Pith Number

pith:OZLX7HA3

pith:2026:OZLX7HA3YXIBF5GH76HSDTRYMI

not attested not anchored not stored refs pending

Stream2LLM: Overlap Context Streaming and Prefill for Reduced Time-to-First-Token (TTFT)

Chengqi Luo, Divya Mahajan, Kexin Rong, Rajveer Bachkaniwala, Richard So

Stream2LLM overlaps incremental context retrieval with LLM prefill to cut time-to-first-token by up to 11x while matching non-streaming throughput.

arxiv:2604.16395 v3 · 2026-03-29 · cs.DB · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{OZLX7HA3YXIBF5GH76HSDTRYMI}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Our evaluation demonstrates that streaming architecture delivers up to 11x TTFT improvements, with cost-aware scheduling providing critical benefits under memory pressure, all while maintaining throughput parity with non-streaming baselines.

C2weakest assumption

The two collected real-world workloads (web crawling and approximate nearest neighbor search) are representative of production streaming patterns, and the adaptive preemption strategies incur acceptable overheads that do not erode the reported TTFT gains in practice.

C3one line summary

Stream2LLM delivers up to 11x lower TTFT by streaming context retrieval, adaptive preemption for append and update modes, and longest-common-prefix reuse in disaggregated LLM deployments while preserving throughput.

Receipt and verification

First computed	2026-05-20T00:04:31.857493Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

76577f9c1bc5d012f4c7ff8f21ce38621db5bd06b819327452f90da7b015b9b3

Aliases

arxiv: 2604.16395 · arxiv_version: 2604.16395v3 · doi: 10.48550/arxiv.2604.16395 · pith_short_12: OZLX7HA3YXIB · pith_short_16: OZLX7HA3YXIBF5GH · pith_short_8: OZLX7HA3

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/OZLX7HA3YXIBF5GH76HSDTRYMI \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 76577f9c1bc5d012f4c7ff8f21ce38621db5bd06b819327452f90da7b015b9b3

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "018b068a404b395d18a282dcfae45ad08bbeb8b5f35a74b5176d4b4f954a889b",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.DB",
    "submitted_at": "2026-03-29T06:49:12Z",
    "title_canon_sha256": "202db2d44f5a870a62c62d6b1f77396d4a4b6774a2324eb2d9b11247275e198d"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.16395",
    "kind": "arxiv",
    "version": 3
  }
}