Skip to content

Commit

Permalink
cuda : reuse ggml-common (cont)
Browse files Browse the repository at this point in the history 
ggml-ci
  • Loading branch information
@ggerganov
ggerganov committed Mar 9, 2024
1 parent 3656c76 commit 40b0a47
Show file tree
Hide file tree
Showing 3 changed files with 158 additions and 117 deletions.
182 changes: 111 additions & 71 deletions ggml-common.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,19 +3,29 @@
#if defined(GGML_COMMON_DECL_C)
#include <stdint.h>

typedef uint16_t ggml_fp16_t;
typedef uint16_t ggml_half;
typedef uint32_t ggml_half2;

#define GGML_COMMON_AGGR

#define GGML_COMMON_DECL
#elif defined(GGML_COMMON_DECL_METAL)
#include <metal_stdlib>

typedef half ggml_fp16_t;
typedef half ggml_half;
typedef half2 ggml_half2;

#define GGML_COMMON_AGGR

#define GGML_COMMON_DECL
#elif defined(GGML_COMMON_DECL_CUDA)
#include <cuda_fp16.h>
#include <cstdint>

typedef half ggml_fp16_t;
typedef half ggml_half;
typedef half2 ggml_half2;

#define GGML_COMMON_AGGR data

#define GGML_COMMON_DECL
#endif
Expand All @@ -40,60 +50,75 @@ typedef half ggml_fp16_t;
#define QI4_0 (QK4_0 / (4 * QR4_0))
#define QR4_0 2
typedef struct {
ggml_fp16_t d; // delta
uint8_t qs[QK4_0 / 2]; // nibbles / quants
ggml_half d; // delta
uint8_t qs[QK4_0 / 2]; // nibbles / quants
} block_q4_0;
static_assert(sizeof(block_q4_0) == sizeof(ggml_fp16_t) + QK4_0 / 2, "wrong q4_0 block size/padding");
static_assert(sizeof(block_q4_0) == sizeof(ggml_half) + QK4_0 / 2, "wrong q4_0 block size/padding");

#define QK4_1 32
#define QI4_1 (QK4_1 / (4 * QR4_1))
#define QR4_1 2
typedef struct {
ggml_fp16_t d; // delta
ggml_fp16_t m; // min
uint8_t qs[QK4_1 / 2]; // nibbles / quants
union {
struct {
ggml_half d; // delta
ggml_half m; // min
} GGML_COMMON_AGGR;
ggml_half2 dm;
};
uint8_t qs[QK4_1 / 2]; // nibbles / quants
} block_q4_1;
static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_fp16_t) + QK4_1 / 2, "wrong q4_1 block size/padding");
static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_half) + QK4_1 / 2, "wrong q4_1 block size/padding");

#define QK5_0 32
#define QI5_0 (QK5_0 / (4 * QR5_0))
#define QR5_0 2
typedef struct {
ggml_fp16_t d; // delta
ggml_half d; // delta
uint8_t qh[4]; // 5-th bit of quants
uint8_t qs[QK5_0 / 2]; // nibbles / quants
} block_q5_0;
static_assert(sizeof(block_q5_0) == sizeof(ggml_fp16_t) + sizeof(uint32_t) + QK5_0 / 2, "wrong q5_0 block size/padding");
static_assert(sizeof(block_q5_0) == sizeof(ggml_half) + sizeof(uint32_t) + QK5_0 / 2, "wrong q5_0 block size/padding");

#define QK5_1 32
#define QI5_1 (QK5_1 / (4 * QR5_1))
#define QR5_1 2
typedef struct {
ggml_fp16_t d; // delta
ggml_fp16_t m; // min
union {
struct {
ggml_half d; // delta
ggml_half m; // min
} GGML_COMMON_AGGR;
ggml_half2 dm;
};
uint8_t qh[4]; // 5-th bit of quants
uint8_t qs[QK5_1 / 2]; // nibbles / quants
} block_q5_1;
static_assert(sizeof(block_q5_1) == 2 * sizeof(ggml_fp16_t) + sizeof(uint32_t) + QK5_1 / 2, "wrong q5_1 block size/padding");
static_assert(sizeof(block_q5_1) == 2 * sizeof(ggml_half) + sizeof(uint32_t) + QK5_1 / 2, "wrong q5_1 block size/padding");

