30 lines
1.0 KiB
Bash
30 lines
1.0 KiB
Bash
# Tunables for frank-default-large (HauhauCS Qwen3.6-27B Q8_K_P, DFlash+ngram-mod, dankclaw)
|
|
# Runtime: llama.cpp llama-server; AMD Strix Halo (gfx1151) 96GB UMA; Ryzen AI MAX+ 395.
|
|
MODEL_DIR=/home/dank/models
|
|
ALIAS=frank-default-large
|
|
HOST=0.0.0.0
|
|
PORT=7770
|
|
# 3 parallel slots x 200700 ctx each = 602100 total context pool.
|
|
CTX_SIZE=602100
|
|
PARALLEL=3
|
|
# Spec: DFlash draft model. Bench (4k, ROCm) showed draft-dflash fastest
|
|
# generation (20.2 t/s): DFlash+ngram 16.9, ngram-only 8.7. ngram layering
|
|
# reduced throughput, so it is not layered in production.
|
|
SPEC_TYPE=draft-dflash
|
|
SPEC_DRAFT_N_MAX=8
|
|
SPEC_NGRAM_MOD_N_MIN=48
|
|
SPEC_NGRAM_MOD_N_MAX=64
|
|
SPEC_NGRAM_MOD_N_MATCH=24
|
|
# Backend device from --list-devices. ROCm wins on generation; Vulkan renders
|
|
# DFlash spec unusable (NONE-op abort) and gen is slower on ngram (6.9 t/s).
|
|
DEVICE=ROCm0
|
|
# Dense model (Qwen3.6-27B, not MoE): no MoE CPU offload. Full offload.
|
|
N_GPU_LAYERS=all
|
|
# KV on GPU, quantized (K q8_0, V f16).
|
|
CACHE_TYPE_K=q8_0
|
|
CACHE_TYPE_V=f16
|
|
BATCH_SIZE=12288
|
|
UBATCH_SIZE=4096
|
|
THREADS=8
|
|
THREADS_HTTP=8
|