Simplify decompose C reference to a single high multiplication by mkannwischer · Pull Request #1177 · pq-code-package/mldsa-native

mkannwischer · 2026-06-14T03:25:07Z

Replace the two-step Barrett division (ceil(a/128) then Barrett-divide
by 2GAMMA2/128) with a direct high multiplication by floor(2^N /
2GAMMA2), mirroring the AArch64 backend. For ML-DSA-44 this is
(a * 1477838209 + 2^47) >> 48; for ML-DSA-65/87 it is
(a * 1074791425 + 2^48) >> 49. Both constants strictly
under-approximate 1/(2*GAMMA2), so half-points round down, matching
the original round-half-down semantics, and the result is exact for
all 0 <= a < Q.

Update the Isabelle attribution in compress/ML-DSA_Compress.thy and
neon_ntt/Barrett_Division_Even.thy.

Resolves Consider simplifying decompose C implementation #652

oqs-bot · 2026-06-14T03:47:22Z

CBMC Results (ML-DSA-65, REDUCE-RAM)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1582s	1504s	+5.2%
`mld_invntt_layer`	✅	178s	163s	+9%
`poly_pointwise_montgomery_c`	✅	143s	126s	+13%
`rej_uniform_native`	✅	129s	119s	+8%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	86s	81s	+6%
`mld_ct_memcmp`	✅	71s	62s	+15%
`mld_ntt_layer`	✅	46s	42s	+10%
`fqmul`	✅	45s	37s	+22%
`polyveck_chknorm`	✅	40s	39s	+3%
`keccakf1600x4_permute_native`	✅	24s	22s	+9%
`mld_attempt_signature_generation`	✅	23s	24s	-4%
`mld_ntt_butterfly_block`	✅	23s	21s	+10%
`poly_chknorm_c`	✅	19s	21s	-10%
`polyt0_unpack`	✅	19s	17s	+12%
`sign_verify_internal`	✅	19s	13s	+46%
`polyveck_decompose`	✅	17s	20s	-15%
`rej_uniform_c`	✅	16s	17s	-6%
`mld_check_pct`	✅	15s	13s	+15%
`polyvecl_chknorm`	✅	15s	13s	+15%
`poly_uniform_eta_4x`	✅	12s	11s	+9%
`keccak_absorb_once_x4`	✅	11s	10s	+10%
`poly_add`	✅	11s	12s	-8%
`rej_uniform`	✅	10s	8s	+25%
`compute_pack_t0_t1`	✅	9s	8s	+12%
`poly_invntt_tomont_c`	✅	9s	6s	+50%
`polyvec_matrix_pointwise_montgomery_row`	✅	9s	7s	+29%
`pointwise_acc_native_x86_64`	✅	8s	6s	+33%
`sign`	✅	8s	8s	+0%
`mld_compute_pack_z`	✅	7s	6s	+17%
`poly_ntt_c`	✅	7s	4s	+75%
`polyeta_unpack`	✅	7s	4s	+75%
`polyveck_caddq`	✅	7s	9s	-22%
`polyveck_reduce`	✅	7s	4s	+75%
`polyvecl_pointwise_acc_montgomery_native`	✅	7s	3s	+133%
`keccak_absorb`	✅	6s	7s	-14%
`mld_keccakf1600_permute_c`	✅	6s	7s	-14%
`nttunpack_native_x86_64`	✅	6s	2s	+200%
`pointwise_acc_native_aarch64`	✅	6s	6s	+0%
`polyt0_pack`	✅	6s	4s	+50%
`polyvecl_ntt`	✅	6s	7s	-14%
`sign_pk_from_sk`	✅	6s	6s	+0%
`pack_sk_rho_key_tr_s2`	✅	5s	4s	+25%
`poly_caddq_c`	✅	5s	4s	+25%
`poly_challenge`	✅	5s	3s	+67%
`poly_shiftl`	✅	5s	6s	-17%
`poly_uniform`	✅	5s	5s	+0%
`polyveck_invntt_tomont`	✅	5s	6s	-17%
`polyvecl_pack_eta`	✅	5s	4s	+25%
`polyz_unpack_c`	✅	5s	6s	-17%
`sign_keypair`	✅	5s	4s	+25%
`sign_keypair_internal`	✅	5s	3s	+67%
`sign_signature_extmu`	✅	5s	3s	+67%
`sign_signature_pre_hash_internal`	✅	5s	4s	+25%
`intt_native_x86_64`	✅	4s	4s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	3s	+33%
`keccak_squeezeblocks_x4`	✅	4s	5s	-20%
`keccakf1600x4_extract_bytes`	✅	4s	5s	-20%
`mld_ct_cmask_nonzero_u32`	✅	4s	3s	+33%
`mld_prepare_domain_separation_prefix`	✅	4s	3s	+33%
`mld_sample_s1_s2`	✅	4s	4s	+0%
`mld_value_barrier_u32`	✅	4s	2s	+100%
`ntt_native_aarch64`	✅	4s	3s	+33%
`pack_sig_h`	✅	4s	5s	-20%
`pack_sk_s1`	✅	4s	4s	+0%
`poly_caddq`	✅	4s	2s	+100%
`poly_chknorm_native_x86_64`	✅	4s	2s	+100%
`poly_decompose_c`	✅	4s	2s	+100%
`poly_decompose_native`	✅	4s	4s	+0%
`poly_ntt_native`	✅	4s	3s	+33%
`poly_power2round`	✅	4s	7s	-43%
`poly_reduce`	✅	4s	5s	-20%
`poly_sub`	✅	4s	3s	+33%
`poly_uniform_gamma1`	✅	4s	2s	+100%
`poly_uniform_gamma1_4x`	✅	4s	2s	+100%
`polyeta_pack`	✅	4s	6s	-33%
`polyveck_pack_w1`	✅	4s	6s	-33%
`polyveck_unpack_eta`	✅	4s	4s	+0%
`polyvecl_uniform_gamma1`	✅	4s	2s	+100%
`polyvecl_uniform_gamma1_serial`	✅	4s	3s	+33%
`polyw1_pack_32`	✅	4s	4s	+0%
`polyw1_pack_88`	✅	4s	4s	+0%
`polyz_pack`	✅	4s	3s	+33%
`polyz_unpack_native`	✅	4s	2s	+100%
`rej_eta_native`	✅	4s	2s	+100%
`rej_uniform_eta_native_aarch64`	✅	4s	4s	+0%
`sign_open`	✅	4s	2s	+100%
`sign_signature_internal`	✅	4s	3s	+33%
`sign_signature_pre_hash_shake256`	✅	4s	6s	-33%
`sign_verify_extmu`	✅	4s	3s	+33%
`sk_s2hat_get_poly`	✅	4s	3s	+33%
`unpack_sk_s2hat`	✅	4s	3s	+33%
`yvec_init`	✅	4s	3s	+33%
`decompose`	✅	3s	2s	+50%
`intt_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	3s	3s	+0%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`keccak_init`	✅	3s	3s	+0%
`keccakf1600_extract_bytes (big endian)`	✅	3s	3s	+0%
`keccakf1600_xor_bytes (big endian)`	✅	3s	2s	+50%
`keccakf1600x4_permute`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes`	✅	3s	3s	+0%
`mld_ct_cmask_nonzero_u8`	✅	3s	5s	-40%
`mld_h`	✅	3s	4s	-25%
`mld_keccakf1600x4_extract_bytes_c`	✅	3s	2s	+50%
`mld_keccakf1600x4_xor_bytes_c`	✅	3s	3s	+0%
`mld_polymat_expand_entry`	✅	3s	3s	+0%
`mld_sample_s1_s2_serial`	✅	3s	2s	+50%
`mld_value_barrier_i64`	✅	3s	4s	-25%
`mld_value_barrier_u8`	✅	3s	2s	+50%
`montgomery_reduce`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	3s	+0%
`pack_sig_c`	✅	3s	5s	-40%
`pointwise_native_aarch64`	✅	3s	4s	-25%
`pointwise_native_x86_64`	✅	3s	3s	+0%
`poly_caddq_native`	✅	3s	3s	+0%
`poly_caddq_native_aarch64`	✅	3s	2s	+50%
`poly_caddq_native_x86_64`	✅	3s	2s	+50%
`poly_chknorm_native`	✅	3s	3s	+0%
`poly_decompose_32_native_aarch64`	✅	3s	3s	+0%
`poly_decompose_88_native_aarch64`	✅	3s	1s	+200%
`poly_invntt_tomont_native`	✅	3s	4s	-25%
`poly_ntt`	✅	3s	2s	+50%
`poly_pointwise_montgomery`	✅	3s	3s	+0%
`poly_uniform_eta`	✅	3s	3s	+0%
`poly_use_hint_native_aarch64`	✅	3s	2s	+50%
`polyt1_pack`	✅	3s	2s	+50%
`polyt1_unpack`	✅	3s	4s	-25%
`polyvecl_unpack_eta`	✅	3s	2s	+50%
`polyvecl_unpack_z`	✅	3s	2s	+50%
`polyw1_pack`	✅	3s	7s	-57%
`polyz_unpack_17_native_aarch64`	✅	3s	1s	+200%
`polyz_unpack_19_native_aarch64`	✅	3s	2s	+50%
`rej_eta`	✅	3s	3s	+0%
`rej_uniform_native_aarch64`	✅	3s	2s	+50%
`shake128_absorb`	✅	3s	4s	-25%
`shake128_finalize`	✅	3s	2s	+50%
`shake128_squeeze`	✅	3s	4s	-25%
`shake256_init`	✅	3s	2s	+50%
`shake256_squeeze`	✅	3s	3s	+0%
`shake256x4_squeezeblocks`	✅	3s	3s	+0%
`sig_unpack_hints`	✅	3s	3s	+0%
`sign_verify`	✅	3s	3s	+0%
`sign_verify_pre_hash_internal`	✅	3s	5s	-40%
`sign_verify_pre_hash_shake256`	✅	3s	5s	-40%
`sk_s1hat_get_poly`	✅	3s	1s	+200%
`sys_check_capability`	✅	3s	4s	-25%
`unpack_sk_s1hat`	✅	3s	3s	+0%
`yvec_get_poly`	✅	3s	1s	+200%
`caddq`	✅	2s	4s	-50%
`fqscale`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	2s	+0%
`keccak_finalize`	✅	2s	5s	-60%
`keccak_squeeze`	✅	2s	2s	+0%
`keccakf1600_permute`	✅	2s	2s	+0%
`keccakf1600_xor_bytes`	✅	2s	1s	+100%
`keccakf1600x4_xor_bytes_native`	✅	2s	2s	+0%
`make_hint`	✅	2s	2s	+0%
`mld_ct_abs_i32`	✅	2s	3s	-33%
`mld_ct_get_optblocker_i64`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u32`	✅	2s	3s	-33%
`mld_ct_get_optblocker_u8`	✅	2s	3s	-33%
`mld_ct_sel_int32`	✅	2s	1s	+100%
`mld_keccakf1600_extract_bytes`	✅	2s	5s	-60%
`pack_sig_z`	✅	2s	5s	-60%
`poly_chknorm`	✅	2s	1s	+100%
`poly_chknorm_native_aarch64`	✅	2s	1s	+100%
`poly_decompose`	✅	2s	3s	-33%
`poly_permute_bitrev_to_custom_optional`	✅	2s	2s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	2s	2s	+0%
`poly_pointwise_montgomery_native`	✅	2s	4s	-50%
`poly_use_hint`	✅	2s	2s	+0%
`poly_use_hint_c`	✅	2s	5s	-60%
`poly_use_hint_native`	✅	2s	2s	+0%
`polyvec_matrix_expand`	✅	2s	4s	-50%
`polyvec_matrix_expand_serial`	✅	2s	2s	+0%
`polyveck_ntt`	✅	2s	2s	+0%
`polyveck_pack_eta`	✅	2s	3s	-33%
`polyvecl_pointwise_acc_montgomery`	✅	2s	2s	+0%
`polyvecl_pointwise_acc_montgomery_c`	✅	2s	4s	-50%
`polyz_unpack`	✅	2s	3s	-33%
`power2round`	✅	2s	3s	-33%
`rej_eta_c`	✅	2s	2s	+0%
`shake128x4_absorb_once`	✅	2s	6s	-67%
`shake128x4_squeezeblocks`	✅	2s	3s	-33%
`shake256_release`	✅	2s	3s	-33%
`shake256x4_absorb_once`	✅	2s	4s	-50%
`sign_signature`	✅	2s	4s	-50%
`sk_t0hat_get_poly`	✅	2s	2s	+0%
`unpack_pk_t1`	✅	2s	4s	-50%
`unpack_sk`	✅	2s	2s	+0%
`keccak_f1600_x1_native_aarch64`	✅	1s	3s	-67%
`keccak_f1600_x1_native_aarch64_v84a`	✅	1s	3s	-67%
`keccakf1600_permute_native`	✅	1s	1s	+0%
`keccakf1600x4_extract_bytes_native`	✅	1s	3s	-67%
`mld_ct_cmask_neg_i32`	✅	1s	2s	-50%
`poly_invntt_tomont`	✅	1s	2s	-50%
`poly_uniform_4x`	✅	1s	2s	-50%
`reduce32`	✅	1s	4s	-75%
`shake128_init`	✅	1s	4s	-75%
`shake128_release`	✅	1s	1s	+0%
`shake256`	✅	1s	3s	-67%
`shake256_absorb`	✅	1s	5s	-80%
`shake256_finalize`	✅	1s	2s	-50%
`unpack_sk_t0hat`	✅	1s	3s	-67%
`use_hint`	✅	1s	3s	-67%

