x86_64 + HOL-Light: Replace poly_decompose AVX2 intrinsics with hand-written assembly and HOL-Light proofs by jakemas · Pull Request #1163 · pq-code-package/mldsa-native

jakemas · 2026-06-10T04:47:26Z

Resolves #420
Resolves #914

Performance

poly_decompose component benchmark, median cycles on AMD EPYC (c6a), OPT=1 CYCLES=PMU:

Variant	AVX2 intrinsics (baseline, `main`)	Hand-written AVX2 asm (this PR)
`decompose_32` (ML-DSA-65/87)	~1606	~1607
`decompose_88` (ML-DSA-44)	~1869	~1870

oqs-bot · 2026-06-10T05:57:16Z

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

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1570s	1507s	+4.2%
`mld_invntt_layer`	✅	171s	157s	+9%
`poly_pointwise_montgomery_c`	✅	131s	125s	+5%
`rej_uniform_native`	✅	128s	123s	+4%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	120s	112s	+7%
`mld_ct_memcmp`	✅	65s	63s	+3%
`mld_ntt_layer`	✅	45s	44s	+2%
`fqmul`	✅	44s	40s	+10%
`sign_verify_internal`	✅	25s	22s	+14%
`mld_attempt_signature_generation`	✅	24s	23s	+4%
`mld_ntt_butterfly_block`	✅	24s	20s	+20%
`keccakf1600x4_permute_native`	✅	23s	21s	+10%
`poly_chknorm_c`	✅	19s	19s	+0%
`polyt0_unpack`	✅	18s	15s	+20%
`rej_uniform_c`	✅	17s	14s	+21%
`mld_check_pct`	✅	14s	13s	+8%
`polyeta_unpack`	✅	14s	14s	+0%
`poly_uniform_eta_4x`	✅	13s	15s	-13%
`poly_add`	✅	11s	13s	-15%
`polyz_unpack_c`	✅	11s	10s	+10%
`compute_pack_t0_t1`	✅	10s	9s	+11%
`keccak_absorb_once_x4`	✅	10s	10s	+0%
`polyveck_chknorm`	✅	10s	11s	-9%
`polyvec_matrix_pointwise_montgomery_row`	✅	9s	6s	+50%
`polyveck_decompose`	✅	8s	7s	+14%
`mld_keccakf1600_permute_c`	✅	7s	6s	+17%
`pointwise_acc_native_x86_64`	✅	7s	4s	+75%
`poly_invntt_tomont_c`	✅	7s	9s	-22%
`rej_uniform`	✅	7s	9s	-22%
`sign`	✅	7s	6s	+17%
`keccak_absorb`	✅	6s	5s	+20%
`keccakf1600_xor_bytes (big endian)`	✅	6s	5s	+20%
`mld_compute_pack_z`	✅	6s	7s	-14%
`mld_sample_s1_s2`	✅	6s	2s	+200%
`poly_challenge`	✅	6s	2s	+200%
`poly_decompose_c`	✅	6s	7s	-14%
`poly_power2round`	✅	6s	6s	+0%
`poly_uniform_eta`	✅	6s	3s	+100%
`polyveck_invntt_tomont`	✅	6s	5s	+20%
`polyvecl_chknorm`	✅	6s	5s	+20%
`keccakf1600x4_permute`	✅	5s	3s	+67%
`keccakf1600x4_xor_bytes_native`	✅	5s	3s	+67%
`mld_h`	✅	5s	3s	+67%
`ntt_native_x86_64`	✅	5s	4s	+25%
`pack_sig_z`	✅	5s	2s	+150%
`pointwise_acc_native_aarch64`	✅	5s	4s	+25%
`poly_permute_bitrev_to_custom_optional`	✅	5s	2s	+150%
`poly_reduce`	✅	5s	6s	-17%
`polyveck_caddq`	✅	5s	5s	+0%
`polyvecl_pack_eta`	✅	5s	3s	+67%
`polyvecl_pointwise_acc_montgomery_native`	✅	5s	4s	+25%
`polyz_pack`	✅	5s	3s	+67%
`rej_eta_c`	✅	5s	3s	+67%
`shake256_absorb`	✅	5s	2s	+150%
`sign_open`	✅	5s	7s	-29%
`sign_signature_pre_hash_internal`	✅	5s	4s	+25%
