Performance measurements › arm_1 › kex-08 · system: x86_1 · arm_1

kex-08 NIIKE — performance on AArch64 (system arm_1)

Independent measurement following the structure of the NICCS ARM self-assessment guide, §3.5 (1)–(7). Not a submitter self-assessment.

1. Basic information

Category: public-key algorithm; function: key exchange
Algorithm: NIIKE
Implementation versions measured: reference
Parameter sets: NIIKE-lv128, NIIKE-lv256
Security evaluation: kex-08 report
Measurement method: arm_1 method
Submission: NICCS page

2. Assessment environment

item	value
processor	Qualcomm Oryon (CPU 2, one core)
machine	ASUS Vivobook S 15
clock	fixed 2.71 GHz (governor performance, minimum = maximum), boost off, SMT none
memory	30562 MiB
OS / kernel	Ubuntu 26.04.1 LTS / 7.0.0-34-generic
compiler / build tool	gcc (Ubuntu 15.2.0-16ubuntu1) 15.2.0 / cmake version 4.2.3
campaign start / end (UTC)	2026-09-28T15:05:46 / 2026-09-29T08:43:52

3. Functional testing (KAT)

Each library was checked against the submitted KAT vectors (SHA-256 manifest kex-08/kat.sha256 in the harness) before timing.

instance	build flags	KAT
`NIIKE-lv128`	guide	PASS
`NIIKE-lv256`	guide	PASS

4. Performance

instance	operation	mean cycles	mean time	ops/s	median trial	n (trials × iters)
`NIIKE-lv128`	exchange	53.18 G	19.7 s	0.0507	19.7 s	15 (5 × 3)
`NIIKE-lv128`	init_a	14.53 G	5.39 s	0.185	5.38 s	15 (5 × 3)
`NIIKE-lv128`	init_b	14.68 G	5.45 s	0.184	5.41 s	15 (5 × 3)
`NIIKE-lv128`	derive_a	12.06 G	4.47 s	0.223	4.47 s	15 (5 × 3)
`NIIKE-lv128`	derive_b	12.15 G	4.51 s	0.222	4.49 s	15 (5 × 3)
`NIIKE-lv256`	exchange	848.72 G	315 s	0.00317	315 s	1 (1 × 1)
`NIIKE-lv256`	init_a	222.60 G	83 s	0.012	83 s	1 (1 × 1)
`NIIKE-lv256`	init_b	222.69 G	83.1 s	0.012	83.1 s	1 (1 × 1)
`NIIKE-lv256`	derive_a	201.98 G	75 s	0.0133	75 s	1 (1 × 1)
`NIIKE-lv256`	derive_b	202.12 G	75 s	0.0133	75 s	1 (1 × 1)

5. Resource consumption

Static memory is approximated by the library's loadable ELF segments; peak memory is the process high-water mark (VmHWM) including the benchmark driver and libc, so both are upper-bound proxies rather than isolated algorithm memory.

instance	operation	static: ELF image (bytes)	baseline RSS	peak RSS
`NIIKE-lv128`	exchange	692528	2580 KiB	2652 KiB
`NIIKE-lv128`	init_a	692528	2580 KiB	2652 KiB
`NIIKE-lv128`	init_b	692528	2580 KiB	4580 KiB
`NIIKE-lv128`	derive_a	692528	2576 KiB	2648 KiB
`NIIKE-lv128`	derive_b	692528	2580 KiB	2652 KiB
`NIIKE-lv256`	exchange	997992	4820 KiB	4888 KiB
`NIIKE-lv256`	init_a	997992	4308 KiB	6356 KiB
`NIIKE-lv256`	init_b	997992	4248 KiB	6360 KiB
`NIIKE-lv256`	derive_a	997992	4820 KiB	4888 KiB
`NIIKE-lv256`	derive_b	997992	4632 KiB	6680 KiB

6. Transmission and storage overhead

Bandwidth counts all specified protocol messages and each required public key once. Public keys are transmitted bytes too. Certificates and transport framing are excluded. The published raw timing records are unchanged.

NIIKE sends no separate protocol message but requires both public keys. The specification lists 4,030/8,943-byte keys, while the submitted external encodings are 4,160/8,960 bytes; this table uses the latter.

instance	passes	messages (bytes; raw API)	protocol-message bytes	public key A / B	bandwidth (bytes)	long-term sk (API cap)	shared secret
`NIIKE-lv128`	0	0 / 0 / 0	0	4160 / 4160	8320	196	64
`NIIKE-lv256`	0	0 / 0 / 0	0	8960 / 8960	17920	336	128

Symmetric primitives

Survey verdict: bypass — no hash or KDF at all: raw j-invariant (known kex-08-1)

Share of each operation spent in the ICCS placeholder hash functions and in the ICCS DRNG (reference build, measured in the same run with link-time wrappers):

instance	operation	pseudohash/XOF/sm3	DRNG	calls per operation
`NIIKE-lv128`	exchange	0.0%	0.0%	drng 174
`NIIKE-lv256`	exchange	0.0%	0.0%	drng 296

7. Raw evidence index

Paths are relative to performance/data/arm_1/ in the harness; each JSON file records its commands, environment and trials.

instance	item	file
`NIIKE-lv128`	KAT log (sha256 `9f077a258eb6d903…`)	`kat/kex-08/NIIKE-lv128.log`
`NIIKE-lv128`	timing derive_a	`records/kex-08/NIIKE-lv128__derive_a.json`
`NIIKE-lv128`	timing derive_b	`records/kex-08/NIIKE-lv128__derive_b.json`
`NIIKE-lv128`	timing exchange	`records/kex-08/NIIKE-lv128__exchange.json`
`NIIKE-lv128`	timing init_a	`records/kex-08/NIIKE-lv128__init_a.json`
`NIIKE-lv128`	timing init_b	`records/kex-08/NIIKE-lv128__init_b.json`
`NIIKE-lv128`	hash profile exchange	`profile/kex-08/NIIKE-lv128__exchange.json`
`NIIKE-lv256`	KAT log (sha256 `de0f3d574c7621e6…`)	`kat/kex-08/NIIKE-lv256.log`
`NIIKE-lv256`	timing derive_a	`records/kex-08/NIIKE-lv256__derive_a.json`
`NIIKE-lv256`	timing derive_b	`records/kex-08/NIIKE-lv256__derive_b.json`
`NIIKE-lv256`	timing exchange	`records/kex-08/NIIKE-lv256__exchange.json`
`NIIKE-lv256`	timing init_a	`records/kex-08/NIIKE-lv256__init_a.json`
`NIIKE-lv256`	timing init_b	`records/kex-08/NIIKE-lv256__init_b.json`
`NIIKE-lv256`	hash profile exchange	`profile/kex-08/NIIKE-lv256__exchange.json`

Scripts: performance/campaign.py, performance/ngcc_perf.c, performance/hashprof/ in the harness.