oqs-bot · 2026-06-14T03:47:39Z

CBMC Results (ML-DSA-44, REDUCE-RAM)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1560s	1453s	+7.4%
`mld_invntt_layer`	✅	169s	147s	+15%
`poly_pointwise_montgomery_c`	✅	134s	125s	+7%
`rej_uniform_native`	✅	125s	119s	+5%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	113s	112s	+1%
`mld_ct_memcmp`	✅	70s	65s	+8%
`mld_ntt_layer`	✅	44s	42s	+5%
`fqmul`	✅	40s	40s	+0%
`mld_attempt_signature_generation`	✅	28s	25s	+12%
`sign_verify_internal`	✅	25s	20s	+25%
`keccakf1600x4_permute_native`	✅	22s	23s	-4%
`mld_ntt_butterfly_block`	✅	22s	21s	+5%
`poly_chknorm_c`	✅	18s	18s	+0%
`polyt0_unpack`	✅	17s	16s	+6%
`mld_check_pct`	✅	16s	12s	+33%
`poly_uniform_eta_4x`	✅	16s	14s	+14%
`rej_uniform_c`	✅	15s	15s	+0%
`polyeta_unpack`	✅	14s	16s	-12%
`polyz_unpack_c`	✅	14s	11s	+27%
`poly_add`	✅	12s	9s	+33%
`compute_pack_t0_t1`	✅	9s	9s	+0%
`keccak_absorb_once_x4`	✅	8s	10s	-20%
`poly_invntt_tomont_c`	✅	8s	8s	+0%
`polyveck_chknorm`	✅	8s	6s	+33%
`rej_uniform`	✅	8s	10s	-20%
`keccak_absorb`	✅	7s	6s	+17%
`mld_h`	✅	7s	3s	+133%
`mld_keccakf1600_permute_c`	✅	7s	8s	-12%
`poly_decompose_c`	✅	7s	7s	+0%
`poly_reduce`	✅	7s	4s	+75%
`polyvec_matrix_pointwise_montgomery_row`	✅	7s	8s	-12%
`sign_verify_pre_hash_internal`	✅	7s	3s	+133%
`poly_ntt`	✅	6s	3s	+100%
`poly_ntt_native`	✅	6s	2s	+200%
`polyveck_decompose`	✅	6s	6s	+0%
`polyvecl_chknorm`	✅	6s	9s	-33%
`sign`	✅	6s	7s	-14%
`sign_pk_from_sk`	✅	6s	7s	-14%
`sign_signature`	✅	6s	3s	+100%
`keccak_squeezeblocks_x4`	✅	5s	4s	+25%
`mld_compute_pack_z`	✅	5s	5s	+0%
`mld_ct_sel_int32`	✅	5s	3s	+67%
`mld_keccakf1600x4_extract_bytes_c`	✅	5s	1s	+400%
`montgomery_reduce`	✅	5s	2s	+150%
`pointwise_acc_native_aarch64`	✅	5s	4s	+25%
`pointwise_native_x86_64`	✅	5s	3s	+67%
`poly_caddq_c`	✅	5s	4s	+25%
`poly_challenge`	✅	5s	5s	+0%
`poly_pointwise_montgomery`	✅	5s	3s	+67%
`poly_power2round`	✅	5s	6s	-17%
`poly_uniform_gamma1`	✅	5s	3s	+67%
`polyt0_pack`	✅	5s	4s	+25%
`polyveck_reduce`	✅	5s	6s	-17%
`polyvecl_unpack_eta`	✅	5s	2s	+150%
`sign_open`	✅	5s	5s	+0%
`sign_signature_internal`	✅	5s	4s	+25%
`keccak_f1600_x1_native_aarch64`	✅	4s	3s	+33%
`make_hint`	✅	4s	4s	+0%
`mld_ct_cmask_nonzero_u8`	✅	4s	1s	+300%
`mld_polymat_expand_entry`	✅	4s	2s	+100%
`mld_sample_s1_s2`	✅	4s	8s	-50%
`mld_value_barrier_i64`	✅	4s	2s	+100%
`nttunpack_native_x86_64`	✅	4s	2s	+100%
`pack_sig_c`	✅	4s	3s	+33%
`pack_sig_h`	✅	4s	5s	-20%
`pack_sig_z`	✅	4s	3s	+33%
`pointwise_acc_native_x86_64`	✅	4s	5s	-20%
`poly_caddq`	✅	4s	3s	+33%
`poly_chknorm_native`	✅	4s	4s	+0%
`poly_decompose_32_native_aarch64`	✅	4s	1s	+300%
`poly_invntt_tomont_native`	✅	4s	2s	+100%
`poly_ntt_c`	✅	4s	3s	+33%
`poly_shiftl`	✅	4s	4s	+0%
`poly_use_hint_native_aarch64`	✅	4s	4s	+0%
`polyt1_unpack`	✅	4s	2s	+100%
`polyveck_caddq`	✅	4s	6s	-33%
`polyveck_ntt`	✅	4s	4s	+0%
`polyveck_pack_eta`	✅	4s	3s	+33%
`polyveck_pack_w1`	✅	4s	2s	+100%
`polyvecl_ntt`	✅	4s	3s	+33%
`polyvecl_pack_eta`	✅	4s	2s	+100%
`polyvecl_pointwise_acc_montgomery_native`	✅	4s	1s	+300%
`polyvecl_uniform_gamma1`	✅	4s	2s	+100%
`polyvecl_uniform_gamma1_serial`	✅	4s	4s	+0%
`polyz_unpack`	✅	4s	3s	+33%
`rej_eta`	✅	4s	4s	+0%
`rej_eta_native`	✅	4s	3s	+33%
`shake256_absorb`	✅	4s	3s	+33%
`shake256_release`	✅	4s	1s	+300%
`shake256x4_absorb_once`	✅	4s	2s	+100%
`sign_keypair`	✅	4s	4s	+0%
`sign_signature_extmu`	✅	4s	4s	+0%
`sign_signature_pre_hash_internal`	✅	4s	4s	+0%
`sk_t0hat_get_poly`	✅	4s	2s	+100%
`sys_check_capability`	✅	4s	3s	+33%
`unpack_pk_t1`	✅	4s	2s	+100%
`caddq`	✅	3s	1s	+200%
`decompose`	✅	3s	1s	+200%
`intt_native_aarch64`	✅	3s	2s	+50%
`intt_native_x86_64`	✅	3s	2s	+50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	3s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	2s	+50%
`keccak_init`	✅	3s	4s	-25%
`keccak_squeeze`	✅	3s	1s	+200%
`keccakf1600_xor_bytes (big endian)`	✅	3s	2s	+50%
`keccakf1600x4_permute`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes_native`	✅	3s	3s	+0%
`mld_ct_abs_i32`	✅	3s	2s	+50%
`mld_ct_cmask_neg_i32`	✅	3s	1s	+200%
`mld_ct_cmask_nonzero_u32`	✅	3s	1s	+200%
`mld_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mld_prepare_domain_separation_prefix`	✅	3s	4s	-25%
`mld_sample_s1_s2_serial`	✅	3s	3s	+0%
`ntt_native_aarch64`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	1s	+200%
`pack_sk_rho_key_tr_s2`	✅	3s	4s	-25%
`pack_sk_s1`	✅	3s	3s	+0%
`pointwise_native_aarch64`	✅	3s	3s	+0%
`poly_caddq_native`	✅	3s	2s	+50%
`poly_caddq_native_aarch64`	✅	3s	1s	+200%
`poly_chknorm_native_aarch64`	✅	3s	5s	-40%
`poly_chknorm_native_x86_64`	✅	3s	4s	-25%
`poly_decompose`	✅	3s	3s	+0%
`poly_decompose_native`	✅	3s	5s	-40%
`poly_permute_bitrev_to_custom_optional`	✅	3s	2s	+50%
`poly_permute_bitrev_to_custom_optional_native`	✅	3s	5s	-40%
`poly_sub`	✅	3s	4s	-25%
`poly_uniform`	✅	3s	5s	-40%
`poly_uniform_4x`	✅	3s	3s	+0%
`poly_uniform_eta`	✅	3s	5s	-40%
`poly_uniform_gamma1_4x`	✅	3s	3s	+0%
`polyeta_pack`	✅	3s	2s	+50%
`polyveck_invntt_tomont`	✅	3s	3s	+0%
`polyvecl_pointwise_acc_montgomery_c`	✅	3s	3s	+0%
`polyvecl_unpack_z`	✅	3s	2s	+50%
`polyw1_pack_32`	✅	3s	3s	+0%
`polyw1_pack_88`	✅	3s	3s	+0%
`polyz_pack`	✅	3s	2s	+50%
`polyz_unpack_17_native_aarch64`	✅	3s	2s	+50%
`power2round`	✅	3s	3s	+0%
`rej_eta_c`	✅	3s	3s	+0%
`rej_uniform_eta_native_aarch64`	✅	3s	2s	+50%
`shake128_init`	✅	3s	6s	-50%
`shake128_squeeze`	✅	3s	2s	+50%
`shake128x4_squeezeblocks`	✅	3s	3s	+0%
`shake256`	✅	3s	2s	+50%
`shake256_squeeze`	✅	3s	4s	-25%
`shake256x4_squeezeblocks`	✅	3s	1s	+200%
`sig_unpack_hints`	✅	3s	2s	+50%
`sign_keypair_internal`	✅	3s	3s	+0%
`sign_signature_pre_hash_shake256`	✅	3s	3s	+0%
`sign_verify`	✅	3s	3s	+0%
`sign_verify_extmu`	✅	3s	2s	+50%
`sign_verify_pre_hash_shake256`	✅	3s	4s	-25%
`sk_s2hat_get_poly`	✅	3s	4s	-25%
`unpack_sk_t0hat`	✅	3s	1s	+200%
`yvec_init`	✅	3s	4s	-25%
`fqscale`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	1s	+100%
`keccak_f1600_x4_native_avx2`	✅	2s	4s	-50%
`keccakf1600_permute`	✅	2s	4s	-50%
`keccakf1600_permute_native`	✅	2s	1s	+100%
`keccakf1600x4_extract_bytes`	✅	2s	4s	-50%
`keccakf1600x4_extract_bytes_native`	✅	2s	1s	+100%
`keccakf1600x4_xor_bytes`	✅	2s	2s	+0%
`mld_ct_get_optblocker_i64`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mld_keccakf1600x4_xor_bytes_c`	✅	2s	1s	+100%
`mld_value_barrier_u32`	✅	2s	1s	+100%
`mld_value_barrier_u8`	✅	2s	2s	+0%
`poly_chknorm`	✅	2s	2s	+0%
`poly_decompose_88_native_aarch64`	✅	2s	4s	-50%
`poly_invntt_tomont`	✅	2s	3s	-33%
`poly_pointwise_montgomery_native`	✅	2s	2s	+0%
`poly_use_hint`	✅	2s	4s	-50%
`poly_use_hint_c`	✅	2s	3s	-33%
`poly_use_hint_native`	✅	2s	4s	-50%
`polyt1_pack`	✅	2s	5s	-60%
`polyvec_matrix_expand`	✅	2s	4s	-50%
`polyvec_matrix_expand_serial`	✅	2s	5s	-60%
`polyveck_unpack_eta`	✅	2s	2s	+0%
`polyvecl_pointwise_acc_montgomery`	✅	2s	3s	-33%
`polyw1_pack`	✅	2s	1s	+100%
`polyz_unpack_19_native_aarch64`	✅	2s	2s	+0%
`polyz_unpack_native`	✅	2s	3s	-33%
`reduce32`	✅	2s	3s	-33%
`rej_uniform_native_aarch64`	✅	2s	5s	-60%
`shake128_finalize`	✅	2s	2s	+0%
`shake128_release`	✅	2s	2s	+0%
`shake128x4_absorb_once`	✅	2s	3s	-33%
`shake256_init`	✅	2s	5s	-60%
`unpack_sk_s1hat`	✅	2s	2s	+0%
`unpack_sk_s2hat`	✅	2s	3s	-33%
`yvec_get_poly`	✅	2s	1s	+100%
`keccak_finalize`	✅	1s	3s	-67%
`keccakf1600_extract_bytes (big endian)`	✅	1s	4s	-75%
`keccakf1600_xor_bytes`	✅	1s	3s	-67%
`mld_ct_get_optblocker_u8`	✅	1s	3s	-67%
`poly_caddq_native_x86_64`	✅	1s	1s	+0%
`shake128_absorb`	✅	1s	3s	-67%
`shake256_finalize`	✅	1s	2s	-50%
`sk_s1hat_get_poly`	✅	1s	3s	-67%
`unpack_sk`	✅	1s	2s	-50%
`use_hint`	✅	1s	1s	+0%

