Pith Number

pith:DXFDO6JM

pith:2026:DXFDO6JMX7ZMDQ4BVRGCXKJFQB

not attested not anchored not stored refs pending

SplitZip: Ultra Fast Lossless KV Compression for Disaggregated LLM Serving

Siddharth Joshi, Yipin Guo

SplitZip achieves over 600 GB/s lossless KV cache compression on GPUs by encoding frequent exponents with fixed-length codes and routing rare ones through a sparse escape stream.

arxiv:2605.01708 v3 · 2026-05-03 · cs.DC · cs.AI · cs.LG

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{DXFDO6JMX7ZMDQ4BVRGCXKJFQB}

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

SplitZip achieves 613.3 GB/s compression throughput and 2181.8 GB/s decompression throughput on real BF16 activation tensors, providing up to 1.32× speedup for BF16 KV cache transfer, 1.30× speedup for TTFT, and 1.23× increase on Request Throughput.

C2weakest assumption

The assumption that an offline calibrated top-16 exponent codebook will effectively capture the distribution of exponents in online KV activations during prefill without significant loss in compression ratio or speed, and that the method integrates seamlessly into existing serving frameworks.

C3one line summary

SplitZip is a new GPU-friendly lossless compressor for KV cache tensors that exploits exponent redundancy to achieve over 600 GB/s compression throughput and up to 1.32x faster transfers in disaggregated LLM serving.

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-06-25T00:18:14.094134Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

1dca37792cbff2c1c381ac4c2ba925805b9af493c0386e681a0230b4b868b621

Aliases

arxiv: 2605.01708 · arxiv_version: 2605.01708v3 · doi: 10.48550/arxiv.2605.01708 · pith_short_12: DXFDO6JMX7ZM · pith_short_16: DXFDO6JMX7ZMDQ4B · pith_short_8: DXFDO6JM

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/DXFDO6JMX7ZMDQ4BVRGCXKJFQB \
  | 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: 1dca37792cbff2c1c381ac4c2ba925805b9af493c0386e681a0230b4b868b621

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "f155781a755c59160d66c77c6c53ce1bb7a5bcaa772ab84b0ab7303920158ce4",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.LG"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.DC",
    "submitted_at": "2026-05-03T04:22:51Z",
    "title_canon_sha256": "a600cd2c62402b48a1aef292d2a03cb0951e2a728aac1f0c8a528c9756aebb87"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.01708",
    "kind": "arxiv",
    "version": 3
  }
}