Performance

Performance Benchmarks

Performance data for cryptographic operations, AI inference, storage throughput, and browser capabilities across the Anticloud ecosystem.

Cryptographic Operations

flowchart LR
    subgraph Metrics[Crypto Performance]
        SHA3[<b>SHA3-256</b><br/>2.1 μs/op<br/>476K ops/s]
        ED[<b>Ed25519 Sign</b><br/>42 μs<br/>23.8K ops/s]
        EDV[<b>Ed25519 Verify</b><br/>78 μs<br/>12.8K ops/s]
        BLAKE[<b>BLAKE3</b><br/>0.8 μs/op<br/>1.25M ops/s]
        AIOSS[<b>.aioss Append</b><br/>89 μs<br/>11.2K ops/s]
    end
    style SHA3 fill:#5856d6,color:#fff
    style ED fill:#ff3b30,color:#fff
    style EDV fill:#ff9f0a,color:#fff
    style BLAKE fill:#34c759,color:#fff
    style AIOSS fill:#0071e3,color:#fff

Loading

Operation	Time	Throughput	Project	Hardware
SHA3-256 (1 KB)	2.1 μs	476,000 ops/s	Libern	AMD Ryzen 9 7950X
Ed25519 Sign (32 B)	42 μs	23,800 ops/s	Libern	AMD Ryzen 9 7950X
Ed25519 Verify (32 B)	78 μs	12,800 ops/s	Libern	AMD Ryzen 9 7950X
BLAKE3 (1 KB)	0.8 μs	1,250,000 ops/s	Kazcade	AMD Ryzen 9 7950X
.aioss Ledger Append	89 μs	11,200 ops/s	aioss-format	AMD Ryzen 9 7950X
AES-256-GCM (1 KB)	18 ns	55,000,000 ops/s	Libern	AES-NI hardware

AI Inference Performance

Model	Precision	Latency	Throughput	Project
Qwen2.5-VL-7B	FP16	320 ms	31 tok/s	Kathon (ad blocking)
Whisper small.en	FP16	180 ms	8.5x realtime	Inte11ect (transcription)
NLLB-200-600M	FP16	245 ms	42 tok/s	Inte11ect (translation)
Embedding (bge-small)	FP16	8 ms	125 queries/s	Kazcade (indexing)
Llama 3.2-3B	INT4	85 ms	94 tok/s	Anticode (local coding)

Ad Blocking Comparison

Engine	Precision	Recall	Latency	Memory
Kathon Vision-LLM	94.3%	91.7%	320 ms	4.2 GB VRAM
uBlock Origin (EasyList)	82.1%	79.4%	<1 ms	48 MB
AdGuard (Base Filter)	84.7%	81.2%	<1 ms	52 MB
Pi-hole (DNS-based)	67.3%	63.8%	<1 ms	system

Kathon's vision-LLM approach trades raw speed for significantly higher precision and recall by visually classifying page elements rather than relying on static filter lists.

Storage Performance

Operation	Latency	Throughput	Project
Vector Embedding (batch 64)	12 ms	5,300 docs/s	Kazcade
CRDT Sync (P2P, 10 peers)	45 ms	2,200 ops/s	Kathon ↔ Kazcade
Semantic Search (rank 10)	28 ms	35 queries/s	Kazcade
.aioss Audit Query (10K entries)	4 ms	250,000 entries/s	aioss-format
MFSO Identity Verify	67 ms	14,900 verifications/s	MFSO

Browser Performance

Metric	Kathon	Chromium	Firefox
Cold start	1.2 s	0.8 s	1.1 s
Tab switch	4 ms	8 ms	6 ms
Memory (10 tabs)	420 MB	580 MB	510 MB
Page load with ads	2.1 s	3.4 s	3.2 s
Page load (Kathon clean)	1.8 s	3.4 s	3.2 s

---

📖 Full docs: Docusaurus Tools · Home · Architecture · Projects · Roadmap · Glossary

.====================================================================.
!  Made in the UAE, Dubai #DubaiIt #Dubai #Dxb #SovereignAI          !
!  Made in The Emirates #Dubai_it                                    !
!                                                                    !
!  Lois-Kleinner Alpasan - The Anticloud 2026-                       !
!                                                                    !
!  0-1.gg ! GitHub ! LinkedIn ! DEV ! GH Pages                       !
!  HuggingFace ! Blog ! Tumblr ! Fandom ! Bluesky ! Mastodon          !
!  Zenodo ! Harvard Dataverse ! Internet Archive ! ORCID ! Figshare   !
!                                                                    !
!  Sovereign AI ! Local-First ! Privacy ! Zero Trust ! No Datacenter !
!  Air-Gapped ! Open Source ! Rust ! Hash Chain ! Single Binary      !
!  Offline LLM ! Crypto Ledger ! P2P ! Federated                     !
'===================================================================='

Lois-Kleinner Alpasan, 22, builds sovereign AI infrastructure and cryptographic audit systems. His work spans formats, proptech, and research platforms serving projects valued at over $1B combined, operating at the intersection of AI, media, and decentralized technology.

References:

1. Lois-Kleinner Zenodo: https://doi.org/10.5281/zenodo.20781790
2. Lois-Kleinner GitHub: https://github.com/kleinnner/Anticloud/tree/main/04-aioss-format
3. Lois-Kleinner Harvard DV: https://doi.org/10.7910/DVN/GDLO0L
4. Lois-Kleinner Internet Arc: https://archive.org/details/aioss-format
5. Lois-Kleinner ORCID: https://orcid.org/0009-0009-2233-6107
6. Lois-Kleinner DEV.to: https://dev.to/kleinner
7. Lois-Kleinner LinkedIn: https://linkedin.com/in/kleinner
8. Lois-Kleinner HuggingFace: https://huggingface.co/Anticloud
9. Lois-Kleinner Tumblr: https://anticloud.tumblr.com
10. Lois-Kleinner Mastodon: https://mastodon.social/@kleinner
11. Lois-Kleinner Bluesky: https://bsky.app/profile/kleinner.bsky.social
12. 0-1.gg: https://0-1.gg
13. Lois-Kleinner Figshare: https://figshare.com/authors/Lois-Kleinner_Alpasan/20849885