diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2025-07-02 09:27:47 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-07-02 09:27:47 +0200 |
| commit | 3248a3599275c3d2a98bc41d512e1d3ba66e32ba (patch) | |
| tree | 14bdaae22a6db922aa9eb0a7131141da81dfd251 /ggml/include/ggml.h | |
| parent | 46f2e5d249568b6395d17a80feb2ba9ab48156e5 (diff) | |
Adding IQ3_KS quants (#566)
* iq3_ks: basics
* iq3_ks: CUDA dequantize
* iq3_ks: CUDA mmvq
* iq3_ks: mmq
* iq3_ks: faster mmq
* iq3_ks: Zen4
* iq3_ks: AVX2 convert to q8_k_r8
This gives usPP-512 = 360 t/s.
* iq3_ks: AVX2 GEMM/GEMV
* iq3_ks: NEON GEMM/GEMV
* iq3_ks: NEON convert to q8_k_r8
This gives us PP-512 = 164 t/s.
* iq3_ks: Metal dequantize
* iq3_ks: Metal gemv - pathetic performance
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'ggml/include/ggml.h')
| -rw-r--r-- | ggml/include/ggml.h | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h index 67551eb2..aea0fac5 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -429,6 +429,7 @@ extern "C" { GGML_TYPE_IQ2_KT = 153, GGML_TYPE_IQ3_KT = 154, GGML_TYPE_IQ4_KT = 155, + GGML_TYPE_IQ3_KS = 156, GGML_TYPE_Q4_0_R8 = 202, GGML_TYPE_Q5_0_R4 = 206, @@ -521,6 +522,7 @@ extern "C" { GGML_FTYPE_MOSTLY_IQ2_KT = 142, // except 1d tensors GGML_FTYPE_MOSTLY_IQ3_KT = 143, // except 1d tensors GGML_FTYPE_MOSTLY_IQ4_KT = 144, // except 1d tensors + GGML_FTYPE_MOSTLY_IQ3_KS = 145, // except 1d tensors // GGML_FTYPE_MOSTLY_Q4_0_R8 = 202, // except 1d tensors GGML_FTYPE_MOSTLY_Q8_0_R8 = 207, // except 1d tensors |