oqs-bot · 2026-06-14T03:48:02Z

CBMC Results (ML-DSA-87, REDUCE-RAM)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1599s	1624s	-1.5%
`mld_invntt_layer`	✅	171s	172s	-1%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	160s	155s	+3%
`poly_pointwise_montgomery_c`	✅	128s	136s	-6%
`rej_uniform_native`	✅	125s	125s	+0%
`mld_ct_memcmp`	✅	67s	65s	+3%
`mld_ntt_layer`	✅	43s	43s	+0%
`fqmul`	✅	40s	43s	-7%
`mld_attempt_signature_generation`	✅	33s	37s	-11%
`keccakf1600x4_permute_native`	✅	22s	26s	-15%
`mld_ntt_butterfly_block`	✅	22s	23s	-4%
`sign_verify_internal`	✅	21s	22s	-5%
`poly_chknorm_c`	✅	18s	18s	+0%
`rej_uniform_c`	✅	18s	16s	+12%
`polyveck_decompose`	✅	17s	18s	-6%
`mld_check_pct`	✅	16s	13s	+23%
`polyeta_unpack`	✅	16s	14s	+14%
`polyt0_unpack`	✅	16s	17s	-6%
`poly_uniform_eta_4x`	✅	14s	13s	+8%
`compute_pack_t0_t1`	✅	12s	11s	+9%
`poly_add`	✅	12s	11s	+9%
`poly_invntt_tomont_c`	✅	11s	9s	+22%
`polyvecl_chknorm`	✅	10s	10s	+0%
`keccak_absorb_once_x4`	✅	9s	10s	-10%
`polyvec_matrix_pointwise_montgomery_row`	✅	8s	8s	+0%
`polyvecl_ntt`	✅	8s	9s	-11%
`polyz_unpack_c`	✅	8s	6s	+33%
`rej_uniform`	✅	8s	8s	+0%
`mld_sample_s1_s2`	✅	7s	4s	+75%
`pointwise_acc_native_aarch64`	✅	7s	8s	-12%
`sign`	✅	7s	8s	-12%
`sign_signature_pre_hash_shake256`	✅	7s	5s	+40%
`sign_verify_pre_hash_internal`	✅	7s	4s	+75%
`keccak_absorb`	✅	6s	7s	-14%
`keccakf1600_permute_native`	✅	6s	3s	+100%
`mld_compute_pack_z`	✅	6s	3s	+100%
`mld_keccakf1600_permute_c`	✅	6s	6s	+0%
`ntt_native_aarch64`	✅	6s	8s	-25%
`pointwise_acc_native_x86_64`	✅	6s	6s	+0%
`poly_shiftl`	✅	6s	3s	+100%
`sign_signature_internal`	✅	6s	5s	+20%
`intt_native_aarch64`	✅	5s	3s	+67%
`keccak_squeezeblocks_x4`	✅	5s	4s	+25%
`keccakf1600x4_extract_bytes_native`	✅	5s	2s	+150%
`nttunpack_native_x86_64`	✅	5s	4s	+25%
`pack_sk_rho_key_tr_s2`	✅	5s	3s	+67%
`poly_challenge`	✅	5s	6s	-17%
`poly_chknorm_native_x86_64`	✅	5s	4s	+25%
`poly_invntt_tomont`	✅	5s	6s	-17%
`poly_power2round`	✅	5s	5s	+0%
`polyveck_caddq`	✅	5s	5s	+0%
`polyveck_invntt_tomont`	✅	5s	7s	-29%
`polyveck_ntt`	✅	5s	2s	+150%
`polyveck_reduce`	✅	5s	6s	-17%
`polyvecl_uniform_gamma1`	✅	5s	3s	+67%
`polyz_unpack_native`	✅	5s	3s	+67%
`rej_eta_c`	✅	5s	4s	+25%
`sk_t0hat_get_poly`	✅	5s	2s	+150%
`unpack_sk`	✅	5s	4s	+25%
`fqscale`	✅	4s	3s	+33%
`keccak_init`	✅	4s	3s	+33%
`mld_ct_cmask_nonzero_u32`	✅	4s	5s	-20%
`mld_sample_s1_s2_serial`	✅	4s	5s	-20%
`montgomery_reduce`	✅	4s	2s	+100%
`pack_sig_c`	✅	4s	3s	+33%
`poly_decompose`	✅	4s	2s	+100%
`poly_decompose_c`	✅	4s	5s	-20%
`poly_ntt_c`	✅	4s	3s	+33%
`poly_permute_bitrev_to_custom_optional_native`	✅	4s	4s	+0%
`poly_pointwise_montgomery`	✅	4s	3s	+33%
`poly_uniform`	✅	4s	4s	+0%
`poly_uniform_eta`	✅	4s	4s	+0%
`poly_use_hint_native_aarch64`	✅	4s	2s	+100%
`polyvecl_unpack_eta`	✅	4s	2s	+100%
`polyw1_pack`	✅	4s	2s	+100%
`polyz_pack`	✅	4s	1s	+300%
`polyz_unpack_17_native_aarch64`	✅	4s	2s	+100%
`polyz_unpack_19_native_aarch64`	✅	4s	3s	+33%
`rej_eta_native`	✅	4s	3s	+33%
`rej_uniform_eta_native_aarch64`	✅	4s	5s	-20%
`shake128_absorb`	✅	4s	2s	+100%
`shake128x4_absorb_once`	✅	4s	2s	+100%
`shake256_init`	✅	4s	4s	+0%
`sig_unpack_hints`	✅	4s	3s	+33%
`sign_pk_from_sk`	✅	4s	6s	-33%
`sign_verify`	✅	4s	6s	-33%
`sign_verify_extmu`	✅	4s	3s	+33%
`sk_s2hat_get_poly`	✅	4s	3s	+33%
`unpack_pk_t1`	✅	4s	3s	+33%
`unpack_sk_s2hat`	✅	4s	3s	+33%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	3s	+0%
`keccak_squeeze`	✅	3s	3s	+0%
`keccakf1600_extract_bytes (big endian)`	✅	3s	4s	-25%
`keccakf1600_xor_bytes`	✅	3s	3s	+0%
`keccakf1600x4_extract_bytes`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes_native`	✅	3s	1s	+200%
`mld_ct_cmask_neg_i32`	✅	3s	2s	+50%
`mld_h`	✅	3s	4s	-25%
`mld_keccakf1600_extract_bytes`	✅	3s	3s	+0%
`mld_prepare_domain_separation_prefix`	✅	3s	3s	+0%
`mld_value_barrier_i64`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	6s	-50%
`pointwise_native_aarch64`	✅	3s	3s	+0%
`poly_caddq`	✅	3s	4s	-25%
`poly_caddq_c`	✅	3s	3s	+0%
`poly_chknorm_native`	✅	3s	4s	-25%
`poly_chknorm_native_aarch64`	✅	3s	4s	-25%
`poly_decompose_native`	✅	3s	2s	+50%
`poly_ntt`	✅	3s	4s	-25%
`poly_permute_bitrev_to_custom_optional`	✅	3s	2s	+50%
`poly_reduce`	✅	3s	4s	-25%
`poly_uniform_gamma1`	✅	3s	3s	+0%
`poly_uniform_gamma1_4x`	✅	3s	5s	-40%
`poly_use_hint_c`	✅	3s	2s	+50%
`poly_use_hint_native`	✅	3s	3s	+0%
`polyeta_pack`	✅	3s	4s	-25%
`polyt0_pack`	✅	3s	4s	-25%
`polyt1_pack`	✅	3s	4s	-25%
`polyt1_unpack`	✅	3s	3s	+0%
`polyvec_matrix_expand`	✅	3s	2s	+50%
`polyveck_chknorm`	✅	3s	3s	+0%
`polyveck_pack_eta`	✅	3s	2s	+50%
`polyvecl_pointwise_acc_montgomery`	✅	3s	4s	-25%
`polyvecl_pointwise_acc_montgomery_c`	✅	3s	3s	+0%
`power2round`	✅	3s	3s	+0%
`rej_uniform_native_aarch64`	✅	3s	4s	-25%
`shake128_finalize`	✅	3s	2s	+50%
`shake128_squeeze`	✅	3s	1s	+200%
`shake256`	✅	3s	4s	-25%
`shake256_absorb`	✅	3s	2s	+50%
`shake256_release`	✅	3s	1s	+200%
`shake256_squeeze`	✅	3s	3s	+0%
`sign_keypair`	✅	3s	3s	+0%
`sign_keypair_internal`	✅	3s	7s	-57%
`sign_signature`	✅	3s	3s	+0%
`sign_signature_extmu`	✅	3s	4s	-25%
`sign_signature_pre_hash_internal`	✅	3s	3s	+0%
`sign_verify_pre_hash_shake256`	✅	3s	3s	+0%
`sk_s1hat_get_poly`	✅	3s	2s	+50%
`sys_check_capability`	✅	3s	4s	-25%
`yvec_get_poly`	✅	3s	2s	+50%
`yvec_init`	✅	3s	2s	+50%
`caddq`	✅	2s	2s	+0%
`decompose`	✅	2s	3s	-33%
`intt_native_x86_64`	✅	2s	4s	-50%
`keccak_f1600_x1_native_aarch64`	✅	2s	2s	+0%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	3s	-33%
`keccak_finalize`	✅	2s	2s	+0%
`keccakf1600_xor_bytes (big endian)`	✅	2s	2s	+0%
`keccakf1600x4_permute`	✅	2s	2s	+0%
`keccakf1600x4_xor_bytes`	✅	2s	2s	+0%
`make_hint`	✅	2s	2s	+0%
`mld_ct_abs_i32`	✅	2s	3s	-33%
`mld_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u8`	✅	2s	2s	+0%
`mld_ct_sel_int32`	✅	2s	2s	+0%
`mld_polymat_expand_entry`	✅	2s	5s	-60%
`mld_value_barrier_u32`	✅	2s	2s	+0%
`mld_value_barrier_u8`	✅	2s	2s	+0%
`pack_sig_h`	✅	2s	3s	-33%
`pack_sig_z`	✅	2s	3s	-33%
`pointwise_native_x86_64`	✅	2s	4s	-50%
`poly_caddq_native`	✅	2s	4s	-50%
`poly_caddq_native_aarch64`	✅	2s	3s	-33%
`poly_caddq_native_x86_64`	✅	2s	3s	-33%
`poly_chknorm`	✅	2s	2s	+0%
`poly_decompose_32_native_aarch64`	✅	2s	2s	+0%
`poly_decompose_88_native_aarch64`	✅	2s	4s	-50%
`poly_invntt_tomont_native`	✅	2s	2s	+0%
`poly_ntt_native`	✅	2s	4s	-50%
`poly_sub`	✅	2s	1s	+100%
`poly_uniform_4x`	✅	2s	3s	-33%
`poly_use_hint`	✅	2s	3s	-33%
`polyvec_matrix_expand_serial`	✅	2s	4s	-50%
`polyveck_pack_w1`	✅	2s	3s	-33%
`polyveck_unpack_eta`	✅	2s	4s	-50%
`polyvecl_unpack_z`	✅	2s	3s	-33%
`polyw1_pack_88`	✅	2s	2s	+0%
`polyz_unpack`	✅	2s	3s	-33%
`reduce32`	✅	2s	5s	-60%
`rej_eta`	✅	2s	3s	-33%
`shake128_init`	✅	2s	2s	+0%
`shake128_release`	✅	2s	2s	+0%
`shake128x4_squeezeblocks`	✅	2s	2s	+0%
`shake256_finalize`	✅	2s	1s	+100%
`shake256x4_absorb_once`	✅	2s	2s	+0%
`shake256x4_squeezeblocks`	✅	2s	3s	-33%
`sign_open`	✅	2s	5s	-60%
`unpack_sk_s1hat`	✅	2s	3s	-33%
`unpack_sk_t0hat`	✅	2s	7s	-71%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	2s	-50%
`keccak_f1600_x4_native_avx2`	✅	1s	4s	-75%
`keccakf1600_permute`	✅	1s	2s	-50%
`mld_ct_cmask_nonzero_u8`	✅	1s	3s	-67%
`mld_ct_get_optblocker_i64`	✅	1s	1s	+0%
`mld_keccakf1600x4_extract_bytes_c`	✅	1s	2s	-50%
`mld_keccakf1600x4_xor_bytes_c`	✅	1s	3s	-67%
`pack_sk_s1`	✅	1s	2s	-50%
`poly_pointwise_montgomery_native`	✅	1s	5s	-80%
`polyvecl_pack_eta`	✅	1s	2s	-50%
`polyvecl_pointwise_acc_montgomery_native`	✅	1s	3s	-67%
`polyvecl_uniform_gamma1_serial`	✅	1s	5s	-80%
`polyw1_pack_32`	✅	1s	3s	-67%
`use_hint`	✅	1s	4s	-75%