`decompose`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	4s	3s	+33%
`keccak_f1600_x1_native_aarch64`	✅	4s	1s	+300%
`keccak_f1600_x1_native_aarch64_v84a`	✅	4s	3s	+33%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	4s	+0%
`keccak_squeezeblocks_x4`	✅	4s	5s	-20%
`mld_prepare_domain_separation_prefix`	✅	4s	3s	+33%
`mld_sample_s1_s2_serial`	✅	4s	3s	+33%
`mld_value_barrier_u32`	✅	4s	2s	+100%
`pointwise_native_aarch64`	✅	4s	3s	+33%
`poly_caddq_c`	✅	4s	5s	-20%
`poly_caddq_native_aarch64`	✅	4s	3s	+33%
`poly_chknorm_native`	✅	4s	6s	-33%
`poly_chknorm_native_aarch64`	✅	4s	1s	+300%
`poly_decompose`	✅	4s	4s	+0%
`poly_invntt_tomont_native`	✅	4s	2s	+100%
`poly_ntt_c`	✅	4s	3s	+33%
`poly_ntt_native`	✅	4s	4s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	4s	5s	-20%
`poly_pointwise_montgomery`	✅	4s	5s	-20%
`poly_pointwise_montgomery_native`	✅	4s	3s	+33%
`poly_uniform`	✅	4s	4s	+0%
`poly_uniform_4x`	✅	4s	3s	+33%
`poly_use_hint`	✅	4s	2s	+100%
`poly_use_hint_native`	✅	4s	4s	+0%
`poly_use_hint_native_aarch64`	✅	4s	3s	+33%
`polyt1_pack`	✅	4s	2s	+100%
`polyt1_unpack`	✅	4s	4s	+0%
`polyvec_matrix_expand`	✅	4s	3s	+33%
`polyvecl_pointwise_acc_montgomery`	✅	4s	2s	+100%
`polyvecl_uniform_gamma1`	✅	4s	2s	+100%
`polyvecl_unpack_eta`	✅	4s	4s	+0%
`polyz_unpack_17_native_aarch64`	✅	4s	2s	+100%
`rej_eta`	✅	4s	4s	+0%
`shake128x4_squeezeblocks`	✅	4s	3s	+33%
`sign_keypair_internal`	✅	4s	4s	+0%
`sign_pk_from_sk`	✅	4s	6s	-33%
`sign_signature_internal`	✅	4s	5s	-20%
`sign_signature_pre_hash_shake256`	✅	4s	2s	+100%
`sign_verify_extmu`	✅	4s	3s	+33%
`sign_verify_pre_hash_internal`	✅	4s	3s	+33%
`sign_verify_pre_hash_shake256`	✅	4s	4s	+0%
`sk_s1hat_get_poly`	✅	4s	2s	+100%
`sk_s2hat_get_poly`	✅	4s	4s	+0%
`sys_check_capability`	✅	4s	2s	+100%
`unpack_sk_s1hat`	✅	4s	3s	+33%
`unpack_sk_s2hat`	✅	4s	4s	+0%
`yvec_get_poly`	✅	4s	5s	-20%
`fqscale`	✅	3s	5s	-40%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	2s	+50%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`keccak_finalize`	✅	3s	4s	-25%
`keccak_init`	✅	3s	2s	+50%
`keccakf1600_permute_native`	✅	3s	2s	+50%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`mld_ct_cmask_nonzero_u32`	✅	3s	3s	+0%
`mld_polymat_expand_entry`	✅	3s	3s	+0%
`nttunpack_native_x86_64`	✅	3s	3s	+0%
`pack_sig_c`	✅	3s	3s	+0%
`pack_sig_h`	✅	3s	2s	+50%
`pack_sk_rho_key_tr_s2`	✅	3s	3s	+0%
`poly_caddq`	✅	3s	3s	+0%
`poly_caddq_native`	✅	3s	6s	-50%
`poly_caddq_native_x86_64`	✅	3s	4s	-25%
`poly_decompose_32_native_aarch64`	✅	3s	2s	+50%
`poly_decompose_native`	✅	3s	4s	-25%
`poly_ntt`	✅	3s	3s	+0%
`poly_shiftl`	✅	3s	5s	-40%
`poly_sub`	✅	3s	3s	+0%
`poly_uniform_gamma1`	✅	3s	2s	+50%
`poly_uniform_gamma1_4x`	✅	3s	5s	-40%
`poly_use_hint_c`	✅	3s	2s	+50%
