vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ggml : make i-quants work with super-blocks of 64 (CPU,Metal) (#5760)

Browse source 
* WIP: make i-quants work for QK_K = 64

* iq2_xs: attempt to fix AVX dot product for QK_K = 64

Tests pass, but I get gibberish.

* QK_K = 64 tests pass on ARM_NEON and Metal

Sadly, that does not mean it actually works.

* Make CUDA compile with QK_K = 64

Tests don't pass, plus we get misaligned access

* Q2_K: fixed bug in imatrix quantization for QK_K = 64

* iq1_s: turn off SIMD implementation for QK_K = 64 (it does not work)

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow 2024-02-28 10:37:02 +02:00 • committed by GitHub
parent cb49e0f8c9
commit 7c4263d426
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 194 additions and 59 deletions
Download patch file Download diff file Expand all files Collapse all files

27
ggml-cuda.cu
View file

@ -544,14 +544,19 @@ static_assert(sizeof(block_iq3_xxs) == sizeof(ggml_fp16_t) + 3*(QK_K/8), "wrong
#define QR3_XS 8
#define QI3_XS (QK_K / (4*QR3_XS))
#if QK_K == 64
#define IQ3S_N_SCALE 2
#else
#define IQ3S_N_SCALE QK_K/64
#endif
typedef struct {
half d;
uint8_t qs[QK_K/4];
uint8_t qh[QK_K/32];
uint8_t signs[QK_K/8];
uint8_t scales[QK_K/64];
uint8_t scales[IQ3S_N_SCALE];
} block_iq3_s;
static_assert(sizeof(block_iq3_s) == sizeof(ggml_fp16_t) + 27*(QK_K/64), "wrong iq3_s block size/padding");
static_assert(sizeof(block_iq3_s) == sizeof(ggml_fp16_t) + 13*(QK_K/32) + IQ3S_N_SCALE, "wrong iq3_s block size/padding");
#define QR1_S 8
#define QI1_S (QK_K / (4*QR1_S))
@ -571,6 +576,11 @@ typedef struct {
} block_iq4_nl;
static_assert(sizeof(block_iq4_nl) == sizeof(ggml_fp16_t) + QK4_NL/2, "wrong iq4_nl block size/padding");
#if QK_K == 64
#define block_iq4_xs block_iq4_nl
#define QR4_XS QR4_NL
#define QI4_XS QI4_NL
#else
// QR4_XS = 8 is very slightly faster than QR4_XS = 4
#define QR4_XS 8
#define QI4_XS (QK_K / (4*QR4_XS))
@ -581,7 +591,7 @@ typedef struct {
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");
#endif
#define WARP_SIZE 32
#define MATRIX_ROW_PADDING 512 // last row of quant. matrices is a multiple of this to avoid out-of-bounds memory accesses
@ -2439,9 +2449,9 @@ static __global__ void dequantize_block_iq4_nl(const void * __restrict__ vx, dst
}
#if QK_K != 64
template<typename dst_t>
static __global__ void dequantize_block_iq4_xs(const void * __restrict__ vx, dst_t * __restrict__ yy) {
const int i = blockIdx.x;
const block_iq4_xs * x = (const block_iq4_xs *)vx;
@ -2455,8 +2465,8 @@ static __global__ void dequantize_block_iq4_xs(const void * __restrict__ vx, dst
y[j+ 0] = d * kvalues_iq4nl[q4[j] & 0xf];
y[j+16] = d * kvalues_iq4nl[q4[j] >> 4];
}
}
#endif
static __global__ void dequantize_mul_mat_vec_q2_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) {
@ -5382,8 +5392,7 @@ static __device__ __forceinline__ float vec_dot_iq4_xs_q8_1(
return 0.f;
#endif
#else
assert(false);
return 0.f;
return vec_dot_iq4_xs_q8_1(vbq, bq8_1, iqs);
#endif
}
@ -7444,7 +7453,11 @@ static void dequantize_row_iq4_nl_cuda(const void * vx, dst_t * y, const int k,
template<typename dst_t>
static void dequantize_row_iq4_xs_cuda(const void * vx, dst_t * y, const int k, cudaStream_t stream) {
const int nb = (k + QK_K - 1) / QK_K;
#if QK_K == 64
dequantize_block_iq4_nl<<<nb, 32, 0, stream>>>(vx, y);
#else
dequantize_block_iq4_xs<<<nb, 32, 0, stream>>>(vx, y);
#endif
}
template <typename src_t, typename dst_t>