#define QK8_0 32
#define QI8_0 (QK8_0 / (4 * QR8_0))
#define QR8_0 1
typedef struct {
ggml_fp16_t d; // delta
int8_t qs[QK8_0]; // quants
ggml_half d; // delta
int8_t qs[QK8_0]; // quants
} block_q8_0;
static_assert(sizeof(block_q8_0) == sizeof(ggml_fp16_t) + QK8_0, "wrong q8_0 block size/padding");
static_assert(sizeof(block_q8_0) == sizeof(ggml_half) + QK8_0, "wrong q8_0 block size/padding");

#define QK8_1 32
#define QI8_1 (QK8_1 / (4 * QR8_1))
#define QR8_1 1
typedef struct {
float d; // delta
float s; // d * sum(qs[i])
int8_t qs[QK8_1]; // quants
union {
struct {
ggml_half xxxd; // delta
ggml_half xxxs; // d * sum(qs[i])
} GGML_COMMON_AGGR;
ggml_half2 ds;
};
int8_t qs[QK8_1]; // quants
} block_q8_1;
static_assert(sizeof(block_q8_1) == 2*sizeof(float) + QK8_1, "wrong q8_1 block size/padding");
static_assert(sizeof(block_q8_1) == 2*sizeof(ggml_half) + QK8_1, "wrong q8_1 block size/padding");

//
// Super-block quantization structures
Expand All @@ -117,10 +142,15 @@ static_assert(sizeof(block_q8_1) == 2*sizeof(float) + QK8_1, "wrong q8_1 block s
typedef struct {
uint8_t scales[QK_K/16]; // scales and mins, quantized with 4 bits
uint8_t qs[QK_K/4]; // quants
ggml_fp16_t d; // super-block scale for quantized scales
ggml_fp16_t dmin; // super-block scale for quantized mins
union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins
} GGML_COMMON_AGGR;
ggml_half2 dm;
};
} block_q2_K;
static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_fp16_t) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding");
static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_half) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding");

// 3-bit quantization
// weight is represented as x = a * q
Expand All @@ -130,20 +160,20 @@ static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_fp16_t) + QK_K/16 + QK_K/4, "w
#define QR3_K 4
#ifdef GGML_QKK_64
typedef struct {
uint8_t hmask[QK_K/8]; // quants - high bit
uint8_t qs[QK_K/4]; // quants - low 2 bits
uint8_t hmask[QK_K/8]; // quants - high bit
uint8_t qs[QK_K/4]; // quants - low 2 bits
uint8_t scales[2];
ggml_fp16_t d; // super-block scale
ggml_half d; // super-block scale
} block_q3_K;
static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 + 2, "wrong q3_K block size/padding");
static_assert(sizeof(block_q3_K) == sizeof(ggml_half) + QK_K / 4 + QK_K / 8 + 2, "wrong q3_K block size/padding");
#else
typedef struct {
uint8_t hmask[QK_K/8]; // quants - high bit
uint8_t qs[QK_K/4]; // quants - low 2 bits
uint8_t scales[12]; // scales, quantized with 6 bits
ggml_fp16_t d; // super-block scale
uint8_t hmask[QK_K/8]; // quants - high bit
uint8_t qs[QK_K/4]; // quants - low 2 bits
uint8_t scales[12]; // scales, quantized with 6 bits
ggml_half d; // super-block scale
} block_q3_K;
static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 + 12, "wrong q3_K block size/padding");
static_assert(sizeof(block_q3_K) == sizeof(ggml_half) + QK_K / 4 + QK_K / 8 + 12, "wrong q3_K block size/padding");
#endif