`polyeta_pack`	✅	3s	2s	+50%
`polyt0_pack`	✅	3s	5s	-40%
`polyveck_pack_eta`	✅	3s	2s	+50%
`polyveck_reduce`	✅	3s	3s	+0%
`polyvecl_ntt`	✅	3s	5s	-40%
`polyvecl_unpack_z`	✅	3s	4s	-25%
`polyw1_pack`	✅	3s	3s	+0%
`power2round`	✅	3s	3s	+0%
`reduce32`	✅	3s	3s	+0%
`rej_eta_native`	✅	3s	4s	-25%
`rej_uniform_eta_native_aarch64`	✅	3s	3s	+0%
`shake128_absorb`	✅	3s	4s	-25%
`shake128_finalize`	✅	3s	3s	+0%
`shake256_init`	✅	3s	1s	+200%
`shake256_squeeze`	✅	3s	3s	+0%
`shake256x4_absorb_once`	✅	3s	2s	+50%
`sign_keypair`	✅	3s	6s	-50%
`sign_signature`	✅	3s	4s	-25%
`sign_signature_extmu`	✅	3s	4s	-25%
`unpack_pk_t1`	✅	3s	1s	+200%
`yvec_init`	✅	3s	3s	+0%
`caddq`	✅	2s	4s	-50%
`intt_native_x86_64`	✅	2s	3s	-33%
`keccak_squeeze`	✅	2s	3s	-33%
`keccakf1600_extract_bytes (big endian)`	✅	2s	5s	-60%
`keccakf1600_permute`	✅	2s	4s	-50%
`keccakf1600_xor_bytes`	✅	2s	4s	-50%
`keccakf1600x4_extract_bytes`	✅	2s	1s	+100%
`keccakf1600x4_xor_bytes`	✅	2s	2s	+0%
`make_hint`	✅	2s	2s	+0%
`mld_ct_abs_i32`	✅	2s	2s	+0%
`mld_ct_cmask_neg_i32`	✅	2s	2s	+0%
`mld_ct_cmask_nonzero_u8`	✅	2s	4s	-50%
`mld_ct_get_optblocker_u32`	✅	2s	1s	+100%
`mld_ct_sel_int32`	✅	2s	1s	+100%
`mld_keccakf1600_extract_bytes`	✅	2s	2s	+0%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	1s	+100%
`mld_keccakf1600x4_xor_bytes_c`	✅	2s	3s	-33%
`mld_value_barrier_i64`	✅	2s	1s	+100%
`mld_value_barrier_u8`	✅	2s	2s	+0%
`ntt_native_aarch64`	✅	2s	3s	-33%
`pack_sk_s1`	✅	2s	2s	+0%
`pointwise_native_x86_64`	✅	2s	6s	-67%
`poly_chknorm`	✅	2s	3s	-33%
`poly_chknorm_native_x86_64`	✅	2s	3s	-33%
`poly_decompose_88_native_aarch64`	✅	2s	3s	-33%
`poly_decompose_native_x86_64`	✅	2s	-	new
`poly_invntt_tomont`	✅	2s	3s	-33%
`polyvec_matrix_expand_serial`	✅	2s	3s	-33%
`polyveck_ntt`	✅	2s	1s	+100%
`polyveck_pack_w1`	✅	2s	2s	+0%
`polyvecl_uniform_gamma1_serial`	✅	2s	3s	-33%
`polyw1_pack_32`	✅	2s	3s	-33%
`polyz_unpack`	✅	2s	2s	+0%
`polyz_unpack_19_native_aarch64`	✅	2s	3s	-33%
`polyz_unpack_native`	✅	2s	2s	+0%
`rej_uniform_native_aarch64`	✅	2s	4s	-50%
`shake128_squeeze`	✅	2s	2s	+0%
`shake128x4_absorb_once`	✅	2s	3s	-33%
`shake256`	✅	2s	4s	-50%
`shake256_release`	✅	2s	1s	+100%
`sig_unpack_hints`	✅	2s	3s	-33%
`sign_verify`	✅	2s	7s	-71%
`sk_t0hat_get_poly`	✅	2s	3s	-33%
`unpack_sk_t0hat`	✅	2s	4s	-50%
`use_hint`	✅	2s	4s	-50%
`mld_ct_get_optblocker_i64`	✅	1s	2s	-50%
`mld_ct_get_optblocker_u8`	✅	1s	3s	-67%
`montgomery_reduce`	✅	1s	4s	-75%
`polyveck_unpack_eta`	✅	1s	4s	-75%
`polyvecl_pointwise_acc_montgomery_c`	✅	1s	5s	-80%
`polyw1_pack_88`	✅	1s	4s	-75%
`shake128_init`	✅	1s	3s	-67%
`shake128_release`	✅	1s	2s	-50%
`shake256_finalize`	✅	1s	2s	-50%
`shake256x4_squeezeblocks`	✅	1s	2s	-50%
`unpack_sk`	✅	1s	2s	-50%

