diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2025-06-21 16:35:08 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-06-21 16:35:08 +0200 |
| commit | 4f97409b80dffa96abe1a31d0a06e6dde78e91b7 (patch) | |
| tree | 533bcb5cc7cc0bccf307317ae5b8d37403fd19b5 /ggml/src/iqk/iqk_mul_mat.cpp | |
| parent | a98b7678a305c560117ce0a63a3529f2aaa17acb (diff) | |
Faster ARM_NEON GEMM implementation for legacy quants (#546)
* iq2_kt and iq3_kt work with new int trellis
Much slower than the fp16 based trellis. I guess, Apple doesn't
have int8_t SIMD on the M2-Max GPU.
* q4_0
83.6 t/s -> 128.4 t/s. q4_0_r8 is at 123.5 t/s
* q5_0
74.2 t/s -> 128.5 t/s. q5_0_r4 is at 111.4 t/s.
* q6_0
74.2 t/s -> 128.8 t/s. q6_0_r4 is at 107.2 t/s.
* q8_0
84.5 -> 128.7 t/s. q8_0_r8 is at 131 t/s.
* iq4_nl
84.5 t/s -> 128.1 t/s. iq4_nl_r4 is at 120.4 t/s
* q4_1
74.4 -> 115.4 t/s. There is no repacked variant
* q5_1
64.2 t/s -> 114.9 t/s. There is no repacked variant.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'ggml/src/iqk/iqk_mul_mat.cpp')
| -rw-r--r-- | ggml/src/iqk/iqk_mul_mat.cpp | 14 |
1 files changed, 11 insertions, 3 deletions
diff --git a/ggml/src/iqk/iqk_mul_mat.cpp b/ggml/src/iqk/iqk_mul_mat.cpp index 41c4f980..ce0753a5 100644 --- a/ggml/src/iqk/iqk_mul_mat.cpp +++ b/ggml/src/iqk/iqk_mul_mat.cpp @@ -271,9 +271,16 @@ struct MulMat { } #else switch (type) { - case GGML_TYPE_IQ2_KT: return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; - case GGML_TYPE_IQ3_KT: return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; - case GGML_TYPE_IQ4_KT: return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_Q4_0 : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_Q4_1 : return nrc_y >= 32 ? GGML_TYPE_Q8_1 : type; + case GGML_TYPE_Q5_0 : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_Q5_1 : return nrc_y >= 32 ? GGML_TYPE_Q8_1 : type; + case GGML_TYPE_Q6_0 : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_Q8_0 : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_IQ4_NL : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_IQ2_KT : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_IQ3_KT : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; + case GGML_TYPE_IQ4_KT : return nrc_y >= 32 ? GGML_TYPE_Q8_0_R8 : type; default: break; } #endif @@ -913,6 +920,7 @@ bool MulMat::prepare(int typeA, int typeB, int ne00, MulMat& m, int /*Ny*/) { case GGML_TYPE_Q5_0_R4: case GGML_TYPE_Q6_0_R4: case GGML_TYPE_Q8_0_R8: + case GGML_TYPE_Q8_1: case GGML_TYPE_IQ4_NL_R4: return iqk_set_kernels_legacy_quants(ne00, typeA, typeB, m.funcs, m.func16); case GGML_TYPE_IQ1_BN: |