// 4-bit quantization
Expand All @@ -154,19 +184,24 @@ static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 +
#define QR4_K 2
#ifdef GGML_QKK_64
typedef struct {
ggml_fp16_t d[2]; // super-block scales/mins
uint8_t scales[2]; // 4-bit block scales/mins
uint8_t qs[QK_K/2]; // 4--bit quants
ggml_half d[2]; // super-block scales/mins
uint8_t scales[2]; // 4-bit block scales/mins
uint8_t qs[QK_K/2]; // 4--bit quants
} block_q4_K;
static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + QK_K/2 + 2, "wrong q4_K block size/padding");
static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_half) + QK_K/2 + 2, "wrong q4_K block size/padding");
#else
typedef struct {
ggml_fp16_t d; // super-block scale for quantized scales
ggml_fp16_t dmin; // super-block scale for quantized mins
union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins
} GGML_COMMON_AGGR;
ggml_half2 dm;
};
uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits
uint8_t qs[QK_K/2]; // 4--bit quants
uint8_t qs[QK_K/2]; // 4--bit quants
} block_q4_K;
static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/2, "wrong q4_K block size/padding");
static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_half) + K_SCALE_SIZE + QK_K/2, "wrong q4_K block size/padding");
#endif

// 5-bit quantization
Expand All @@ -177,21 +212,26 @@ static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/
#define QR5_K 2
#ifdef GGML_QKK_64
typedef struct {
ggml_fp16_t d; // super-block scale
int8_t scales[QK_K/16]; // 8-bit block scales
uint8_t qh[QK_K/8]; // quants, high bit
uint8_t qs[QK_K/2]; // quants, low 4 bits
ggml_half d; // super-block scale
int8_t scales[QK_K/16]; // 8-bit block scales
uint8_t qh[QK_K/8]; // quants, high bit
uint8_t qs[QK_K/2]; // quants, low 4 bits
} block_q5_K;
static_assert(sizeof(block_q5_K) == sizeof(ggml_fp16_t) + QK_K/2 + QK_K/8 + QK_K/16, "wrong q5_K block size/padding");
static_assert(sizeof(block_q5_K) == sizeof(ggml_half) + QK_K/2 + QK_K/8 + QK_K/16, "wrong q5_K block size/padding");
#else
typedef struct {
ggml_fp16_t d; // super-block scale for quantized scales
ggml_fp16_t dmin; // super-block scale for quantized mins
uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits
uint8_t qh[QK_K/8]; // quants, high bit
uint8_t qs[QK_K/2]; // quants, low 4 bits
union {
struct {
ggml_half d; // super-block scale for quantized scales
ggml_half dmin; // super-block scale for quantized mins
} GGML_COMMON_AGGR;
ggml_half2 dm;
};
uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits
uint8_t qh[QK_K/8]; // quants, high bit
uint8_t qs[QK_K/2]; // quants, low 4 bits
} block_q5_K;
static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/2 + QK_K/8, "wrong q5_K block size/padding");
static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_half) + K_SCALE_SIZE + QK_K/2 + QK_K/8, "wrong q5_K block size/padding");
#endif

// 6-bit quantization
Expand All @@ -204,9 +244,9 @@ typedef struct {
uint8_t ql[QK_K/2]; // quants, lower 4 bits
uint8_t qh[QK_K/4]; // quants, upper 2 bits
int8_t scales[QK_K/16]; // scales, quantized with 8 bits
ggml_fp16_t d; // super-block scale
ggml_half d; // super-block scale
} block_q6_K;
static_assert(sizeof(block_q6_K) == sizeof(ggml_fp16_t) + QK_K / 16 + 3*QK_K/4, "wrong q6_K block size/padding");
static_assert(sizeof(block_q6_K) == sizeof(ggml_half) + QK_K / 16 + 3*QK_K/4, "wrong q6_K block size/padding");

// This is only used for intermediate quantization and dot products
typedef struct {
Expand All @@ -222,42 +262,42 @@ static_assert(sizeof(block_q8_K) == sizeof(float) + QK_K + QK_K/16*sizeof(int16_
#define QI2_XXS (QK_K / (4*QR2_XXS))
#define QR2_XXS 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint16_t qs[QK_K/8];
} block_iq2_xxs;
static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding");
static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_half) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding");

// 2.3125 bpw quants
#define QI2_XS (QK_K / (4*QR2_XS))
#define QR2_XS 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint16_t qs[QK_K/8];
uint8_t scales[QK_K/32];
} block_iq2_xs;
static_assert(sizeof(block_iq2_xs) == sizeof(ggml_fp16_t) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding");
static_assert(sizeof(block_iq2_xs) == sizeof(ggml_half) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding");