oqs-bot · 2026-06-14T03:49:10Z

CBMC Results (ML-DSA-44)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1727s	1752s	-1.4%
`mld_invntt_layer`	✅	284s	284s	+0%
`rej_uniform_native`	✅	144s	143s	+1%
`polyvecl_pointwise_acc_montgomery_c`	✅	116s	123s	-6%
`poly_pointwise_montgomery_c`	✅	91s	88s	+3%
`mld_ct_memcmp`	✅	67s	67s	+0%
`mld_attempt_signature_generation`	✅	58s	62s	-6%
`mld_ntt_layer`	✅	42s	47s	-11%
`fqmul`	✅	41s	43s	-5%
`sign_verify_internal`	✅	28s	27s	+4%
`polyvec_matrix_expand`	✅	27s	26s	+4%
`keccakf1600x4_permute_native`	✅	22s	24s	-8%
`mld_ntt_butterfly_block`	✅	21s	22s	-5%
`rej_uniform`	✅	21s	23s	-9%
`sign_signature_internal`	✅	18s	18s	+0%
`poly_chknorm_c`	✅	17s	19s	-11%
`polyt0_unpack`	✅	17s	15s	+13%
`mld_check_pct`	✅	15s	14s	+7%
`polyeta_unpack`	✅	15s	15s	+0%
`compute_pack_t0_t1`	✅	13s	15s	-13%
`poly_uniform_eta_4x`	✅	13s	12s	+8%
`rej_uniform_c`	✅	13s	12s	+8%
`polyz_unpack_c`	✅	12s	10s	+20%
`poly_invntt_tomont_c`	✅	10s	7s	+43%
`poly_uniform_4x`	✅	10s	10s	+0%
`polyveck_chknorm`	✅	10s	11s	-9%
`sign`	✅	10s	7s	+43%
`keccak_absorb_once_x4`	✅	9s	14s	-36%
`mld_compute_pack_z`	✅	9s	11s	-18%
`poly_add`	✅	9s	10s	-10%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	9s	10s	-10%
`polyvec_matrix_expand_serial`	✅	8s	6s	+33%
`pointwise_acc_native_aarch64`	✅	7s	8s	-12%
`poly_use_hint_c`	✅	7s	6s	+17%
`mld_h`	✅	6s	6s	+0%
`mld_keccakf1600_permute_c`	✅	6s	8s	-25%
`pointwise_acc_native_x86_64`	✅	6s	4s	+50%
`poly_chknorm_native`	✅	6s	2s	+200%
`poly_uniform_gamma1_4x`	✅	6s	4s	+50%
`polyt0_pack`	✅	6s	6s	+0%
`polyveck_decompose`	✅	6s	10s	-40%
`polyveck_invntt_tomont`	✅	6s	6s	+0%
`sign_keypair_internal`	✅	6s	4s	+50%
`fqscale`	✅	5s	3s	+67%
`keccak_absorb`	✅	5s	6s	-17%
`keccakf1600x4_extract_bytes_native`	✅	5s	3s	+67%
`mld_prepare_domain_separation_prefix`	✅	5s	4s	+25%
`poly_caddq_c`	✅	5s	4s	+25%
`poly_ntt`	✅	5s	4s	+25%
`poly_uniform`	✅	5s	7s	-29%
`polyt1_pack`	✅	5s	4s	+25%
`polyvecl_uniform_gamma1_serial`	✅	5s	3s	+67%
`polyw1_pack`	✅	5s	4s	+25%
`sign_keypair`	✅	5s	4s	+25%
`sign_verify_pre_hash_internal`	✅	5s	6s	-17%
`sign_verify_pre_hash_shake256`	✅	5s	4s	+25%
`unpack_sk_s1hat`	✅	5s	3s	+67%
`use_hint`	✅	5s	2s	+150%
`intt_native_x86_64`	✅	4s	4s	+0%
`keccak_f1600_x1_native_aarch64_v84a`	✅	4s	2s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	4s	2s	+100%
`keccakf1600_xor_bytes (big endian)`	✅	4s	3s	+33%
`keccakf1600x4_permute`	✅	4s	3s	+33%
`make_hint`	✅	4s	4s	+0%
`mld_ct_get_optblocker_i64`	✅	4s	2s	+100%
`mld_value_barrier_i64`	✅	4s	1s	+300%
`nttunpack_native_x86_64`	✅	4s	5s	-20%
`pack_sig_c`	✅	4s	2s	+100%
`pointwise_native_x86_64`	✅	4s	2s	+100%
`poly_caddq_native`	✅	4s	4s	+0%
`poly_challenge`	✅	4s	4s	+0%
`poly_decompose_88_native_aarch64`	✅	4s	2s	+100%
`poly_decompose_c`	✅	4s	4s	+0%
`poly_uniform_gamma1`	✅	4s	4s	+0%
`poly_use_hint_native_aarch64`	✅	4s	4s	+0%
`polyeta_pack`	✅	4s	4s	+0%
`polyveck_ntt`	✅	4s	3s	+33%
`polyvecl_chknorm`	✅	4s	5s	-20%
`polyvecl_pointwise_acc_montgomery_native`	✅	4s	2s	+100%
`rej_eta_native`	✅	4s	4s	+0%
`shake256`	✅	4s	1s	+300%
`sign_pk_from_sk`	✅	4s	6s	-33%
`sign_signature`	✅	4s	4s	+0%
`sign_signature_extmu`	✅	4s	3s	+33%
`sign_signature_pre_hash_internal`	✅	4s	4s	+0%
`sign_signature_pre_hash_shake256`	✅	4s	5s	-20%
`sign_verify_extmu`	✅	4s	4s	+0%
`sk_t0hat_get_poly`	✅	4s	4s	+0%
`keccak_f1600_x1_native_aarch64`	✅	3s	3s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	3s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	3s	2s	+50%
`keccak_finalize`	✅	3s	5s	-40%
`keccakf1600_extract_bytes (big endian)`	✅	3s	2s	+50%
`keccakf1600_permute_native`	✅	3s	6s	-50%
`keccakf1600_xor_bytes`	✅	3s	1s	+200%
`keccakf1600x4_extract_bytes`	✅	3s	4s	-25%
`keccakf1600x4_xor_bytes`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes_native`	✅	3s	2s	+50%
`mld_ct_abs_i32`	✅	3s	2s	+50%
`mld_keccakf1600x4_xor_bytes_c`	✅	3s	4s	-25%
`mld_polymat_expand_entry`	✅	3s	2s	+50%
`mld_sample_s1_s2`	✅	3s	3s	+0%
`mld_sample_s1_s2_serial`	✅	3s	3s	+0%
`ntt_native_aarch64`	✅	3s	4s	-25%
`ntt_native_x86_64`	✅	3s	3s	+0%
`pack_sig_z`	✅	3s	3s	+0%
`pack_sk_rho_key_tr_s2`	✅	3s	2s	+50%
`pack_sk_s1`	✅	3s	2s	+50%
`poly_caddq`	✅	3s	3s	+0%
`poly_chknorm`	✅	3s	4s	-25%
`poly_chknorm_native_aarch64`	✅	3s	4s	-25%
`poly_chknorm_native_x86_64`	✅	3s	3s	+0%
`poly_invntt_tomont`	✅	3s	3s	+0%
`poly_invntt_tomont_native`	✅	3s	5s	-40%
`poly_pointwise_montgomery`	✅	3s	2s	+50%
`poly_power2round`	✅	3s	5s	-40%
`poly_reduce`	✅	3s	4s	-25%
`poly_shiftl`	✅	3s	2s	+50%
`poly_sub`	✅	3s	2s	+50%
`poly_uniform_eta`	✅	3s	3s	+0%
`poly_use_hint_native`	✅	3s	2s	+50%
`polyvec_matrix_pointwise_montgomery_row`	✅	3s	3s	+0%
`polyveck_caddq`	✅	3s	2s	+50%
`polyveck_reduce`	✅	3s	4s	-25%
`polyveck_unpack_eta`	✅	3s	2s	+50%
`polyvecl_ntt`	✅	3s	5s	-40%
`polyvecl_pack_eta`	✅	3s	3s	+0%
`polyvecl_pointwise_acc_montgomery`	✅	3s	3s	+0%
`polyvecl_uniform_gamma1`	✅	3s	3s	+0%
`polyvecl_unpack_z`	✅	3s	3s	+0%
`polyw1_pack_88`	✅	3s	4s	-25%
`polyz_unpack_native`	✅	3s	4s	-25%
`reduce32`	✅	3s	1s	+200%
`rej_uniform_eta_native_aarch64`	✅	3s	2s	+50%
`rej_uniform_native_aarch64`	✅	3s	2s	+50%
`shake128_absorb`	✅	3s	2s	+50%
`shake128_finalize`	✅	3s	2s	+50%
`shake128_init`	✅	3s	3s	+0%
`shake128_release`	✅	3s	3s	+0%
`shake128x4_squeezeblocks`	✅	3s	3s	+0%
`shake256_release`	✅	3s	5s	-40%
`shake256_squeeze`	✅	3s	1s	+200%
`shake256x4_squeezeblocks`	✅	3s	1s	+200%
`sig_unpack_hints`	✅	3s	3s	+0%
`sign_verify`	✅	3s	6s	-50%
`unpack_pk_t1`	✅	3s	2s	+50%
`unpack_sk`	✅	3s	3s	+0%
`unpack_sk_s2hat`	✅	3s	1s	+200%
`unpack_sk_t0hat`	✅	3s	4s	-25%
`yvec_init`	✅	3s	3s	+0%
`caddq`	✅	2s	2s	+0%
`decompose`	✅	2s	3s	-33%
`intt_native_aarch64`	✅	2s	3s	-33%
`keccak_f1600_x4_native_avx2`	✅	2s	2s	+0%
`keccak_init`	✅	2s	1s	+100%
`keccak_squeeze`	✅	2s	2s	+0%
`keccak_squeezeblocks_x4`	✅	2s	5s	-60%
`mld_ct_cmask_neg_i32`	✅	2s	4s	-50%
`mld_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u8`	✅	2s	2s	+0%
`mld_ct_sel_int32`	✅	2s	2s	+0%
`mld_keccakf1600_extract_bytes`	✅	2s	2s	+0%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mld_value_barrier_u32`	✅	2s	3s	-33%
`mld_value_barrier_u8`	✅	2s	4s	-50%
`montgomery_reduce`	✅	2s	2s	+0%
`pointwise_native_aarch64`	✅	2s	3s	-33%
`poly_caddq_native_aarch64`	✅	2s	2s	+0%
`poly_caddq_native_x86_64`	✅	2s	4s	-50%
`poly_decompose`	✅	2s	4s	-50%
`poly_decompose_32_native_aarch64`	✅	2s	1s	+100%
`poly_decompose_native`	✅	2s	2s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	2s	6s	-67%
`poly_pointwise_montgomery_native`	✅	2s	1s	+100%
`poly_use_hint`	✅	2s	3s	-33%
`polyt1_unpack`	✅	2s	2s	+0%
`polyveck_pack_eta`	✅	2s	2s	+0%
`polyveck_pack_w1`	✅	2s	4s	-50%
`polyvecl_unpack_eta`	✅	2s	4s	-50%
`polyw1_pack_32`	✅	2s	4s	-50%
`polyz_pack`	✅	2s	4s	-50%
`polyz_unpack`	✅	2s	2s	+0%
`polyz_unpack_17_native_aarch64`	✅	2s	2s	+0%
`polyz_unpack_19_native_aarch64`	✅	2s	4s	-50%
`power2round`	✅	2s	2s	+0%
`rej_eta`	✅	2s	1s	+100%
`rej_eta_c`	✅	2s	5s	-60%
`shake256_absorb`	✅	2s	2s	+0%
`shake256_init`	✅	2s	1s	+100%
`sign_open`	✅	2s	6s	-67%
`sk_s1hat_get_poly`	✅	2s	1s	+100%
`sk_s2hat_get_poly`	✅	2s	3s	-33%
`yvec_get_poly`	✅	2s	2s	+0%
`keccakf1600_permute`	✅	1s	1s	+0%
`mld_ct_cmask_nonzero_u32`	✅	1s	2s	-50%
`pack_sig_h`	✅	1s	1s	+0%
`poly_ntt_c`	✅	1s	3s	-67%
`poly_ntt_native`	✅	1s	1s	+0%
`poly_permute_bitrev_to_custom_optional`	✅	1s	6s	-83%
`shake128_squeeze`	✅	1s	2s	-50%
`shake128x4_absorb_once`	✅	1s	3s	-67%
`shake256_finalize`	✅	1s	3s	-67%
`shake256x4_absorb_once`	✅	1s	4s	-75%
`sys_check_capability`	✅	1s	2s	-50%

