loom-fun/bartowski-Spark-X2.6-4B-loom

Spark-X2.6-4B (Q4_K_M) for AMD Loom

A Loom package (.loom programs and a model.json card) for the unmodified GGUF checkpoint bartowski/Spark-X2.6-4B-GGUF (Spark-X2.6-4B-Q4_K_M.gguf), generated on an AMD Ryzen AI Max+ 395 "Strix Halo" with the loom-skills workflow.

Quick start

pip install "hrx-loom[hf]"
loom-fetch https://huggingface.co/loom-fun/bartowski-Spark-X2.6-4B-loom/resolve/main/model.json --materialize=copy --output=Spark-X2.6-4B
loom-serve --model=Spark-X2.6-4B/model.json '--device=amdgpu://?completion_spin_ns=4294967295' --memory_bytes=55000000000

The server listens on 127.0.0.1:8080 and answers OpenAI-style /v1/chat/completions requests.

Performance vs llama.cpp

metric	llama.cpp	Loom	margin	roofline	Loom % of roofline	llama.cpp Vulkan
decode tok/s, short prompt	58.1	69.3	+19 %	87.5	79 %	61.7
decode tok/s, 2k context	49.2	64.4	+31 %	85.6	75 %	59.2
decode tok/s, 32k context	33.4	42.8	+28 %	60.9	70 %	41.9
prefill tok/s, 1k document	1374.5	2789.5	+103 %	7518.2	37 %	1520.5
prefill tok/s, 2k document	1394.0	2823.3	+103 %	7443.3	38 %	1563.7
prefill tok/s, 8k document	1431.0	2631.2	+84 %	7049.8	37 %	1569.0
prefill tok/s, 16k document	1114.0	2476.9	+122 %	6585.7	38 %	1418.2
prefill tok/s, 32k document	780.1	2116.8	+171 %	5819.4	36 %	1164.8
4 streams, aggregate decode tok/s	153.7	176.4	+15 %	314.3	56 %	161.6
speculative decoding, drafted tokens accepted	—	14 % ¹	—	n/a	n/a	n/a

• llama.cpp: build llamacpp-rocm/b1370, server arguments -ngl 99 -cram 0 -c 65536 -kvu -np 4.
• Loom: speculative decoding on (n-gram prompt lookup, depth 2).
• Same GGUF on both engines, tokens per second. Loom: client-side timing; llama.cpp: server-side timings. Runs per value: decode and prefill 2 (mean), 4 streams 3 (median).
• Measured on AMD DevLab Dispatch Strix Halo runners (devlab-perf jobs): Loom, served with the active wait, and llama.cpp on the same runner; runners differ between sessions.
• Measured 2026-11-11; device AMD Strix Halo (DevLab Dispatch runner X2-LA03-S04-D06), Linux; hrx-loom 0.1.0 (hrx-system d01d1c1, recovered); source: DevLab Dispatch run 43033815593.
• ¹ Measured 2026-11-08; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1); source: DevLab Dispatch run 42419939771.
• llama.cpp Vulkan: build llamacpp-vulkan/b12211 (archive sha256 9759acd), server arguments -ngl 99 -cram 0 -c 147456 -np 4; same GGUF, prompts and settings as llama.cpp; measured 2026-11-09 in its own DevLab job (another runner than the other columns: compare the engines through ratios, not across jobs); source: DevLab Dispatch run 42569157613. Differs from llama.cpp: KV cache: 4 per-slot caches (-np 4, -c 147456, no -kvu) instead of the ROCm baseline's unified pool: b12211's Vulkan backend has no GPU kernel for the fused flash attention over K/V rows that the unified pool uses.
• Drafted tokens accepted: the share of the drafter's proposed tokens the model accepted during the spec gate's requests.
• Roofline: the hardware ceiling for this GGUF on AMD Ryzen AI Max+ 395 "Strix Halo" (gfx1151) at 216 GB/s and 59 TFLOP/s, from loom-roofline roofline.py (loom-skills 14669f7); decode and 4 streams with speculative decoding: verify passes × (1 + acceptance × depth 2) tokens per pass, the drafter's cost not counted; acceptance measured in the row's own session: decode 3 %, decode at 2k 4 %, decode at 32k 4 %, 4 streams 3 %.
• — not measured; n/a not applicable.

Features

feature	status	detail
API	✅	OpenAI-style /v1/chat/completions
tool calling	✅	OpenAI-style tools and tool_calls
reasoning / thinking	✅	on by default; chat_template_kwargs: {"enable_thinking": false} turns it off
reasoning effort	✅	reasoning_effort: low, high (default high)
speculative decoding	✅	n-gram prompt lookup, depth 2, on by default
image input	—	not applicable
audio input / output	—	not applicable
sampling	✅	temperature, top_p, top_k, seed
context size	✅	262,144 tokens per request
concurrency	✅	4 concurrent requests
max output tokens	✅	4,096 (preset max_tokens)

Correctness

gate	description	result
kernel_tests	Unit tests of the generated kernels against reference implementations	✅ pass (57/57) ¹
golden	llama.cpp reference recorded on the same GGUF: greedy tokens and top-5 log-probs, short prompts and 1k-32k documents	✅ pass ²
compare	Greedy output vs the llama.cpp reference, token by token; every divergence is a near-tie	✅ pass ²
vision	Image input vs llama.cpp with the same projector: identical prompt tokens, replies equal up to near-ties	— not applicable
consistency	Concurrent requests give the same output as the same requests run alone	✅ pass ²
spec	Speculative decoding on gives exactly the output of speculative decoding off	✅ pass ²
agent	Multi-turn tool-calling loop: exact prompt rendering, history reuse, valid tool calls, output vs llama.cpp	✅ pass ²
agent_think_off	The tool-calling loop with thinking off	✅ pass ²
sampling	temperature, top_k and seed behave as specified; first-token distribution matches llama.cpp	✅ pass ²
needle	Needle-in-a-haystack prompt at up to half the context, compared with llama.cpp; over-capacity prompt refused	✅ pass ³
needle_full	Needle-in-a-haystack prompt at the full context is served	✅ pass ²
decision	Decision answers (choice, score, noul probabilities) vs llama.cpp on the same prompts	— not applicable

• ¹ DevLab Dispatch run 42391288313, 2026-11-07; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• ² DevLab Dispatch run 42419939771, LEVEL=full, 2026-11-08; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• ³ DevLab Dispatch run 42391288313, LEVEL=full, 2026-11-07; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• Gates are the loom-benchmark correctness gates (gates.sh); a gate without a recorded run shows no evidence.

Files

• model.json / model.json.lock: the card and the pinned source files (sha256 checked by loom-fetch).
• program/*.loom, weights.loom: the generated Loom programs.
• tokenizer.json: the tokenizer.
• llm_config.json: the configuration the programs were generated from.
• info.json: the measurements, gate evidence and versions this card shows; everything else comes from the files above.
• REPORT.md: how the values above were measured: setup, commands and the value of every run.

Versions: card schema loom-card/2; loom-skills 02927f3; hrx-loom 0.1.0 (hrx-system d01d1c1); card generator 14669f7.

Source files, fetched by loom-fetch from their own repositories:

• checkpoint: bartowski/Spark-X2.6-4B-GGUF @ 17682e3, Spark-X2.6-4B-Q4_K_M.gguf