mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-05-01 22:54:05 +00:00
d28961d81e
* llama : enable chunked fused GDN path
* models : avoid Q and K repeats when using fused GDA
* cont : fix comment
Co-authored-by: Aman Gupta <amangupta052@gmail.com>
* cont : fix the fix
Co-authored-by: Aman Gupta <amangupta052@gmail.com>
* cont : fix
* metal : add GDN kernel (#20361)
* metal : add Metal backend for GGML_OP_GATED_DELTA_NET
Add a fused Metal kernel for the gated delta net recurrence op
(#19504), enabling GPU-accelerated inference for DeltaNet-based
models (Qwen3.5, etc.) on Apple Silicon.
Supports both GDA (scalar gate) and KDA (per-row gate) modes
with head_size 64 and 128. Unsupported configurations (head_size
32, non-contiguous tensors) gracefully fall back to CPU.
Performance: Qwen3.5-0.8B Q4_K_M on M4 Max
tg128: 170 -> 213 t/s (+25%)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* metal : validate contiguity of all input tensors in supports_op
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* metal : add algorithm equivalence comment for GDA decay path
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* cont : unslop + optimize
* cont : clean-up
---------
Co-authored-by: Paul Flynn <paul@arkavo.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* CUDA: AR gated delta net improvements (#20391)
* Add FastDiv to gated_delta_net_cuda
* Shard columns across warps
This reduces register pressure (avoids spill for S_v = 128) and gives
the warp-scheduler more CTAs to schedule (thus hiding data-access
latencies).
* Remove unneded include in gated_delta_net.cu
* Improve comments
* Apply code-formating
* Make sharding HIP-compatible
1. Use ggml_cuda_get_physical_warp_size() to determine warp size flexibly
2. Add test with partial warp to test sum reduction on CUDA
* Remove fastdiv_s64, as we can treat neqk1 and rq3 as uint32_t
* Rename variables
* Enable GDN also for prefill, move TODO for chunked_GDN
* Actually remove the TODO from 2068908975
* Get warp size at runtime
warp_size is not known at compile time in hip host code.
* Don't expose ggml_cuda_get_physical_warp_size on host
---------
Co-authored-by: uvos <devnull@uvos.xyz>
* llama : refactor llm_build_delta_net_base API
---------
Co-authored-by: Aman Gupta <amangupta052@gmail.com>
Co-authored-by: Paul Flynn <paul@arkavo.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Oliver Simons <osimons@nvidia.com>
Co-authored-by: uvos <devnull@uvos.xyz>
76 lines
2.0 KiB
C++
76 lines
2.0 KiB
C++
#pragma once
|
|
|
|
#include "ggml.h" // for ggml_log_level
|
|
|
|
#include <string>
|
|
#include <vector>
|
|
|
|
#ifdef __GNUC__
|
|
# if defined(__MINGW32__) && !defined(__clang__)
|
|
# define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
|
|
# else
|
|
# define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
|
|
# endif
|
|
#else
|
|
# define LLAMA_ATTRIBUTE_FORMAT(...)
|
|
#endif
|
|
|
|
//
|
|
// logging
|
|
//
|
|
|
|
LLAMA_ATTRIBUTE_FORMAT(2, 3)
|
|
void llama_log_internal (ggml_log_level level, const char * format, ...);
|
|
void llama_log_callback_default(ggml_log_level level, const char * text, void * user_data);
|
|
|
|
#define LLAMA_LOG(...) llama_log_internal(GGML_LOG_LEVEL_NONE , __VA_ARGS__)
|
|
#define LLAMA_LOG_INFO(...) llama_log_internal(GGML_LOG_LEVEL_INFO , __VA_ARGS__)
|
|
#define LLAMA_LOG_WARN(...) llama_log_internal(GGML_LOG_LEVEL_WARN , __VA_ARGS__)
|
|
#define LLAMA_LOG_ERROR(...) llama_log_internal(GGML_LOG_LEVEL_ERROR, __VA_ARGS__)
|
|
#define LLAMA_LOG_DEBUG(...) llama_log_internal(GGML_LOG_LEVEL_DEBUG, __VA_ARGS__)
|
|
#define LLAMA_LOG_CONT(...) llama_log_internal(GGML_LOG_LEVEL_CONT , __VA_ARGS__)
|
|
|
|
//
|
|
// helpers
|
|
//
|
|
|
|
template <typename T>
|
|
struct no_init {
|
|
T value;
|
|
no_init() = default;
|
|
};
|
|
|
|
struct time_meas {
|
|
time_meas(int64_t & t_acc, bool disable = false);
|
|
~time_meas();
|
|
|
|
const int64_t t_start_us;
|
|
|
|
int64_t & t_acc;
|
|
};
|
|
|
|
template <typename T>
|
|
struct buffer_view {
|
|
T * data;
|
|
size_t size = 0;
|
|
|
|
bool has_data() const {
|
|
return data && size > 0;
|
|
}
|
|
};
|
|
|
|
void replace_all(std::string & s, const std::string & search, const std::string & replace);
|
|
|
|
// TODO: rename to llama_format ?
|
|
LLAMA_ATTRIBUTE_FORMAT(1, 2)
|
|
std::string format(const char * fmt, ...);
|
|
|
|
std::string llama_format_tensor_shape(const std::vector<int64_t> & ne);
|
|
std::string llama_format_tensor_shape(const struct ggml_tensor * t);
|
|
|
|
std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i);
|
|
|
|
#define LLAMA_TENSOR_NAME_FATTN "__fattn__"
|
|
#define LLAMA_TENSOR_NAME_FGDN_AR "__fgdn_ar__"
|
|
#define LLAMA_TENSOR_NAME_FGDN_CH "__fgdn_ch__"
|