml_backend_dev_t ggml_backend_dev_by_type(enum ggml_backend_dev_type type) {
return NULL;
}
void ggml_backend_set_log_callback(ggml_log_callback log_callback, void * user_data) {
for (size_t i = 0; i < ggml_backend_reg_count(); i++) {
ggml_backend_reg_t reg = ggml_backend_reg_get(i);
ggml_backend_reg_set_log_callback(reg, log_callback, user_data);
}
}
// Convenience functions
ggml_backend_t ggml_backend_init_by_name(const char * name, const char * params) {
ggml_backend_dev_t dev = ggml_backend_dev_by_name(name);
@ -1021,6 +1022,70 @@ ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(void * ptr, size_t size)
////////////////////////
struct ggml_backend_cpu_device_context {
std::string description = "CPU";
ggml_backend_cpu_device_context() {
#ifdef __APPLE__
size_t len = 0;
if (!sysctlbyname("machdep.cpu.brand_string", NULL, &len, NULL, 0)) {
description.resize(len);
sysctlbyname("machdep.cpu.brand_string", &description[0], &len, NULL, 0); // NOLINT
}
#elif defined(__linux__)
FILE * f = fopen("/proc/cpuinfo", "r");
if (f) {
char buf[1024];
while (fgets(buf, sizeof(buf), f)) {
if (strncmp(buf, "model name", 10) == 0) {
char * p = strchr(buf, ':');
if (p) {
p++;
while (std::isspace(*p)) {
p++;
}
while (std::isspace(p[strlen(p) - 1])) {
p[strlen(p) - 1] = '\0';
}
description = p;
break;
}
}
}
fclose(f);
}
#elif defined(_WIN32)
HKEY hKey;
if (RegOpenKeyEx(HKEY_LOCAL_MACHINE,
TEXT("HARDWARE\\DESCRIPTION\\System\\CentralProcessor\\0"),
0,
KEY_READ,
&hKey) == ERROR_SUCCESS) {
DWORD cpu_brand_size = 0;
if (RegQueryValueExA(hKey,
TEXT("ProcessorNameString"),
NULL,
NULL,
NULL,
&cpu_brand_size) == ERROR_SUCCESS) {
description.resize(cpu_brand_size);
if (RegQueryValueExA(hKey,
TEXT("ProcessorNameString"),
NULL,
NULL,
(LPBYTE)&description[0], // NOLINT
&cpu_brand_size) == ERROR_SUCCESS) {
if (description.find('\0') != std::string::npos) {
description.resize(description.find('\0'));
}
}
}
RegCloseKey(hKey);
}
#endif
}
};
static const char * ggml_backend_cpu_device_get_name(ggml_backend_dev_t dev) {
return "CPU";
@ -1028,10 +1093,9 @@ static const char * ggml_backend_cpu_device_get_name(ggml_backend_dev_t dev) {
}
static const char * ggml_backend_cpu_device_get_description(ggml_backend_dev_t dev) {
// TODO
return "CPU";
struct ggml_backend_cpu_device_context * ctx = (struct ggml_backend_cpu_device_context *)dev->context;
GGML_UNUSED(dev);
return ctx->description.c_str();
}
static void ggml_backend_cpu_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) {
@ -1144,10 +1208,11 @@ static size_t ggml_backend_cpu_reg_get_device_count(ggml_backend_reg_t reg) {
static ggml_backend_dev_t ggml_backend_cpu_reg_get_device(ggml_backend_reg_t reg, size_t index) {
GGML_ASSERT(index == 0);
static ggml_backend_cpu_device_context ctx;
static ggml_backend_device ggml_backend_cpu_device = {
/* .iface = */ ggml_backend_cpu_device_i,
/* .reg = */ reg,
/* .context = */ NULL,
/* .context = */ &ctx,
};
return &ggml_backend_cpu_device;
@ -1161,7 +1226,6 @@ static const struct ggml_backend_reg_i ggml_backend_cpu_reg_i = {
/* .get_device_count = */ ggml_backend_cpu_reg_get_device_count,
/* .get_device = */ ggml_backend_cpu_reg_get_device,
/* .get_proc_address = */ NULL,
/* .set_log_callback = */ NULL,
};
ggml_backend_reg_t ggml_backend_cpu_reg(void) {