oqs-bot · 2026-06-10T05:57:23Z

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

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1643s	1654s	-0.7%
`mld_invntt_layer`	✅	173s	179s	-3%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	157s	163s	-4%
`poly_pointwise_montgomery_c`	✅	133s	140s	-5%
`rej_uniform_native`	✅	128s	131s	-2%
`mld_ct_memcmp`	✅	68s	74s	-8%
`mld_ntt_layer`	✅	45s	44s	+2%
`fqmul`	✅	42s	43s	-2%
`mld_attempt_signature_generation`	✅	36s	35s	+3%
`mld_ntt_butterfly_block`	✅	23s	26s	-12%
`keccakf1600x4_permute_native`	✅	22s	22s	+0%
`sign_verify_internal`	✅	22s	22s	+0%
`poly_chknorm_c`	✅	20s	20s	+0%
`poly_uniform_eta_4x`	✅	18s	15s	+20%
`polyt0_unpack`	✅	18s	18s	+0%
`polyveck_decompose`	✅	16s	16s	+0%
`rej_uniform_c`	✅	16s	17s	-6%
`polyeta_unpack`	✅	15s	14s	+7%
`mld_check_pct`	✅	14s	16s	-12%
`compute_pack_t0_t1`	✅	12s	9s	+33%
`polyvecl_chknorm`	✅	11s	10s	+10%
`poly_add`	✅	10s	12s	-17%
`poly_invntt_tomont_c`	✅	10s	8s	+25%
`keccak_absorb_once_x4`	✅	9s	10s	-10%
`mld_prepare_domain_separation_prefix`	✅	9s	5s	+80%
`sign`	✅	9s	8s	+12%
`polyvecl_ntt`	✅	8s	8s	+0%
`keccak_absorb`	✅	7s	8s	-12%
`keccak_squeeze`	✅	7s	3s	+133%
`poly_caddq_c`	✅	7s	3s	+133%
`poly_power2round`	✅	7s	9s	-22%
`polyveck_caddq`	✅	7s	8s	-12%
`polyveck_invntt_tomont`	✅	7s	8s	-12%
`mld_keccakf1600_permute_c`	✅	6s	7s	-14%
`mld_sample_s1_s2`	✅	6s	5s	+20%
`pointwise_acc_native_aarch64`	✅	6s	7s	-14%
`pointwise_acc_native_x86_64`	✅	6s	8s	-25%
`poly_shiftl`	✅	6s	5s	+20%
`polyveck_chknorm`	✅	6s	5s	+20%
`polyz_unpack_c`	✅	6s	8s	-25%
`rej_uniform`	✅	6s	6s	+0%
`sign_keypair_internal`	✅	6s	3s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	5s	2s	+150%
`keccak_f1600_x4_native_avx2`	✅	5s	3s	+67%
`keccak_finalize`	✅	5s	3s	+67%
`keccak_squeezeblocks_x4`	✅	5s	5s	+0%
`mld_compute_pack_z`	✅	5s	5s	+0%
`mld_h`	✅	5s	6s	-17%
`poly_chknorm_native_aarch64`	✅	5s	4s	+25%
`poly_reduce`	✅	5s	5s	+0%
`poly_uniform_eta`	✅	5s	5s	+0%
`polyvec_matrix_expand_serial`	✅	5s	4s	+25%
`polyvec_matrix_pointwise_montgomery_row`	✅	5s	6s	-17%
`polyveck_reduce`	✅	5s	5s	+0%
`polyw1_pack`	✅	5s	4s	+25%
`power2round`	✅	5s	2s	+150%
`reduce32`	✅	5s	3s	+67%
`rej_eta_c`	✅	5s	2s	+150%
`sign_open`	✅	5s	3s	+67%
`sign_pk_from_sk`	✅	5s	4s	+25%
`sign_signature_internal`	✅	5s	7s	-29%
`yvec_get_poly`	✅	5s	4s	+25%
`keccak_f1600_x1_native_aarch64_v84a`	✅	4s	2s	+100%
`keccak_init`	✅	4s	4s	+0%
`keccakf1600_xor_bytes`	✅	4s	3s	+33%
`mld_ct_cmask_nonzero_u32`	✅	4s	3s	+33%
`mld_keccakf1600x4_xor_bytes_c`	✅	4s	1s	+300%