oqs-bot · 2026-06-14T03:49:27Z

CBMC Results (ML-DSA-65)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	2190s	2040s	+7.4%
`mld_invntt_layer`	✅	311s	286s	+9%
`polyvecl_pointwise_acc_montgomery_c`	✅	232s	193s	+20%
`rej_uniform_native`	✅	154s	149s	+3%
`polyvec_matrix_expand`	✅	137s	130s	+5%
`poly_pointwise_montgomery_c`	✅	109s	99s	+10%
`mld_ct_memcmp`	✅	76s	67s	+13%
`mld_attempt_signature_generation`	✅	67s	66s	+2%
`sign_verify_internal`	✅	64s	64s	+0%
`sign_signature_internal`	✅	48s	48s	+0%
`mld_ntt_layer`	✅	47s	43s	+9%
`fqmul`	✅	43s	43s	+0%
`polyvec_matrix_expand_serial`	✅	28s	25s	+12%
`mld_ntt_butterfly_block`	✅	24s	22s	+9%
`keccakf1600x4_permute_native`	✅	22s	23s	-4%
`poly_chknorm_c`	✅	22s	19s	+16%
`rej_uniform`	✅	22s	22s	+0%
`mld_check_pct`	✅	19s	13s	+46%
`compute_pack_t0_t1`	✅	18s	16s	+12%
`polyvecl_chknorm`	✅	18s	16s	+12%
`polyt0_unpack`	✅	16s	18s	-11%
`polyveck_decompose`	✅	15s	15s	+0%
`rej_uniform_c`	✅	14s	12s	+17%
`poly_add`	✅	13s	10s	+30%
`polyveck_chknorm`	✅	13s	11s	+18%
`keccak_absorb_once_x4`	✅	12s	12s	+0%
`poly_uniform_4x`	✅	12s	11s	+9%
`poly_uniform_eta_4x`	✅	12s	12s	+0%
`sign`	✅	10s	7s	+43%
`mld_compute_pack_z`	✅	9s	10s	-10%
`polyveck_caddq`	✅	9s	7s	+29%
`polyveck_invntt_tomont`	✅	9s	8s	+12%
`poly_invntt_tomont_c`	✅	8s	8s	+0%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	8s	10s	-20%
`polyveck_ntt`	✅	8s	9s	-11%
`sign_pk_from_sk`	✅	8s	7s	+14%
`unpack_sk_t0hat`	✅	8s	6s	+33%
`mld_keccakf1600_permute_c`	✅	7s	7s	+0%
`mld_sample_s1_s2_serial`	✅	7s	9s	-22%
`polyvecl_pointwise_acc_montgomery_native`	✅	7s	3s	+133%
`keccak_absorb`	✅	6s	5s	+20%
`mld_prepare_domain_separation_prefix`	✅	6s	3s	+100%
`pointwise_acc_native_aarch64`	✅	6s	6s	+0%
`pointwise_acc_native_x86_64`	✅	6s	6s	+0%
`poly_caddq_c`	✅	6s	7s	-14%
`poly_challenge`	✅	6s	6s	+0%
`poly_uniform`	✅	6s	3s	+100%
`polyvecl_ntt`	✅	6s	6s	+0%
`sign_keypair_internal`	✅	6s	2s	+200%
`sign_open`	✅	6s	3s	+100%
`sign_signature_pre_hash_internal`	✅	6s	7s	-14%
`sign_verify_pre_hash_internal`	✅	6s	5s	+20%
`unpack_sk`	✅	6s	1s	+500%
`keccakf1600_extract_bytes (big endian)`	✅	5s	2s	+150%
`keccakf1600x4_extract_bytes_native`	✅	5s	7s	-29%
`keccakf1600x4_xor_bytes_native`	✅	5s	2s	+150%
`poly_sub`	✅	5s	3s	+67%
`poly_uniform_gamma1_4x`	✅	5s	3s	+67%
`poly_use_hint_native_aarch64`	✅	5s	2s	+150%
`polyt0_pack`	✅	5s	6s	-17%
`polyz_unpack_c`	✅	5s	7s	-29%
`sig_unpack_hints`	✅	5s	3s	+67%
`sign_keypair`	✅	5s	3s	+67%
`sk_s1hat_get_poly`	✅	5s	2s	+150%
`intt_native_aarch64`	✅	4s	2s	+100%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	2s	+100%
`keccak_squeezeblocks_x4`	✅	4s	5s	-20%
`keccakf1600x4_xor_bytes`	✅	4s	3s	+33%
`mld_ct_get_optblocker_u32`	✅	4s	4s	+0%
`mld_h`	✅	4s	3s	+33%
`montgomery_reduce`	✅	4s	2s	+100%
`ntt_native_aarch64`	✅	4s	5s	-20%
`ntt_native_x86_64`	✅	4s	3s	+33%
`pack_sig_c`	✅	4s	4s	+0%
`pack_sig_h`	✅	4s	4s	+0%
`poly_caddq_native_aarch64`	✅	4s	2s	+100%
`poly_chknorm_native`	✅	4s	4s	+0%
`poly_chknorm_native_aarch64`	✅	4s	4s	+0%
`poly_decompose_88_native_aarch64`	✅	4s	2s	+100%
`poly_ntt_native`	✅	4s	2s	+100%
`poly_permute_bitrev_to_custom_optional`	✅	4s	3s	+33%
`poly_permute_bitrev_to_custom_optional_native`	✅	4s	3s	+33%
`poly_uniform_gamma1`	✅	4s	2s	+100%
`polyveck_pack_eta`	✅	4s	1s	+300%
`polyveck_pack_w1`	✅	4s	3s	+33%
`polyveck_reduce`	✅	4s	3s	+33%
`polyveck_unpack_eta`	✅	4s	4s	+0%
`polyvecl_pack_eta`	✅	4s	5s	-20%
`polyvecl_uniform_gamma1_serial`	✅	4s	3s	+33%
`polyw1_pack_88`	✅	4s	3s	+33%
`polyz_unpack_17_native_aarch64`	✅	4s	3s	+33%
`rej_eta_c`	✅	4s	2s	+100%
`rej_eta_native`	✅	4s	4s	+0%
`rej_uniform_eta_native_aarch64`	✅	4s	4s	+0%
`rej_uniform_native_aarch64`	✅	4s	4s	+0%
`shake256_release`	✅	4s	2s	+100%
`sign_signature_extmu`	✅	4s	2s	+100%
`sign_signature_pre_hash_shake256`	✅	4s	4s	+0%
`sign_verify_pre_hash_shake256`	✅	4s	4s	+0%
`yvec_get_poly`	✅	4s	4s	+0%
`caddq`	✅	3s	6s	-50%
`decompose`	✅	3s	4s	-25%
`fqscale`	✅	3s	5s	-40%
`intt_native_x86_64`	✅	3s	3s	+0%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`keccak_squeeze`	✅	3s	1s	+200%
`keccakf1600_xor_bytes`	✅	3s	2s	+50%
`make_hint`	✅	3s	3s	+0%
`mld_ct_cmask_nonzero_u32`	✅	3s	4s	-25%
`mld_ct_cmask_nonzero_u8`	✅	3s	2s	+50%
`mld_keccakf1600x4_xor_bytes_c`	✅	3s	3s	+0%
`mld_polymat_expand_entry`	✅	3s	2s	+50%
`nttunpack_native_x86_64`	✅	3s	6s	-50%
`pack_sk_rho_key_tr_s2`	✅	3s	5s	-40%
`pointwise_native_aarch64`	✅	3s	3s	+0%
`pointwise_native_x86_64`	✅	3s	3s	+0%
`poly_chknorm`	✅	3s	2s	+50%
`poly_decompose_32_native_aarch64`	✅	3s	3s	+0%
`poly_decompose_c`	✅	3s	5s	-40%
`poly_decompose_native`	✅	3s	2s	+50%
`poly_invntt_tomont`	✅	3s	4s	-25%
`poly_invntt_tomont_native`	✅	3s	2s	+50%
`poly_ntt`	✅	3s	3s	+0%
`poly_ntt_c`	✅	3s	2s	+50%
`poly_pointwise_montgomery_native`	✅	3s	3s	+0%
`poly_power2round`	✅	3s	4s	-25%
`poly_reduce`	✅	3s	2s	+50%
`poly_shiftl`	✅	3s	7s	-57%
`poly_uniform_eta`	✅	3s	3s	+0%
`poly_use_hint`	✅	3s	3s	+0%
`poly_use_hint_c`	✅	3s	3s	+0%
`poly_use_hint_native`	✅	3s	3s	+0%
`polyeta_pack`	✅	3s	3s	+0%
`polyt1_pack`	✅	3s	2s	+50%
`polyt1_unpack`	✅	3s	4s	-25%
`polyvecl_pointwise_acc_montgomery`	✅	3s	2s	+50%
`polyvecl_uniform_gamma1`	✅	3s	4s	-25%
`polyvecl_unpack_eta`	✅	3s	3s	+0%
`polyw1_pack_32`	✅	3s	3s	+0%
`polyz_pack`	✅	3s	4s	-25%
`polyz_unpack_19_native_aarch64`	✅	3s	3s	+0%
`reduce32`	✅	3s	3s	+0%
`shake128_init`	✅	3s	5s	-40%
`shake128_release`	✅	3s	2s	+50%
`shake256`	✅	3s	3s	+0%
`shake256_finalize`	✅	3s	3s	+0%
`sign_signature`	✅	3s	3s	+0%
`sign_verify_extmu`	✅	3s	4s	-25%
`unpack_pk_t1`	✅	3s	3s	+0%
`unpack_sk_s1hat`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	1s	+100%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	2s	+0%
`keccak_finalize`	✅	2s	2s	+0%
`keccak_init`	✅	2s	3s	-33%
`keccakf1600_permute`	✅	2s	3s	-33%
`keccakf1600_permute_native`	✅	2s	3s	-33%
`keccakf1600x4_extract_bytes`	✅	2s	2s	+0%
`keccakf1600x4_permute`	✅	2s	2s	+0%
`mld_ct_cmask_neg_i32`	✅	2s	3s	-33%
`mld_ct_get_optblocker_u8`	✅	2s	1s	+100%
`mld_ct_sel_int32`	✅	2s	1s	+100%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mld_sample_s1_s2`	✅	2s	4s	-50%
`mld_value_barrier_u8`	✅	2s	2s	+0%
`pack_sig_z`	✅	2s	2s	+0%
`pack_sk_s1`	✅	2s	3s	-33%
`poly_caddq`	✅	2s	2s	+0%
`poly_caddq_native`	✅	2s	3s	-33%
`poly_caddq_native_x86_64`	✅	2s	2s	+0%
`poly_chknorm_native_x86_64`	✅	2s	6s	-67%
`poly_decompose`	✅	2s	2s	+0%
`poly_pointwise_montgomery`	✅	2s	2s	+0%
`polyvec_matrix_pointwise_montgomery_row`	✅	2s	2s	+0%
`polyvecl_unpack_z`	✅	2s	3s	-33%
`polyw1_pack`	✅	2s	3s	-33%
`polyz_unpack`	✅	2s	2s	+0%
`polyz_unpack_native`	✅	2s	2s	+0%
`rej_eta`	✅	2s	3s	-33%
`shake128_finalize`	✅	2s	3s	-33%
`shake128_squeeze`	✅	2s	1s	+100%
`shake128x4_absorb_once`	✅	2s	2s	+0%
`shake128x4_squeezeblocks`	✅	2s	2s	+0%
`shake256_absorb`	✅	2s	2s	+0%
`shake256_init`	✅	2s	1s	+100%
`shake256_squeeze`	✅	2s	3s	-33%
`shake256x4_absorb_once`	✅	2s	2s	+0%
`shake256x4_squeezeblocks`	✅	2s	3s	-33%
`sign_verify`	✅	2s	5s	-60%
`unpack_sk_s2hat`	✅	2s	4s	-50%
`yvec_init`	✅	2s	1s	+100%
`keccakf1600_xor_bytes (big endian)`	✅	1s	3s	-67%
`mld_ct_abs_i32`	✅	1s	2s	-50%
`mld_ct_get_optblocker_i64`	✅	1s	1s	+0%
`mld_keccakf1600_extract_bytes`	✅	1s	2s	-50%
`mld_value_barrier_i64`	✅	1s	3s	-67%
`mld_value_barrier_u32`	✅	1s	2s	-50%
`polyeta_unpack`	✅	1s	1s	+0%
`power2round`	✅	1s	2s	-50%
`shake128_absorb`	✅	1s	3s	-67%
`sk_s2hat_get_poly`	✅	1s	2s	-50%
`sk_t0hat_get_poly`	✅	1s	3s	-67%
`sys_check_capability`	✅	1s	2s	-50%
`use_hint`	✅	1s	2s	-50%

