diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2025-06-18 07:29:33 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-06-18 07:29:33 +0300 |
| commit | dc96820ddb45c639ea4e149e4bbfcb0b67fbcc2b (patch) | |
| tree | 2ac3011164d541f5899db1afdad375cc59bfc142 /src/llama.cpp | |
| parent | 8b3002bba2ea64b1de9ca2ff87207d8c37b0f08e (diff) | |
Much faster CPU prompt processing (part 2) (#533)
* iq4_ks
203 t/s -> 357 t/s. iq4_ks_r4 is 242 t/s.
* iq4_k
175 t/s -> 353 t/s. iq4_k_r4 is 208 t/s.
PPL is actually lower!
* iq5_ks
180 t/s -> 359 t/s. iq5_ks_r4 is 210 t/s.
PPL is actually lower - 7.4160 vs 7.4494 for LlaMA-3.1-8B-Instruct
* iq5_k - accuracy loss is too big
* iq5_k - there was a bug with the shifts
...and that's why PPL was so high. It is also high on main.
This fixes it.
* iq6_k
148 t/s -> 350 t/s. There is no iq6_k_r4
PPL is actually lower because we have a bug in the existing
implementation!
* iq3_k
169 t/s -> 363 t/s. iq3_k_r4 is at 200 t/s.
* iq2_k
190 t/s -> 364 t/s. iq2_k_r4 is at 232 t/s.
* iq2_ks
200 t/s -> 367 t/s. There is no iq2_ks_r4.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'src/llama.cpp')
0 files changed, 0 insertions, 0 deletions