leanVM-b

Warning: highly experimental.

Benchmarks

Machine: Mac M4 Max

XMSS aggregation

Our XMSS is specified in XMSS.pdf.

cargo run --release -- aggregate --xmss 900 --log-inv-rate 1 --repeat 3

aggregation, 900 XMSS signatures
  cycles (VM steps)           : 1,541,705 = 2^20.556
    details                   : DEREF 2^18.993 (33.9%)  SET 2^18.381 (22.1%)  MUL 2^18.204 (19.6%)  BLAKE2S 2^16.994 (8.5%)  XOR 2^16.97 (8.3%)  JUMP 2^16.843 (7.6%)  MEMORY 2^21.282  TOTAL_COMMITTED 2^26.195
  proof size                  : 295.2 KiB
  proving time                : 0.797 s ± 4.4%      peak memory 9.037 GiB
  per signature               : 1,128.79 signatures/s
  verifying                   : 0.0165 s

SPHINCS aggregation

Our SPHINCS is specified in SPHINCS.pdf.

cargo run --release -- aggregate --sphincs 245 --log-inv-rate 1 --repeat 3

aggregation, 245 SPHINCS signatures
  cycles (VM steps)           : 2,602,803 = 2^21.312
    details                   : DEREF 2^19.444 (27.4%)  XOR 2^19.238 (23.8%)  MUL 2^19.216 (23.4%)  SET 2^18.735 (16.8%)  BLAKE2S 2^16.995 (5.0%)  JUMP 2^16.545 (3.7%)  MEMORY 2^21.78  TOTAL_COMMITTED 2^26.664
  proof size                  : 321.5 KiB
  proving time                : 1.053 s ± 1.6%      peak memory 12.634 GiB
  per signature               : 232.661 signatures/s
  verifying                   : 0.0175 s

Recursion

cargo run --release -- recursion --n 2 --xmss-per-leaf 900 --log-inv-rate 2 --repeat 3

recursion 2→1, over leaves of 900 XMSS signatures
  cycles (VM steps)           : 711,107 = 2^19.44
    details                   : DEREF 2^17.852 (33.3%)  MUL 2^17.731 (30.6%)  XOR 2^17.366 (23.8%)  SET 2^15.258 (5.5%)  JUMP 2^14.804(4.0%)  BLAKE2S 2^14.302 (2.8%)  MEMORY 2^19.722  TOTAL_COMMITTED 2^24.656
  proof size                  : 205.8 KiB
  proving time                : 0.396 s ± 2.5%      peak memory 10.504 GiB
  verifying                   : 0.0158 s

Fibonacci

cargo run --release -- fibonacci --n 2000000 --log-inv-rate 1 --repeat 3

Fibonacci (in the exponent, i.e. modulo 2^64 - 1), N = 2,000,000
  cycles (VM steps)           : 2,127,880
    details                   : MUL 2^20.936 (98.7%)  DEREF 2^13.967 (0.8%)  SET 2^12.552 (0.3%)  JUMP 2^10.968 (0.1%)  XOR 2^10.966 (0.1%)  MEMORY 2^20.957  TOTAL_COMMITTED 2^25.263
  proof size                  : 285.4 KiB
  proving                     : 0.4 s ± 4.1%   5,320,879 cycles/s      peak memory 5.14 GiB
  verifying                   : 0.00294 s

Batch proving BLAKE2s

BENCH_REPEAT=3 BENCH_COOLDOWN=2 FLOCK_N_LOG=18 cargo test --release --package flock --test batch_proving_hashes -- hash_batch_prove_verify --exact --nocapture --include-ignored

Flock BLAKE2s batch proving, 262,144 compressions (2^18 slots)
  setup (preprocessing, excluded) :      0.0 ms
  witness-gen                     :     64.1 ms ± 8.1%   10.6%
  commit                          :    100.5 ms ± 0.7%   16.6%
  zerocheck                       :    243.6 ms ± 7.5%   40.3%
  lincheck                        :     20.7 ms ± 16.6%   3.4%
  pcs opening                     :    175.6 ms ± 2.4%   29.1%
  other                           :      0.0 ms           0.0%
  ------------------------------------------
  prove TOTAL (witness excluded)  :    540.5 ms ± 4.6%   89.4%
  verify                          :      1.9 ms
  throughput                      :        485,033 compressions/s ± 4.6%
  (~3322.1 XMSS/s equivalent at 146 compressions/signature)

Security

• 128-bit (LDR Johnson, no proximity gaps conjecture)

Snark machinery

• Binary field of 192 bits
• PCS: WHIR (aka Ligerito)
• Proving BLAKE2s by Flock
• RingSwitching, M3 arithmetisation, (and more) by Binius / Binius64 (see DP23 and DP24)