oqs-bot · 2026-06-14T03:50:35Z

CBMC Results (ML-DSA-87)

Full Results (204 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	2379s	2406s	-1.1%
`polyvecl_pointwise_acc_montgomery_c`	✅	307s	319s	-4%
`mld_invntt_layer`	✅	285s	291s	-2%
`polyvec_matrix_expand`	✅	219s	223s	-2%
`rej_uniform_native`	✅	149s	153s	-3%
`mld_attempt_signature_generation`	✅	100s	105s	-5%
`poly_pointwise_montgomery_c`	✅	99s	94s	+5%
`mld_ct_memcmp`	✅	67s	64s	+5%
`sign_signature_internal`	✅	64s	68s	-6%
`sign_verify_internal`	✅	60s	61s	-2%
`polyvec_matrix_expand_serial`	✅	49s	48s	+2%
`mld_ntt_layer`	✅	45s	44s	+2%
`fqmul`	✅	40s	40s	+0%
`compute_pack_t0_t1`	✅	32s	30s	+7%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	28s	30s	-7%
`mld_ntt_butterfly_block`	✅	24s	22s	+9%
`keccakf1600x4_permute_native`	✅	22s	25s	-12%
`rej_uniform`	✅	21s	23s	-9%
`poly_chknorm_c`	✅	20s	17s	+18%
`mld_check_pct`	✅	16s	15s	+7%
`polyeta_unpack`	✅	16s	14s	+14%
`polyt0_unpack`	✅	16s	19s	-16%
`poly_uniform_eta_4x`	✅	13s	12s	+8%
`rej_uniform_c`	✅	11s	13s	-15%
`poly_add`	✅	10s	11s	-9%
`keccak_absorb_once_x4`	✅	9s	10s	-10%
`poly_invntt_tomont_c`	✅	9s	8s	+12%
`poly_uniform_4x`	✅	9s	12s	-25%
`polyveck_decompose`	✅	9s	8s	+12%
`polyveck_invntt_tomont`	✅	9s	9s	+0%
`polyvecl_ntt`	✅	9s	8s	+12%
`mld_keccakf1600_permute_c`	✅	8s	7s	+14%
`polyveck_caddq`	✅	8s	8s	+0%
`keccak_absorb`	✅	7s	6s	+17%
`mld_sample_s1_s2`	✅	7s	8s	-12%
`pointwise_acc_native_aarch64`	✅	7s	7s	+0%
`pointwise_acc_native_x86_64`	✅	7s	8s	-12%
`poly_challenge`	✅	7s	5s	+40%
`polyveck_ntt`	✅	7s	7s	+0%
`sign_pk_from_sk`	✅	7s	5s	+40%
`sign_signature_extmu`	✅	7s	4s	+75%
`sign_verify_pre_hash_shake256`	✅	7s	4s	+75%
`mld_compute_pack_z`	✅	6s	7s	-14%
`ntt_native_x86_64`	✅	6s	3s	+100%
`poly_caddq_c`	✅	6s	7s	-14%
`poly_decompose_88_native_aarch64`	✅	6s	4s	+50%
`poly_decompose_c`	✅	6s	8s	-25%
`poly_use_hint_native`	✅	6s	2s	+200%
`polyt0_pack`	✅	6s	3s	+100%
`polyveck_chknorm`	✅	6s	5s	+20%
`sign`	✅	6s	11s	-45%
`sign_verify_pre_hash_internal`	✅	6s	5s	+20%
`use_hint`	✅	6s	2s	+200%
`intt_native_aarch64`	✅	5s	3s	+67%
`keccak_finalize`	✅	5s	1s	+400%
`pointwise_native_x86_64`	✅	5s	3s	+67%
`poly_chknorm`	✅	5s	3s	+67%
`poly_decompose_32_native_aarch64`	✅	5s	6s	-17%
`polyveck_pack_w1`	✅	5s	3s	+67%
`polyz_unpack_c`	✅	5s	4s	+25%
`reduce32`	✅	5s	3s	+67%
`rej_eta_c`	✅	5s	2s	+150%
`sign_keypair_internal`	✅	5s	5s	+0%
`sign_signature_pre_hash_internal`	✅	5s	5s	+0%
`sign_signature_pre_hash_shake256`	✅	5s	7s	-29%
`sign_verify_extmu`	✅	5s	5s	+0%
`unpack_sk_t0hat`	✅	5s	5s	+0%
`intt_native_x86_64`	✅	4s	4s	+0%
`keccak_init`	✅	4s	2s	+100%
`keccakf1600_permute`	✅	4s	2s	+100%
`keccakf1600_xor_bytes`	✅	4s	3s	+33%
`keccakf1600_xor_bytes (big endian)`	✅	4s	3s	+33%
`keccakf1600x4_extract_bytes`	✅	4s	2s	+100%
`keccakf1600x4_xor_bytes`	✅	4s	2s	+100%
`make_hint`	✅	4s	4s	+0%
`mld_h`	✅	4s	2s	+100%
`mld_polymat_expand_entry`	✅	4s	2s	+100%
`mld_sample_s1_s2_serial`	✅	4s	5s	-20%
`mld_value_barrier_u32`	✅	4s	3s	+33%
`poly_caddq`	✅	4s	2s	+100%
`poly_caddq_native`	✅	4s	5s	-20%
`poly_caddq_native_aarch64`	✅	4s	4s	+0%
`poly_decompose_native`	✅	4s	5s	-20%
`poly_invntt_tomont_native`	✅	4s	4s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	4s	4s	+0%
`poly_power2round`	✅	4s	4s	+0%
`poly_shiftl`	✅	4s	3s	+33%
`polyvec_matrix_pointwise_montgomery_row`	✅	4s	7s	-43%
`polyveck_pack_eta`	✅	4s	2s	+100%
`polyvecl_uniform_gamma1`	✅	4s	3s	+33%
`polyvecl_uniform_gamma1_serial`	✅	4s	2s	+100%
`polyw1_pack_88`	✅	4s	3s	+33%
`polyz_unpack_17_native_aarch64`	✅	4s	1s	+300%
`polyz_unpack_19_native_aarch64`	✅	4s	3s	+33%
`rej_eta`	✅	4s	3s	+33%
`rej_eta_native`	✅	4s	5s	-20%
`shake128_absorb`	✅	4s	3s	+33%
`shake128_release`	✅	4s	3s	+33%
`shake256`	✅	4s	2s	+100%
`sign_keypair`	✅	4s	5s	-20%
`sign_signature`	✅	4s	4s	+0%
`sign_verify`	✅	4s	3s	+33%
`sk_s2hat_get_poly`	✅	4s	4s	+0%
`unpack_pk_t1`	✅	4s	3s	+33%
`yvec_get_poly`	✅	4s	4s	+0%
`caddq`	✅	3s	3s	+0%
`decompose`	✅	3s	2s	+50%
`fqscale`	✅	3s	3s	+0%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	1s	+200%
`keccak_f1600_x4_native_avx2`	✅	3s	3s	+0%
`keccak_squeezeblocks_x4`	✅	3s	5s	-40%
`keccakf1600_extract_bytes (big endian)`	✅	3s	2s	+50%
`keccakf1600_permute_native`	✅	3s	5s	-40%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`mld_ct_abs_i32`	✅	3s	3s	+0%
`mld_ct_cmask_nonzero_u32`	✅	3s	4s	-25%
`mld_ct_cmask_nonzero_u8`	✅	3s	2s	+50%
`mld_keccakf1600_extract_bytes`	✅	3s	5s	-40%
`mld_prepare_domain_separation_prefix`	✅	3s	4s	-25%
`montgomery_reduce`	✅	3s	2s	+50%
`ntt_native_aarch64`	✅	3s	3s	+0%
`nttunpack_native_x86_64`	✅	3s	5s	-40%
`pack_sig_c`	✅	3s	5s	-40%
`pack_sig_h`	✅	3s	3s	+0%
`pack_sig_z`	✅	3s	2s	+50%
`pack_sk_s1`	✅	3s	3s	+0%
`pointwise_native_aarch64`	✅	3s	4s	-25%
`poly_caddq_native_x86_64`	✅	3s	2s	+50%
`poly_chknorm_native`	✅	3s	4s	-25%
`poly_chknorm_native_aarch64`	✅	3s	4s	-25%
`poly_decompose`	✅	3s	3s	+0%
`poly_pointwise_montgomery`	✅	3s	2s	+50%
`poly_pointwise_montgomery_native`	✅	3s	3s	+0%
`poly_reduce`	✅	3s	3s	+0%
`poly_uniform`	✅	3s	4s	-25%
`poly_uniform_eta`	✅	3s	4s	-25%
`poly_uniform_gamma1`	✅	3s	4s	-25%
`poly_uniform_gamma1_4x`	✅	3s	4s	-25%
`poly_use_hint_c`	✅	3s	3s	+0%
`polyveck_reduce`	✅	3s	4s	-25%
`polyvecl_chknorm`	✅	3s	4s	-25%
`polyvecl_pack_eta`	✅	3s	3s	+0%
`polyvecl_pointwise_acc_montgomery`	✅	3s	3s	+0%
`polyvecl_pointwise_acc_montgomery_native`	✅	3s	4s	-25%
`polyvecl_unpack_z`	✅	3s	4s	-25%
`polyz_pack`	✅	3s	3s	+0%
`power2round`	✅	3s	2s	+50%
`rej_uniform_eta_native_aarch64`	✅	3s	2s	+50%
`rej_uniform_native_aarch64`	✅	3s	3s	+0%
`shake128_squeeze`	✅	3s	3s	+0%
`shake256_absorb`	✅	3s	3s	+0%
`shake256_squeeze`	✅	3s	4s	-25%
`shake256x4_absorb_once`	✅	3s	2s	+50%
`shake256x4_squeezeblocks`	✅	3s	3s	+0%
`sign_open`	✅	3s	3s	+0%
`sk_t0hat_get_poly`	✅	3s	4s	-25%
`sys_check_capability`	✅	3s	3s	+0%
`yvec_init`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	3s	-33%
`keccak_squeeze`	✅	2s	5s	-60%
`keccakf1600x4_xor_bytes_native`	✅	2s	3s	-33%
`mld_ct_cmask_neg_i32`	✅	2s	3s	-33%
`mld_ct_get_optblocker_u32`	✅	2s	1s	+100%
`mld_ct_get_optblocker_u8`	✅	2s	3s	-33%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mld_value_barrier_i64`	✅	2s	4s	-50%
`pack_sk_rho_key_tr_s2`	✅	2s	1s	+100%
`poly_invntt_tomont`	✅	2s	3s	-33%
`poly_ntt`	✅	2s	5s	-60%
`poly_ntt_c`	✅	2s	2s	+0%
`poly_ntt_native`	✅	2s	3s	-33%
`poly_permute_bitrev_to_custom_optional`	✅	2s	2s	+0%
`poly_sub`	✅	2s	4s	-50%
`poly_use_hint`	✅	2s	2s	+0%
`poly_use_hint_native_aarch64`	✅	2s	2s	+0%
`polyeta_pack`	✅	2s	3s	-33%
`polyt1_pack`	✅	2s	3s	-33%
`polyt1_unpack`	✅	2s	2s	+0%
`polyveck_unpack_eta`	✅	2s	5s	-60%
`polyvecl_unpack_eta`	✅	2s	3s	-33%
`polyw1_pack`	✅	2s	1s	+100%
`polyw1_pack_32`	✅	2s	3s	-33%
`polyz_unpack`	✅	2s	3s	-33%
`polyz_unpack_native`	✅	2s	2s	+0%
`shake128_finalize`	✅	2s	4s	-50%
`shake128_init`	✅	2s	3s	-33%
`shake128x4_absorb_once`	✅	2s	1s	+100%
`shake128x4_squeezeblocks`	✅	2s	3s	-33%
`shake256_init`	✅	2s	3s	-33%
`shake256_release`	✅	2s	2s	+0%
`sig_unpack_hints`	✅	2s	3s	-33%
`sk_s1hat_get_poly`	✅	2s	3s	-33%
`unpack_sk`	✅	2s	4s	-50%
`unpack_sk_s1hat`	✅	2s	3s	-33%
`unpack_sk_s2hat`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	1s	+0%
`keccakf1600x4_permute`	✅	1s	3s	-67%
`mld_ct_get_optblocker_i64`	✅	1s	3s	-67%
`mld_ct_sel_int32`	✅	1s	5s	-80%
`mld_keccakf1600x4_xor_bytes_c`	✅	1s	3s	-67%
`mld_value_barrier_u8`	✅	1s	2s	-50%
`poly_chknorm_native_x86_64`	✅	1s	3s	-67%
`shake256_finalize`	✅	1s	2s	-50%