`mld_polymat_expand_entry`	✅	4s	3s	+33%
`mld_sample_s1_s2_serial`	✅	4s	6s	-33%
`ntt_native_aarch64`	✅	4s	4s	+0%
`ntt_native_x86_64`	✅	4s	4s	+0%
`poly_caddq_native_aarch64`	✅	4s	3s	+33%
`poly_challenge`	✅	4s	4s	+0%
`poly_chknorm_native`	✅	4s	3s	+33%
`poly_decompose_32_native_aarch64`	✅	4s	1s	+300%
`poly_decompose_88_native_aarch64`	✅	4s	4s	+0%
`poly_decompose_native`	✅	4s	3s	+33%
`poly_invntt_tomont`	✅	4s	2s	+100%
`poly_uniform_4x`	✅	4s	2s	+100%
`poly_use_hint_c`	✅	4s	4s	+0%
`polyveck_pack_eta`	✅	4s	4s	+0%
`polyveck_pack_w1`	✅	4s	3s	+33%
`polyvecl_pointwise_acc_montgomery`	✅	4s	4s	+0%
`polyvecl_pointwise_acc_montgomery_c`	✅	4s	1s	+300%
`polyvecl_pointwise_acc_montgomery_native`	✅	4s	3s	+33%
`polyvecl_uniform_gamma1_serial`	✅	4s	2s	+100%
`polyvecl_unpack_z`	✅	4s	2s	+100%
`polyw1_pack_32`	✅	4s	1s	+300%
`polyz_pack`	✅	4s	3s	+33%
`polyz_unpack_native`	✅	4s	2s	+100%
`rej_eta`	✅	4s	3s	+33%
`shake256x4_squeezeblocks`	✅	4s	2s	+100%
`sign_keypair`	✅	4s	2s	+100%
`sign_signature_pre_hash_shake256`	✅	4s	5s	-20%
`sign_verify`	✅	4s	4s	+0%
`sign_verify_pre_hash_internal`	✅	4s	3s	+33%
`sign_verify_pre_hash_shake256`	✅	4s	2s	+100%
`sys_check_capability`	✅	4s	5s	-20%
`unpack_sk`	✅	4s	2s	+100%
`unpack_sk_s2hat`	✅	4s	3s	+33%
`decompose`	✅	3s	1s	+200%
`intt_native_aarch64`	✅	3s	2s	+50%
`intt_native_x86_64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	3s	2s	+50%
`keccakf1600x4_extract_bytes`	✅	3s	3s	+0%
`keccakf1600x4_extract_bytes_native`	✅	3s	1s	+200%
`keccakf1600x4_xor_bytes`	✅	3s	4s	-25%
`keccakf1600x4_xor_bytes_native`	✅	3s	2s	+50%
`mld_ct_cmask_nonzero_u8`	✅	3s	3s	+0%
`mld_ct_get_optblocker_i64`	✅	3s	4s	-25%
`mld_ct_get_optblocker_u8`	✅	3s	3s	+0%
`mld_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mld_value_barrier_u32`	✅	3s	2s	+50%
`mld_value_barrier_u8`	✅	3s	2s	+50%
`pack_sig_c`	✅	3s	5s	-40%
`pack_sig_h`	✅	3s	4s	-25%
`pack_sk_rho_key_tr_s2`	✅	3s	1s	+200%
`pack_sk_s1`	✅	3s	3s	+0%
`pointwise_native_x86_64`	✅	3s	4s	-25%
`poly_caddq_native`	✅	3s	2s	+50%
`poly_chknorm`	✅	3s	5s	-40%
`poly_decompose`	✅	3s	2s	+50%
`poly_decompose_native_x86_64`	✅	3s	-	new
`poly_invntt_tomont_native`	✅	3s	2s	+50%
`poly_ntt`	✅	3s	3s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	3s	2s	+50%
`poly_pointwise_montgomery_native`	✅	3s	3s	+0%
`poly_sub`	✅	3s	2s	+50%
`poly_uniform_gamma1`	✅	3s	2s	+50%
`poly_use_hint_native`	✅	3s	5s	-40%
`polyt0_pack`	✅	3s	4s	-25%
`polyveck_ntt`	✅	3s	5s	-40%
`polyveck_unpack_eta`	✅	3s	4s	-25%
`polyvecl_uniform_gamma1`	✅	3s	4s	-25%
