dank/serverAppConfigs
1
0
Fork 0
Code Issues 6 Pull Requests 1 Actions Packages Projects Releases Wiki Activity
32 Commits 48 Branches 8 Tags
841d872aa7cc972e09dbb427dfd3b7b145d174c2
Commit Graph
22 Commits
Author SHA1 Message Date
Frank 841d872aa7 sync scripts from remotes: ROCmFPX llama-server.service (dankclaw), dankserver vLLM serve/validate; drop failed-attempt scripts 
- llama-server.service: dankclaw migrated to ROCmFPX runtime (Ling-3.0-flash,
  draft-mtp, no-mmap/direct-io) from regular llama.cpp Qwen3.6-27B-DFlash
- serve-vllm-qwen36-dflash.sh: dankserver version (relies on image 'vllm serve'
  entrypoint; HF cache mounted read-write)
- validate-vllm.sh: dankserver multi-stream vLLM validation (4 unique prompts,
  1-4 streams, prompt+completion speed)
- remove superseded scripts: serve-llamacpp-dflash.sh, validate.sh,
  vllm-awq-qwen36-dflash-ampere.sh, vllm-qwen36.service
 2026-08-10 19:48:46 -07:00
Frank ced4c6f46e sync: update vllm-awq-qwen36-dflash-ampere.sh from dankserver 
- Model: Avesed/Qwen3.6-35B-A3B-INT8-W8A8 (was AWQ 4bit)
- Image: vllm/vllm-openai:nightly (was :latest)
- Spec method: mtp with 2 tokens (was dflash with 12)
- Max model len: 200700 (was 32768)
- Max num seqs: 3 (was 1)
- Max batched tokens: 3072 (was 4096)
- GPU mem util: 0.97 (was 0.96)
- CPU offload: 30GB (was 40GB)
- Add --enable-prefix-caching
- Add --kv-cache-dtype auto
 2026-08-10 03:23:09 -07:00
Frank 11d66c3823 Set max-model-len to 32768 for testing 2026-08-09 22:19:32 -07:00
Frank 6bdabfc65b Add VLLM_USE_V2_MODEL_RUNNER=0 to skip V2 model runner MARLIN repacking OOM 2026-08-09 21:49:25 -07:00
Frank 83b0bea29b Fix: move commented speculative config to end to not break bash line continuation 2026-08-09 12:27:28 -07:00
Frank 88aa18ef34 Reduce GPU memory to 75%, add expandable_segments to fix OOM 2026-08-09 12:11:48 -07:00
Frank 9dfa16fae9 Disable speculative decoding - OOM on stable vLLM too 2026-08-09 12:08:26 -07:00
Frank 578c3e8f71 Enable DFlash speculative decoding (12 tokens), increase CPU offload to 20GB 2026-08-09 11:54:40 -07:00
Frank 8c13c7bbaa Add AGENTS.md git workflow rules and initial vllm-awq-qwen36-dflash-ampere.sh from dankserver 2026-08-09 11:54:04 -07:00
Frank 0f97dcf449 Update validation script to test parallel streams (4 concurrent requests) 2026-08-08 00:30:08 -07:00
Frank 9971bdef0f feat: initial vLLM Qwen3.6-35B-A3B-FP8 serve script with DFlash drafting on dankserver 
- Podman container with CUDA GPU access
- DFlash speculative decoding via --spec-method dflash
- FP8 KV cache, 4 parallel streams, 262K context
- CPU expert offloading via --cpu-offload-gb
- systemd service for persistent runtime
 2026-08-07 18:28:38 -07:00
Frank 5452bb63f1 Add vLLM serve script, systemd service, and validation test for Qwen3.6-35B-A3B-FP8 on dankserver 2026-08-07 18:19:40 -07:00
Frank 663e47866b Bump context to 600000 (200000/stream x 3 parallel) 2026-08-07 17:18:10 -07:00
Frank a5f72627bb Use --alias frank-default-large, multi-stream validation script 
- Replace --override-kv with --alias for model naming
- Validation script tests 1, 2, 3 parallel streams sequentially
- Model name confirmed: frank-default-large
 2026-08-07 17:15:09 -07:00
Frank 471e963350 Fix parallel validation, remove ngram-mod 
- Rewrite validation script with temp files for true parallel curl requests
- Both prompts now complete 1024 tokens each (was cutting first short)
- Remove ngram-mod from spec pipeline (was -1.2 tok/s regression)
- True parallel throughput: 15.68 tok/s across 2 streams
 2026-08-07 16:44:12 -07:00
Frank 2426ca563a Add ngram-mod to spec pipeline: -1.2 tok/s regression 
- DFlash+ngram-mod: 17.11 tok/s vs DFlash alone: 18.34 tok/s
- ngram overhead not offset by acceptance rate at current settings
- Keeping in place for now, may tune ngram-mod params later
 2026-08-07 16:25:49 -07:00
Frank 764a65080f Validation: 2 parallel prompts, ~18 tok/s with DFlash 
- Add second parallel prompt (Byzantine Empire history vs ML)
- Both hit 1024 tokens, combined 2048 in 111.6s = 18.34 tok/s
- Parallel throughput matches single-stream baseline
 2026-08-07 16:21:20 -07:00
Frank 25d57184a6 Enable DFlash speculative decoding, remove mlock 
- Add --spec-type draft-dflash (was missing, DFlash was inactive)
- Change --load-mode from mlock to none (RLIMIT_MEMLOCK too small for 2.7GB)
- Increase validation max_tokens from 256 to 1024 with complex ML prompt
- Baseline: 18.76 tok/s (was 7.62 tok/s without DFlash)
 2026-08-07 16:10:55 -07:00
Frank 1949f0651b Pivot to llama.cpp with DFlash speculative decoding on ROCm 
- Abandon vLLM/FP8 (Triton compilation hangs, ~0.5 tok/s on Strix Halo)
- llama.cpp v10133 with ROCm0 offload, DFlash draft model
- Main: unsloth/Qwen3.6-27B-MTP-GGUF:Q8_0 (~22GB)
- Draft: Alittlehammmer/Qwen3.6-27B-DFlash-GGUF-llama.cpp:Q8_0 (~1.8GB)
- 3 parallel streams, 131072 context/stream
- systemd service with minimal sandboxing (ROCm requires /dev/kfd access)
- Baseline: 7.62 tok/s (threshold 5.0 tok/s)
 2026-08-07 15:41:27 -07:00
Frank 2bb1b2ae1d Baseline: ThinkingCap 27B FP8 on dankclaw with enforce-eager and validation timeout 2026-08-07 08:58:20 -07:00
Frank e4fba35e89 Baseline: ThinkingCap 27B FP8 on dankclaw 
- vLLM 0.26.0+rocm723
- --kv-cache-dtype fp8 --gpu-memory-utilization 0.80
- --max-model-len 32768 --max-num-batched-tokens 2048
- --max-num-seqs 32 --reasoning-parser qwen3 --language-model-only
- Baseline response time: 219s for 50 tokens
 2026-08-07 08:49:59 -07:00
Frank 7dce375ab4 docs: fix dankclaw specs (30GiB RAM, 96GiB GPU-visible, ROCm 7.2.4) + add validate.sh 2026-08-06 22:40:44 -07:00