github-actions

Mac Mini (M1, 2020) benchmarks (opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`46485` cycles	`46485` cycles	`1`
`ML-DSA-44 sign`	`131073` cycles	`131058` cycles	`1.00`
`ML-DSA-44 verify`	`47306` cycles	`47305` cycles	`1.00`
`ML-DSA-65 keypair`	`81684` cycles	`81689` cycles	`1.00`
`ML-DSA-65 sign`	`215307` cycles	`215323` cycles	`1.00`
`ML-DSA-65 verify`	`79296` cycles	`79302` cycles	`1.00`
`ML-DSA-87 keypair`	`132405` cycles	`132467` cycles	`1.00`
`ML-DSA-87 sign`	`277644` cycles	`277549` cycles	`1.00`
`ML-DSA-87 verify`	`134050` cycles	`134119` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Mac Mini (M1, 2020) benchmarks (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`113139` cycles	`112736` cycles	`1.00`
`ML-DSA-44 sign`	`399645` cycles	`400878` cycles	`1.00`
`ML-DSA-44 verify`	`119028` cycles	`119432` cycles	`1.00`
`ML-DSA-65 keypair`	`193658` cycles	`192975` cycles	`1.00`
`ML-DSA-65 sign`	`644542` cycles	`649981` cycles	`0.99`
`ML-DSA-65 verify`	`191939` cycles	`192870` cycles	`1.00`
`ML-DSA-87 keypair`	`318942` cycles	`318797` cycles	`1.00`
`ML-DSA-87 sign`	`822796` cycles	`828723` cycles	`0.99`
`ML-DSA-87 verify`	`325445` cycles	`326796` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`43417` cycles	`43553` cycles	`1.00`
`ML-DSA-44 sign`	`130691` cycles	`131114` cycles	`1.00`
`ML-DSA-44 verify`	`45128` cycles	`45587` cycles	`0.99`
`ML-DSA-65 keypair`	`75814` cycles	`75908` cycles	`1.00`
`ML-DSA-65 sign`	`215254` cycles	`215310` cycles	`1.00`
`ML-DSA-65 verify`	`74377` cycles	`74636` cycles	`1.00`
`ML-DSA-87 keypair`	`123621` cycles	`123588` cycles	`1.00`
`ML-DSA-87 sign`	`271463` cycles	`272387` cycles	`1.00`
`ML-DSA-87 verify`	`120973` cycles	`120775` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Intel Xeon 4th gen (c7i) (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`91331` cycles	`91530` cycles	`1.00`
`ML-DSA-44 sign`	`351183` cycles	`352430` cycles	`1.00`
`ML-DSA-44 verify`	`99090` cycles	`99915` cycles	`0.99`
`ML-DSA-65 keypair`	`154160` cycles	`153947` cycles	`1.00`
`ML-DSA-65 sign`	`570281` cycles	`572172` cycles	`1.00`
`ML-DSA-65 verify`	`159498` cycles	`159760` cycles	`1.00`
`ML-DSA-87 keypair`	`255196` cycles	`254922` cycles	`1.00`
`ML-DSA-87 sign`	`722193` cycles	`726518` cycles	`0.99`
`ML-DSA-87 verify`	`263128` cycles	`263741` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`55431` cycles	`55164` cycles	`1.00`
`ML-DSA-44 sign`	`159071` cycles	`159123` cycles	`1.00`
`ML-DSA-44 verify`	`57563` cycles	`57907` cycles	`0.99`
`ML-DSA-65 keypair`	`96226` cycles	`95536` cycles	`1.01`
`ML-DSA-65 sign`	`263817` cycles	`263540` cycles	`1.00`
`ML-DSA-65 verify`	`96247` cycles	`96187` cycles	`1.00`
`ML-DSA-87 keypair`	`154986` cycles	`154719` cycles	`1.00`
`ML-DSA-87 sign`	`322492` cycles	`322467` cycles	`1.00`
`ML-DSA-87 verify`	`151333` cycles	`150859` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

AMD EPYC 3rd gen (c6a) (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`133161` cycles	`133031` cycles	`1.00`
`ML-DSA-44 sign`	`516727` cycles	`518961` cycles	`1.00`
`ML-DSA-44 verify`	`146251` cycles	`146356` cycles	`1.00`
`ML-DSA-65 keypair`	`224163` cycles	`223780` cycles	`1.00`
`ML-DSA-65 sign`	`844780` cycles	`843408` cycles	`1.00`
`ML-DSA-65 verify`	`233848` cycles	`234043` cycles	`1.00`
`ML-DSA-87 keypair`	`370291` cycles	`367594` cycles	`1.01`
`ML-DSA-87 sign`	`1065740` cycles	`1061107` cycles	`1.00`
`ML-DSA-87 verify`	`382861` cycles	`380814` cycles	`1.01`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton2

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`112523` cycles	`112521` cycles	`1.00`
`ML-DSA-44 sign`	`354964` cycles	`354071` cycles	`1.00`
`ML-DSA-44 verify`	`117553` cycles	`117392` cycles	`1.00`
`ML-DSA-65 keypair`	`194393` cycles	`194710` cycles	`1.00`
`ML-DSA-65 sign`	`584208` cycles	`584562` cycles	`1.00`
`ML-DSA-65 verify`	`193325` cycles	`193311` cycles	`1.00`
`ML-DSA-87 keypair`	`321448` cycles	`321242` cycles	`1.00`
`ML-DSA-87 sign`	`749057` cycles	`747207` cycles	`1.00`
`ML-DSA-87 verify`	`318617` cycles	`318958` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`47023` cycles	`46737` cycles	`1.01`
`ML-DSA-44 sign`	`140001` cycles	`139162` cycles	`1.01`
`ML-DSA-44 verify`	`49133` cycles	`49579` cycles	`0.99`
`ML-DSA-65 keypair`	`82397` cycles	`82474` cycles	`1.00`
`ML-DSA-65 sign`	`228366` cycles	`228196` cycles	`1.00`
`ML-DSA-65 verify`	`82016` cycles	`81813` cycles	`1.00`
`ML-DSA-87 keypair`	`130805` cycles	`129675` cycles	`1.01`
`ML-DSA-87 sign`	`281569` cycles	`279317` cycles	`1.01`
`ML-DSA-87 verify`	`130616` cycles	`128340` cycles	`1.02`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton4

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`67406` cycles	`67286` cycles	`1.00`
`ML-DSA-44 sign`	`198364` cycles	`198339` cycles	`1.00`
`ML-DSA-44 verify`	`70204` cycles	`70253` cycles	`1.00`
`ML-DSA-65 keypair`	`119285` cycles	`119300` cycles	`1.00`
`ML-DSA-65 sign`	`326202` cycles	`325686` cycles	`1.00`
`ML-DSA-65 verify`	`116868` cycles	`116842` cycles	`1.00`
`ML-DSA-87 keypair`	`196485` cycles	`196569` cycles	`1.00`
`ML-DSA-87 sign`	`421245` cycles	`421883` cycles	`1.00`
`ML-DSA-87 verify`	`193229` cycles	`193336` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

AMD EPYC 4th gen (c7a) (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`118231` cycles	`118077` cycles	`1.00`
`ML-DSA-44 sign`	`455256` cycles	`458854` cycles	`0.99`
`ML-DSA-44 verify`	`130717` cycles	`130666` cycles	`1.00`
`ML-DSA-65 keypair`	`200754` cycles	`201340` cycles	`1.00`
`ML-DSA-65 sign`	`742358` cycles	`741817` cycles	`1.00`
`ML-DSA-65 verify`	`208879` cycles	`209500` cycles	`1.00`
`ML-DSA-87 keypair`	`330201` cycles	`330842` cycles	`1.00`
`ML-DSA-87 sign`	`938760` cycles	`936928` cycles	`1.00`
`ML-DSA-87 verify`	`342173` cycles	`343616` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`212090` cycles	`211907` cycles	`1.00`
`ML-DSA-44 sign`	`753861` cycles	`760155` cycles	`0.99`
`ML-DSA-44 verify`	`228257` cycles	`229379` cycles	`1.00`
`ML-DSA-65 keypair`	`376614` cycles	`378157` cycles	`1.00`
`ML-DSA-65 sign`	`1232908` cycles	`1250904` cycles	`0.99`
`ML-DSA-65 verify`	`370936` cycles	`372722` cycles	`1.00`
`ML-DSA-87 keypair`	`603139` cycles	`600977` cycles	`1.00`
`ML-DSA-87 sign`	`1572054` cycles	`1584763` cycles	`0.99`
`ML-DSA-87 verify`	`616641` cycles	`616480` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`127569` cycles	`127637` cycles	`1.00`
`ML-DSA-44 sign`	`436957` cycles	`441211` cycles	`0.99`
`ML-DSA-44 verify`	`135453` cycles	`136398` cycles	`0.99`
`ML-DSA-65 keypair`	`220890` cycles	`220759` cycles	`1.00`
`ML-DSA-65 sign`	`707720` cycles	`713650` cycles	`0.99`
`ML-DSA-65 verify`	`220274` cycles	`220740` cycles	`1.00`
`ML-DSA-87 keypair`	`364568` cycles	`365093` cycles	`1.00`
`ML-DSA-87 sign`	`908294` cycles	`921276` cycles	`0.99`
`ML-DSA-87 verify`	`369293` cycles	`370783` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Intel Xeon 3rd gen (c6i)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`61928` cycles	`61599` cycles	`1.01`
`ML-DSA-44 sign`	`189843` cycles	`189177` cycles	`1.00`
`ML-DSA-44 verify`	`66558` cycles	`66462` cycles	`1.00`
`ML-DSA-65 keypair`	`111312` cycles	`110803` cycles	`1.00`
`ML-DSA-65 sign`	`316194` cycles	`315558` cycles	`1.00`
`ML-DSA-65 verify`	`110980` cycles	`111490` cycles	`1.00`
`ML-DSA-87 keypair`	`170913` cycles	`171547` cycles	`1.00`
`ML-DSA-87 sign`	`378782` cycles	`378591` cycles	`1.00`
`ML-DSA-87 verify`	`169723` cycles	`169379` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Intel Xeon 3rd gen (c6i) (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`154749` cycles	`154402` cycles	`1.00`
`ML-DSA-44 sign`	`587983` cycles	`589538` cycles	`1.00`
`ML-DSA-44 verify`	`168810` cycles	`169828` cycles	`0.99`
`ML-DSA-65 keypair`	`262287` cycles	`263339` cycles	`1.00`
`ML-DSA-65 sign`	`957036` cycles	`966082` cycles	`0.99`
`ML-DSA-65 verify`	`271611` cycles	`272946` cycles	`1.00`
`ML-DSA-87 keypair`	`432882` cycles	`432716` cycles	`1.00`
`ML-DSA-87 sign`	`1209381` cycles	`1216186` cycles	`0.99`
`ML-DSA-87 verify`	`447652` cycles	`447711` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton3