`polyw1_pack_88`	✅	3s	2s	+50%
`polyz_unpack`	✅	3s	2s	+50%
`polyz_unpack_17_native_aarch64`	✅	3s	3s	+0%
`rej_eta_native`	✅	3s	4s	-25%
`rej_uniform_eta_native_aarch64`	✅	3s	5s	-40%
`rej_uniform_native_aarch64`	✅	3s	3s	+0%
`shake256_absorb`	✅	3s	3s	+0%
`sign_signature_extmu`	✅	3s	4s	-25%
`sign_signature_pre_hash_internal`	✅	3s	6s	-50%
`sign_verify_extmu`	✅	3s	5s	-40%
`sk_t0hat_get_poly`	✅	3s	1s	+200%
`unpack_pk_t1`	✅	3s	4s	-25%
`unpack_sk_s1hat`	✅	3s	5s	-40%
`unpack_sk_t0hat`	✅	3s	2s	+50%
`use_hint`	✅	3s	1s	+200%
`yvec_init`	✅	3s	2s	+50%
`caddq`	✅	2s	4s	-50%
`fqscale`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	1s	+100%
`keccakf1600_permute`	✅	2s	1s	+100%
`keccakf1600_permute_native`	✅	2s	3s	-33%
`keccakf1600x4_permute`	✅	2s	2s	+0%
`make_hint`	✅	2s	2s	+0%
`mld_ct_abs_i32`	✅	2s	2s	+0%
`mld_ct_cmask_neg_i32`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mld_ct_sel_int32`	✅	2s	1s	+100%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mld_value_barrier_i64`	✅	2s	4s	-50%
`montgomery_reduce`	✅	2s	3s	-33%
`pack_sig_z`	✅	2s	4s	-50%
`pointwise_native_aarch64`	✅	2s	3s	-33%
`poly_caddq`	✅	2s	3s	-33%
`poly_caddq_native_x86_64`	✅	2s	3s	-33%
`poly_chknorm_native_x86_64`	✅	2s	3s	-33%
`poly_decompose_c`	✅	2s	6s	-67%
`poly_ntt_c`	✅	2s	4s	-50%
`poly_ntt_native`	✅	2s	3s	-33%
`poly_permute_bitrev_to_custom_optional`	✅	2s	4s	-50%
`poly_pointwise_montgomery`	✅	2s	4s	-50%
`poly_uniform`	✅	2s	4s	-50%
`poly_use_hint`	✅	2s	4s	-50%
`poly_use_hint_native_aarch64`	✅	2s	3s	-33%
`polyt1_pack`	✅	2s	3s	-33%
`polyt1_unpack`	✅	2s	2s	+0%
`polyvec_matrix_expand`	✅	2s	3s	-33%
`polyvecl_pack_eta`	✅	2s	3s	-33%
`polyz_unpack_19_native_aarch64`	✅	2s	2s	+0%
`shake128_squeeze`	✅	2s	1s	+100%
`shake128x4_squeezeblocks`	✅	2s	1s	+100%
`shake256`	✅	2s	4s	-50%
`shake256_finalize`	✅	2s	4s	-50%
`shake256_init`	✅	2s	2s	+0%
`shake256_release`	✅	2s	4s	-50%
`shake256_squeeze`	✅	2s	3s	-33%
`shake256x4_absorb_once`	✅	2s	2s	+0%
`sig_unpack_hints`	✅	2s	4s	-50%
`sign_signature`	✅	2s	3s	-33%
`sk_s1hat_get_poly`	✅	2s	2s	+0%
`sk_s2hat_get_poly`	✅	2s	2s	+0%
`keccakf1600_extract_bytes (big endian)`	✅	1s	2s	-50%
`keccakf1600_xor_bytes (big endian)`	✅	1s	3s	-67%
`nttunpack_native_x86_64`	✅	1s	3s	-67%
`poly_uniform_gamma1_4x`	✅	1s	4s	-75%
`polyeta_pack`	✅	1s	3s	-67%
`polyvecl_unpack_eta`	✅	1s	3s	-67%
`shake128_absorb`	✅	1s	3s	-67%
`shake128_finalize`	✅	1s	2s	-50%
`shake128_init`	✅	1s	2s	-50%
`shake128_release`	✅	1s	2s	-50%
`shake128x4_absorb_once`	✅	1s	1s	+0%

