attention_oracle_ggml.h File Reference

Go to the source code of this file.

Macros
#define	CK_ENABLE_LLAMA_CPP_PARITY   0

Functions
static int	ck_attention_full_ggml_graph_oracle_multihead (const float *q, const float *k, const float *v, float *output, int num_heads, int num_kv_heads, int num_tokens, int head_dim, int aligned_head_dim, int kv_stride_tokens, float scale)
static int	ck_attention_head_full_ggml_graph_oracle_regular (const float *q_head, const float *k_head, const float *v_head, float *out_head, int num_tokens, int head_dim, int aligned_head_dim, float scale)

Macro Definition Documentation

CK_ENABLE_LLAMA_CPP_PARITY

#define CK_ENABLE_LLAMA_CPP_PARITY   0

Definition at line 5 of file attention_oracle_ggml.h.

Function Documentation

ck_attention_full_ggml_graph_oracle_multihead()

static int ck_attention_full_ggml_graph_oracle_multihead ( const float *  q, const float *  k, const float *  v, float *  output, int  num_heads, int  num_kv_heads, int  num_tokens, int  head_dim, int  aligned_head_dim, int  kv_stride_tokens, float  scale  )

inlinestatic

Definition at line 53 of file attention_oracle_ggml.h.

{
    (void) q;
    (void) k;
    (void) v;
    (void) output;
    (void) num_heads;
    (void) num_kv_heads;
    (void) num_tokens;
    (void) head_dim;
    (void) aligned_head_dim;
    (void) kv_stride_tokens;
    (void) scale;
    return 0;
}

ck_attention_head_full_ggml_graph_oracle_regular()

static int ck_attention_head_full_ggml_graph_oracle_regular ( const float *  q_head, const float *  k_head, const float *  v_head, float *  out_head, int  num_tokens, int  head_dim, int  aligned_head_dim, float  scale  )

inlinestatic

Definition at line 33 of file attention_oracle_ggml.h.

{
    (void) q_head;
    (void) k_head;
    (void) v_head;
    (void) out_head;
    (void) num_tokens;
    (void) head_dim;
    (void) aligned_head_dim;
    (void) scale;
    return 0;
}