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`71369` cycles	`71552` cycles	`1.00`
`ML-DSA-44 sign`	`208953` cycles	`208994` cycles	`1.00`
`ML-DSA-44 verify`	`74780` cycles	`74736` cycles	`1.00`
`ML-DSA-65 keypair`	`125947` cycles	`125905` cycles	`1.00`
`ML-DSA-65 sign`	`345638` cycles	`345393` cycles	`1.00`
`ML-DSA-65 verify`	`124113` cycles	`124212` cycles	`1.00`
`ML-DSA-87 keypair`	`207065` cycles	`206527` cycles	`1.00`
`ML-DSA-87 sign`	`444028` cycles	`439858` cycles	`1.01`
`ML-DSA-87 verify`	`204117` cycles	`204429` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`137911` cycles	`138034` cycles	`1.00`
`ML-DSA-44 sign`	`482215` cycles	`486060` cycles	`0.99`
`ML-DSA-44 verify`	`148228` cycles	`149072` cycles	`0.99`
`ML-DSA-65 keypair`	`240961` cycles	`241789` cycles	`1.00`
`ML-DSA-65 sign`	`784814` cycles	`791605` cycles	`0.99`
`ML-DSA-65 verify`	`240936` cycles	`241325` cycles	`1.00`
`ML-DSA-87 keypair`	`395872` cycles	`396314` cycles	`1.00`
`ML-DSA-87 sign`	`1006622` cycles	`1019348` cycles	`0.99`
`ML-DSA-87 verify`	`402277` cycles	`403755` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Arm Cortex-A76 (Raspberry Pi 5) benchmarks (opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`112170` cycles	`112136` cycles	`1.00`
`ML-DSA-44 sign`	`353500` cycles	`353794` cycles	`1.00`
`ML-DSA-44 verify`	`117008` cycles	`117198` cycles	`1.00`
`ML-DSA-65 keypair`	`194777` cycles	`194374` cycles	`1.00`
`ML-DSA-65 sign`	`583911` cycles	`583728` cycles	`1.00`
`ML-DSA-65 verify`	`192722` cycles	`193104` cycles	`1.00`
`ML-DSA-87 keypair`	`320921` cycles	`320068` cycles	`1.00`
`ML-DSA-87 sign`	`747350` cycles	`747194` cycles	`1.00`
`ML-DSA-87 verify`	`318786` cycles	`317899` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Arm Cortex-A76 (Raspberry Pi 5) benchmarks (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`211603` cycles	`211623` cycles	`1.00`
`ML-DSA-44 sign`	`751832` cycles	`760075` cycles	`0.99`
`ML-DSA-44 verify`	`227491` cycles	`229458` cycles	`0.99`
`ML-DSA-65 keypair`	`375371` cycles	`378162` cycles	`0.99`
`ML-DSA-65 sign`	`1232745` cycles	`1247139` cycles	`0.99`
`ML-DSA-65 verify`	`369289` cycles	`371939` cycles	`0.99`
`ML-DSA-87 keypair`	`600293` cycles	`601521` cycles	`1.00`
`ML-DSA-87 sign`	`1568166` cycles	`1582092` cycles	`0.99`
`ML-DSA-87 verify`	`613423` cycles	`617517` cycles	`0.99`

This comment was automatically generated by workflow using github-action-benchmark.

hanno-becker

This looks right to me.

github-actions

Arm Cortex-A55 (Snapdragon 888) benchmarks (opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `0f9bd57`	Ratio
`ML-DSA-44 keypair`	`268375` cycles	`267651` cycles	`1.00`
`ML-DSA-44 sign`	`809831` cycles	`811144` cycles	`1.00`
`ML-DSA-44 verify`	`269969` cycles	`270460` cycles	`1.00`
`ML-DSA-65 keypair`	`459702` cycles	`459568` cycles	`1.00`
`ML-DSA-65 sign`	`1315458` cycles	`1315519` cycles	`1.00`
`ML-DSA-65 verify`	`445789` cycles	`445438` cycles	`1.00`
`ML-DSA-87 keypair`	`796724` cycles	`786888` cycles	`1.01`
`ML-DSA-87 sign`	`1805086` cycles	`1790709` cycles	`1.01`
`ML-DSA-87 verify`	`777883` cycles	`767214` cycles	`1.01`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Arm Cortex-A55 (Snapdragon 888) benchmarks (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `0f9bd57`	Ratio
`ML-DSA-44 keypair`	`463089` cycles	`462914` cycles	`1.00`
`ML-DSA-44 sign`	`2118538` cycles	`2132089` cycles	`0.99`
`ML-DSA-44 verify`	`550906` cycles	`554284` cycles	`0.99`
`ML-DSA-65 keypair`	`781134` cycles	`780479` cycles	`1.00`
`ML-DSA-65 sign`	`3454267` cycles	`3479571` cycles	`0.99`
`ML-DSA-65 verify`	`860012` cycles	`864154` cycles	`1.00`
`ML-DSA-87 keypair`	`1261699` cycles	`1265342` cycles	`1.00`
`ML-DSA-87 sign`	`4274189` cycles	`4320477` cycles	`0.99`
`ML-DSA-87 verify`	`1380932` cycles	`1388614` cycles	`0.99`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

SpacemiT K1 8 (Banana Pi F3) benchmarks (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`760536` cycles	`760114` cycles	`1.00`
`ML-DSA-44 sign`	`3115896` cycles	`3141235` cycles	`0.99`
`ML-DSA-44 verify`	`855663` cycles	`859284` cycles	`1.00`
`ML-DSA-65 keypair`	`1285920` cycles	`1285249` cycles	`1.00`
`ML-DSA-65 sign`	`5085189` cycles	`5074895` cycles	`1.00`
`ML-DSA-65 verify`	`1364204` cycles	`1363443` cycles	`1.00`
`ML-DSA-87 keypair`	`2113049` cycles	`2114103` cycles	`1.00`
`ML-DSA-87 sign`	`6359782` cycles	`6362879` cycles	`1.00`
`ML-DSA-87 verify`	`2229223` cycles	`2230277` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Arm Cortex-A72 (Raspberry Pi 4) benchmarks (opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `ec15c59`	Ratio
`ML-DSA-44 keypair`	`223726` cycles	`221766` cycles	`1.01`
`ML-DSA-44 sign`	`613726` cycles	`614083` cycles	`1.00`
`ML-DSA-44 verify`	`218800` cycles	`233383` cycles	`0.94`
`ML-DSA-65 keypair`	`384926` cycles	`393803` cycles	`0.98`
`ML-DSA-65 sign`	`998355` cycles	`1023539` cycles	`0.98`
`ML-DSA-65 verify`	`368420` cycles	`377414` cycles	`0.98`
`ML-DSA-87 keypair`	`663434` cycles	`669482` cycles	`0.99`
`ML-DSA-87 sign`	`1356972` cycles	`1392850` cycles	`0.97`
`ML-DSA-87 verify`	`644004` cycles	`643114` cycles	`1.00`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

Arm Cortex-A72 (Raspberry Pi 4) benchmarks (no-opt)

Details

Benchmark suite	Current: `9bb8f8d`	Previous: `0f9bd57`	Ratio
`ML-DSA-44 keypair`	`302274` cycles	`301193` cycles	`1.00`
`ML-DSA-44 sign`	`1139196` cycles	`1138793` cycles	`1.00`
`ML-DSA-44 verify`	`322598` cycles	`328764` cycles	`0.98`
`ML-DSA-65 keypair`	`558400` cycles	`549022` cycles	`1.02`
`ML-DSA-65 sign`	`1865495` cycles	`1904832` cycles	`0.98`
`ML-DSA-65 verify`	`547034` cycles	`529988` cycles	`1.03`
`ML-DSA-87 keypair`	`890507` cycles	`850322` cycles	`1.05`
`ML-DSA-87 sign`	`2420330` cycles	`2373299` cycles	`1.02`
`ML-DSA-87 verify`	`899250` cycles	`877505` cycles	`1.02`

This comment was automatically generated by workflow using github-action-benchmark.

github-actions

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Arm Cortex-A72 (Raspberry Pi 4) benchmarks (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `9bb8f8d`	Previous: `0f9bd57`	Ratio
`ML-DSA-65 verify`	`547034` cycles	`529988` cycles	`1.03`
`ML-DSA-87 keypair`	`890507` cycles	`850322` cycles	`1.05`

This comment was automatically generated by workflow using github-action-benchmark.

Replace the two-step Barrett division (ceil(a/128) then Barrett-divide by 2*GAMMA2/128) with a direct high multiplication by floor(2^N / 2*GAMMA2), mirroring the AArch64 backend. For ML-DSA-44 this is `(a * 1477838209 + 2^47) >> 48`; for ML-DSA-65/87 it is `(a * 1074791425 + 2^48) >> 49`. Both constants strictly under-approximate 1/(2*GAMMA2), so half-points round down, matching the original round-half-down semantics, and the result is exact for all 0 <= a < Q. This collapses a five-op dependency chain (add, asr, mul, add, asr) into a single signed multiply-add and one shift. On Graviton2 the scalar a1 step is ~31% faster (1.39 ns -> 0.96 ns per call) for both parameter sets; the loop is scalar in practice because the ct value-barriers in mld_decompose block auto-vectorization. AArch64 and x86_64 builds use their native asm backends and are unaffected. Update the Isabelle attribution in compress/ML-DSA_Compress.thy and neon_ntt/Barrett_Division_Even.thy to group the C reference with the AArch64 backend (direct Barrett division) rather than with AVX2 (divide-by-128 first). The corollary names keep their `aarch64` suffix for stability; a note in the prose explains they cover the C reference as well. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

hanno-becker force-pushed the mjk/issue-652 branch from c5826af to 9bb8f8d Compare June 14, 2026 05:47

hanno-becker added the benchmark label Jun 14, 2026

github-actions Bot reviewed Jun 14, 2026

View reviewed changes

oqs-bot reviewed Jun 14, 2026

View reviewed changes

github-actions Bot reviewed Jun 14, 2026

View reviewed changes

hanno-becker approved these changes Jun 14, 2026

View reviewed changes

github-actions Bot reviewed Jun 14, 2026

View reviewed changes

hanno-becker force-pushed the mjk/issue-652 branch from 9bb8f8d to e5044f8 Compare June 15, 2026 04:29

Conversation

mkannwischer commented Jun 14, 2026

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-65, REDUCE-RAM)

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-44, REDUCE-RAM)

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-87, REDUCE-RAM)

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-44)

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-65)

Uh oh!

oqs-bot commented Jun 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-DSA-87)

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Mac Mini (M1, 2020) benchmarks (opt)

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Mac Mini (M1, 2020) benchmarks (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Intel Xeon 4th gen (c7i)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Intel Xeon 4th gen (c7i) (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

AMD EPYC 3rd gen (c6a)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

AMD EPYC 3rd gen (c6a) (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton2

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

AMD EPYC 4th gen (c7a)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton4

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

AMD EPYC 4th gen (c7a) (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton2 (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton4 (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Intel Xeon 3rd gen (c6i)

Uh oh!

oqs-bot left a comment

oqs-bot commented Jun 14, 2026 •

edited

Loading

oqs-bot commented Jun 14, 2026 •

edited

Loading

oqs-bot commented Jun 14, 2026 •

edited

Loading

oqs-bot commented Jun 14, 2026 •

edited

Loading

oqs-bot commented Jun 14, 2026 •

edited

Loading

oqs-bot commented Jun 14, 2026 •

edited

Loading