// 2.5625 bpw quants
#define QI2_S (QK_K / (4*QR2_S))
#define QR2_S 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint8_t qs[QK_K/4];
uint8_t qh[QK_K/32];
uint8_t scales[QK_K/32];
} block_iq2_s;
static_assert(sizeof(block_iq2_s) == sizeof(ggml_fp16_t) + QK_K/4 + QK_K/16, "wrong iq2_s block size/padding");
static_assert(sizeof(block_iq2_s) == sizeof(ggml_half) + QK_K/4 + QK_K/16, "wrong iq2_s block size/padding");

// (Almost) "true" 3-bit quantization.
// Due to the need to use blocks as per ggml design, it ends up using
// 3.0625 bpw because of the 16-bit scale for each block of 256.
#define QI3_XXS (QK_K / (4*QR3_XXS))
#define QR3_XXS 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint8_t qs[3*QK_K/8];
} block_iq3_xxs;
static_assert(sizeof(block_iq3_xxs) == sizeof(ggml_fp16_t) + 3*(QK_K/8), "wrong iq3_xxs block size/padding");
static_assert(sizeof(block_iq3_xxs) == sizeof(ggml_half) + 3*(QK_K/8), "wrong iq3_xxs block size/padding");

// 3.4375 bpw
#if QK_K == 64
Expand All @@ -268,32 +308,32 @@ static_assert(sizeof(block_iq3_xxs) == sizeof(ggml_fp16_t) + 3*(QK_K/8), "wrong
#define QI3_XS (QK_K / (4*QR3_XS))
#define QR3_XS 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint8_t qs[QK_K/4];
uint8_t qh[QK_K/32];
uint8_t signs[QK_K/8];
uint8_t scales[IQ3S_N_SCALE];
} block_iq3_s;
static_assert(sizeof(block_iq3_s) == sizeof(ggml_fp16_t) + 13*(QK_K/32) + IQ3S_N_SCALE, "wrong iq3_s block size/padding");
static_assert(sizeof(block_iq3_s) == sizeof(ggml_half) + 13*(QK_K/32) + IQ3S_N_SCALE, "wrong iq3_s block size/padding");

#define QI1_S (QK_K / (4*QR1_S))
#define QR1_S 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint8_t qs[QK_K/8];
uint8_t scales[QK_K/16];
} block_iq1_s;
static_assert(sizeof(block_iq1_s) == sizeof(ggml_fp16_t) + QK_K/8 + QK_K/16, "wrong iq1_s block size/padding");
static_assert(sizeof(block_iq1_s) == sizeof(ggml_half) + QK_K/8 + QK_K/16, "wrong iq1_s block size/padding");

// Non-linear quants
#define QK4_NL 32
#define QI4_NL (QK4_NL / (4*QR4_NL))
#define QR4_NL 2
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint8_t qs[QK4_NL/2];
} block_iq4_nl;
static_assert(sizeof(block_iq4_nl) == sizeof(ggml_fp16_t) + QK4_NL/2, "wrong iq4_nl block size/padding");
static_assert(sizeof(block_iq4_nl) == sizeof(ggml_half) + QK4_NL/2, "wrong iq4_nl block size/padding");

#if QK_K == 64
#define block_iq4_xs block_iq4_nl
Expand All @@ -304,12 +344,12 @@ static_assert(sizeof(block_iq4_nl) == sizeof(ggml_fp16_t) + QK4_NL/2, "wrong iq4
#define QI4_XS (QK_K / (4*QR4_XS))
#define QR4_XS 8
typedef struct {
ggml_fp16_t d;
ggml_half d;
uint16_t scales_h;
uint8_t scales_l[QK_K/64];
uint8_t qs[QK_K/2];
} block_iq4_xs;
static_assert(sizeof(block_iq4_xs) == sizeof(ggml_fp16_t) + sizeof(uint16_t) + QK_K/64 + QK_K/2, "wrong iq4_xs block size/padding");
static_assert(sizeof(block_iq4_xs) == sizeof(ggml_half) + sizeof(uint16_t) + QK_K/64 + QK_K/2, "wrong iq4_xs block size/padding");
#endif

#endif // GGML_COMMON_DECL
Expand Down
Loading

0 comments on commit 40b0a47

Please sign in to comment.