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. Learn more on the Loom ecosystem website.

Quick start

pip install "hrx-loom[hf]==0.1.0"
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	61.5	72.6	+18 %	82.6	88 %	61.9
decode tok/s, 2k context	55.9	68.6	+23 %	78.7	87 %	57.9
decode tok/s, 32k context	42.2	48.0	+14 %	56.1	86 %	42.6
prefill tok/s, 1k document	1750.6	2807.3	+60 %	7518.2	37 %	1478.8
prefill tok/s, 2k document	1851.2	2820.8	+52 %	7443.3	38 %	1563.4
prefill tok/s, 8k document	1804.3	2599.3	+44 %	7049.8	37 %	1575.6
prefill tok/s, 16k document	1632.4	2444.2	+50 %	6585.7	37 %	1411.3
prefill tok/s, 32k document	1339.9	2112.6	+58 %	5819.4	36 %	1158.7
4 streams, aggregate decode tok/s	155.6	251.6	+62 %	294.8	85 %	164.9
speculative decoding, drafted tokens accepted	n/a	n/a	n/a	n/a	n/a	n/a

Methodology

• Compared: Loom against llama.cpp ROCm llamacpp-rocm/b1374, llama.cpp Vulkan llamacpp-vulkan/b12211; the same GGUF, the same DevLab runner and job, the same prompts; Loom served with the active wait. Values are the mean of 3 rounds (4 streams: median of 3); Loom is timed by the client, llama.cpp by its server.
• Drafting, 1-stream rows (decode, decode_2k, decode_32k, prefill): no drafting on either engine (no MTP head in the GGUF and no drafter published for this exact model). Each engine is loaded once for all these rows.
• 4 streams: both engines loaded with drafting off, 4 parallel requests, aggregate tokens per second.
• Rows: decode = short-context decode; decode_2k / decode_32k = decode after a ~2k / ~32k-token context; prefill_Nk = prompt processing of an N k-token prompt; streams4 = 4-stream aggregate decode. A margin within +-3 % (the measured same-runner noise) is a tie.
• llama.cpp ROCm server arguments: 1-stream rows: -ngl 99 -cram 0 -c 65536 -np 1; 4 streams: -ngl 99 -cram 0 -c 65536 -kvu -np 4.
• llama.cpp Vulkan server arguments: 1-stream rows: -ngl 99 -cram 0 -c 36864 -np 1; 4 streams: -ngl 99 -cram 0 -c 147456 -np 4.
• Measured on AMD DevLab Dispatch Strix Halo runners (devlab-perf jobs); runners differ between sessions.
• Measured 2026-11-13; device AMD Strix Halo (DevLab Dispatch runner X2-LA03-S04-D04), Linux; hrx-loom 0.1.0 (hrx-system d01d1c1, recovered); source: DevLab Dispatch run 43390795404.
• llama.cpp Vulkan: measured 2026-11-13; source: DevLab Dispatch run 43390795404. Archive sha256 9759acd.
• 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 acd9ac7).
• — 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	—	not applicable
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 (78/78) 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	— not applicable
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 ²
needle_full	Needle-in-a-haystack prompt at the full context is served	? not re-run on this package (the last pass was on a previous build, sha256 d262555; the nightly gates re-run it)
decision	Decision answers (choice, score, noul probabilities) vs llama.cpp on the same prompts	— not applicable

• ¹ DevLab Dispatch run 43377116500, 2026-11-13; 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 43377116500, LEVEL=full, 2026-11-13; 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 1364b4e; hrx-loom 0.1.0 (hrx-system d01d1c1); card generator acd9ac7.

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