High-Precision Exact FHE Made Simple, General, and Fast¶
Authors: Peikert, Zarchy, Zyskind
Year: 2026
Venue: CRYPTO 2026
type: scheme collection: schemes status: auto
Links: ePrint 2025/2321
Visual lesson: High-precision exact FHE · Scheme table row: High-precision exact FHE
Summary¶
Auto-extracted summary (unreviewed)
The paper gives a simple and general technique for exact FHE over high-precision plaintext rings Z_p, in which the fresh-ciphertext error rate and the error growth under multiplication match those of classic schemes at exponentially smaller precision, so runtimes scale only quasi-linearly (rather than quadratically) in log p. Unlike prior work it supports any integer modulus and any underlying ring, including plain LWE with no ring structure, and remains compatible with packing, SIMD and bootstrapping; a preliminary single-threaded implementation performs homomorphic Z_{2^64} multiplication in tens of milliseconds and gains a four- to five-fold increase in multiplicative depth over classic FHE.
Abstract¶
Many important applications of fully homomorphic encryption (FHE) require arithmetic on highprecision plaintexts, e.g., from the ring Zp for a huge prime or power-of-two modulus p. The classic FHE schemes are poorly suited to this, because the inverse error rate of fresh ciphertexts, and the error growth under homomorphic multiplication, are both larger than p, which results in large and inefficient parameters. While there are now several works addressing this problem, the landscape for exact (as opposed to approximate) FHE is highly fragmented: known solutions either work only for certain rare plaintext moduli having very special forms (sometimes using non-standard ciphertext rings that lack other important features for FHE), or have quite complicated and high-latency constructions. This work gives a very simple, general, and efficient technique for high-precision exact FHE, in which the error rates and growth match those of classic schemes for exponentially smaller precision. The runtimes can scale only quasi-linearly (versus quadratically for classic schemes) with the plaintext precision log p, and are fast in practice. Also in contrast to all prior works, our technique works for any integer modulus and over any underlying (number) ring—or even with no structured ring at all, making it the first solution that can be based on plain LWE. Moreover, it is fully compatible with prior FHE techniques for fast ring arithmetic, plaintext packing and SIMD operations, bootstrapping, etc. For typical parameters and security levels, our (preliminary, unoptimized, single-threaded) implementation does homomorphic Z264-multiplication in just tens of milliseconds, and obtains a four- to five-fold increase in multiplicative depth versus classic FHE schemes.
Source: pdf
Extracted details¶
| Field | Value | Source | Confidence |
|---|---|---|---|
| contribution | general technique for high-precision exact FHE over any integer modulus and any (number) ring, including plain LWE | llm | 0.70 |
| contribution | error rate and growth matching classic schemes at exponentially smaller precision; quasi-linear runtime in log p | llm | 0.70 |
| contribution | compatibility with existing packing, SIMD and bootstrapping techniques | llm | 0.70 |
| contribution | preliminary implementation with benchmarks | llm | 0.70 |
| schemes | high-precision exact FHE (proposed) | llm | 0.70 |
| schemes | LWE | llm | 0.70 |
| schemes | Ring-LWE | llm | 0.70 |
| tasks | high-precision exact integer arithmetic on encrypted data | llm | 0.70 |
| threat model | untrusted server computing on client's encrypted data (client-server) | llm | 0.70 |
| interaction | non-interactive FHE | llm | 0.70 |
| limitations | implementation is preliminary, unoptimized and single-threaded | llm | 0.70 |
| llm relevance | low | llm | 0.70 |
| key result | homomorphic Z_{2^64} multiplication latency: tens of milliseconds (preliminary, unoptimized, single-threaded implementation at typical parameters) | llm | 0.70 |
| key result | multiplicative depth increase: 4-5x (vs classic FHE schemes at the same parameters) | llm | 0.70 |
Scheme card: High-precision exact FHE¶
| Group | 2. Exact word-wise FHE (BGV / BFV family) |
| Kind | Scheme |
| Hardness assumption | (R)LWE |
| Plaintext type | Exact arithmetic mod huge p or 2^k |
| SIMD packing | Yes |
| Bootstrapping | Yes |
| Post-quantum | Yes |
| Status | Active SOTA (emerging) |
| Fit for LLM inference | Exact large-modulus arithmetic. Could replace CKKS where exact fixed-point semantics matter. |
Keyword mentions¶
Counts of vocabulary terms in the full text (first page of occurrence in brackets).
- scheme: BFV 122 [p.2], GSW 30 [p.2], CKKS 23 [p.2], TFHE 21 [p.2], BGV 19 [p.2], FHEW 11 [p.2], NTRU 3 [p.11]
- library: TFHE-rs 2 [p.6], HElib 1 [p.40], Zama 1 [p.6]
- hardware: GPU 3 [p.4], NVIDIA A100 1 [p.6], AMD EPYC 1 [p.6]
- technique: Bootstrapping 100 [p.1], Key switching 22 [p.7], SIMD packing 12 [p.1], Programmable bootstrapping 11 [p.10], NTT 5 [p.11], Modulus switching 3 [p.6], Rotations 2 [p.37], RNS 2 [p.52]
- assumption: Lattice-based 90 [p.4], LWE 78 [p.1], RLWE 8 [p.13], IND-CPA 4 [p.46]
- task: Language modeling / generation 12 [p.14]
- topic: Homomorphic encryption 72 [p.1]
Files¶
| Path (under papers/) | Collection | Pages | Size | SHA-256 | Note |
|---|---|---|---|---|---|
references/schemes/s18-0_high-precision-exact-fhe-made-simple-general-and-fast.pdf |
schemes | 52 | 685.9 KB | a5d73a31df10 |
Database key: high-precision-exact-fhe-made-simple-general-and-fast · created 2026-11-05 · updated 2026-11-05 · edit with python3 tools/paperdb.py set high-precision-exact-fhe-made-simple-general-and-fast COLUMN VALUE