The ROCm HIP runtime (libamdhip64) reads the TEMP environment variable as
its JIT kernel-cache temp directory. Commit db3eb01 added TEMP=1.0 to the
env files as a sampling default; the serve scripts export it via set -a.
HIP then used the path "1.0" for its cache, failed, and segfaulted with a
NULL-refcount unref (offset 0x446153 in libamdhip64) during model tensor
upload (ggml_backend_cuda_buffer_set_tensor). This crashed frank-default-
large on every restart from Aug 13 16:22 onward; the earlier c984a3e
"fix" (KV/batch changes) was unrelated.
Rename the sampling variable TEMP -> TEMPERATURE in both env files and
serve scripts. llama-server still receives --temp from the renamed var, so
sampling defaults are unchanged. The HIP runtime no longer sees a bogus
TEMP value. Affects frank-default-large (dankclaw) and frank-default-small
(dankserver).
The Aug 13 change crashed frank-default-large in a restart loop. The service
segfaulted ~15s into model load (SIGSEGV in hipMalloc during load). The
benchmark (frank-default-large-benchmark.md) already proved ubatch 2048 OOMs
(40.9GB prefill buffer) and batch 2048 / ubatch 512 is the fits-and-loads
config on the 96GiB UMA budget. Restore q8_0/q8_0 KV cache (FA pairing) and
batch 2048 / ubatch 512.
Also fix deployment.md: KV is q8_0 K / q8_0 V (not f16 V) since 334b6af.
- Remove CACHE_TYPE_K/CACHE_TYPE_V (fp16 default on GPU)
- Increase BATCH_SIZE from 2048 to 8192
- Increase UBATCH_SIZE from 512 to 2048
- Remove --cache-type-k and --cache-type-v from llama-server flags
- frank-default-large: add TEMP=1.0, TOP_K=20, TOP_P=0.95, MIN_P=0.0
(per HauhauCS Qwen3.6-27B model card, thinking-general mode)
- frank-default-small: add same sampling defaults
(per HauhauCS Qwen3.6-35B-A3B model card)
- frank-default-small: reduce N_CPU_MOE from 33 to 34 (6 expert
layers on GPU instead of 7) to prevent OOM during high-load sessions
- Both serve scripts pass --temp, --top-k, --top-p, --min-p to
llama-server; clients can still override per-request via API
Root cause of ~1 t/s long-context decode: llama.cpp build has
GGML_CUDA_FA_ALL_QUANTS=OFF, and FA requires K type == V type.
Mixed cache q8_0/f16 silently disabled FA, collapsing decode at
200k ctx. Set V to q8_0 (halves V memory too) and retune
threads 12/1 for 16 Zen5 cores.
- PARALLEL=4 (800000 total ctx), batch=8192, ubatch=2048
- n-cpu-moe=32 (moved 8 expert layers back to GPU)
- ctx-checkpoints=4 for headroom during larger batch compute
- All other tunables unchanged (ngram-mod, no-mmap, no-kv-offload, q8_0/f16 KV)
Lower each stream to 200000 ctx (pool 600000) to free KV VRAM for the larger
batch. Add --no-mmap so the CPU-resident MoE experts load eagerly into RAM
(avoids decode page faults). Keep KV on GPU. Experts already fully on CPU.
Bigger logical/physical batches for prefill throughput. Enable
--ctx-checkpoints to park attention checkpoints in RAM, holding peak GPU
memory down so the larger ubatch fits the 24GB VRAM alongside 3x262144 KV
without OOM. Weight split unchanged (experts CPU, non-experts GPU).
Retire vllm-awq-qwen36 (AWQ) from target maps, README, deployment, KB index,
Dankserver specs, DFlash and FP8 docs. New dankserver runtime: upstream
llama.cpp, port 7770, alias frank-default-small, 3x262144 ctx, DFlash Q8_0
draft, MoE experts offloaded to CPU, attention+KV on GPU (K q8_0, V f16).
deploy.sh now maps dankserver to frank-default-small and symlinks its unit.
models.json: dankserver baseUrl -> :7770, contextWindow -> 262144.
llama.cpp splits --ctx-size across --parallel slots, so 262144 total gave
87552 per stream. Non-negotiable is 262144 per stream: total pool must be
786432.
llama.cpp service replacing retired vllm-awq-qwen36 on port 7770 with alias
frank-default-small. Wrapper sources frank-default-small.env for tunables:
3 parallel slots at 262144 ctx, draft-dflash n-max 7, KV K=q8_0 V=f16 on GPU,
all MoE experts on CPU (--n-cpu-moe 40), non-expert tensors on GPU, mmproj
vision + jinja. Unit: frank-default-small.service (ExecStart in dankserver
repo clone /home/dank/vLLM/scripts).
Update target maps (AGENTS, README, deployment, deploy.sh) from the retired
ROCmFPX llama.cpp stack to the active vLLM-ROCm CIRU INT4 service on port 7770
with MTP speculative decoding. Align Dankserver to the running
vllm-awq-qwen36 (AWQ, port 34401). Refresh KB pages: Dankclaw system specs
(gpu-mem-util 0.95, CIRU venv), quantization choices (INT4 primary, MXFP4
experiment), DFlash/DSpark Dankclaw section, and the KB index topology.
The launcher does not change directory, so process cwd becomes sys.path[0]
for the 'python -m ...registry' model-inspection subprocess. From the CIRU
package root, the top-level vllm/ repo dir shadows the editable install as a
namespace package, breaking lazy imports (SamplingParams) and failing model
architecture inspection. cd to /home/dank/ling3-ciru before exec.
- Normalize 'VLLM' to 'vLLM' in prose across 6 docs (32 occurrences; wikilink
targets untouched)
- Remove duplicated Environment Variables block in 27B Dankclaw doc
- Update stale model refs Qwen3-30B-A3B -> Qwen3.6-35B-A3B in example commands
- Fix VLLM Index: -- separators to em-dashes; correct model names on links
- README: document active services (ROCmFPX llama-server on Dankclaw, vLLM
AWQ on Dankserver), replace broken [[docs/README.md]] link with docs/index.md
- scripts/deploy.sh: target-aware deploy (dankclaw/dankserver). Pushes main,
pulls on the target, installs/enables systemd unit, restarts service, and
checks the health endpoint; --dry-run preview; pre-flight dirty-tree check
- docs/deployment.md: full deployment/update process with mermaid workflow,
target map, manual update commands, service units, and troubleshooting
- AGENTS.md: add Deployment & Update Process section (deploy.sh usage, target
map, mermaid workflow) and a Summary table row
- docs/index.md: crosslink deployment doc
Frank