oqs-bot · 2026-06-10T05:57:44Z

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

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1503s	1601s	-6.1%
`mld_invntt_layer`	✅	156s	181s	-14%
`poly_pointwise_montgomery_c`	✅	131s	140s	-6%
`rej_uniform_native`	✅	122s	133s	-8%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	81s	85s	-5%
`mld_ct_memcmp`	✅	63s	73s	-14%
`mld_ntt_layer`	✅	45s	44s	+2%
`fqmul`	✅	40s	44s	-9%
`polyveck_chknorm`	✅	38s	40s	-5%
`keccakf1600x4_permute_native`	✅	23s	23s	+0%
`mld_attempt_signature_generation`	✅	22s	25s	-12%
`mld_ntt_butterfly_block`	✅	21s	25s	-16%
`poly_chknorm_c`	✅	16s	21s	-24%
`polyvecl_chknorm`	✅	15s	16s	-6%
`sign_verify_internal`	✅	15s	15s	+0%
`mld_check_pct`	✅	14s	14s	+0%
`polyt0_unpack`	✅	14s	16s	-12%
`polyveck_decompose`	✅	14s	16s	-12%
`rej_uniform_c`	✅	14s	19s	-26%
`poly_add`	✅	11s	11s	+0%
`poly_uniform_eta_4x`	✅	11s	14s	-21%
`polyvec_matrix_pointwise_montgomery_row`	✅	10s	8s	+25%
`pointwise_acc_native_x86_64`	✅	8s	7s	+14%
`polyeta_unpack`	✅	8s	7s	+14%
`polyvecl_ntt`	✅	8s	9s	-11%
`rej_uniform`	✅	8s	9s	-11%
`keccak_absorb`	✅	7s	7s	+0%
`keccak_absorb_once_x4`	✅	7s	9s	-22%
`mld_h`	✅	7s	4s	+75%
`mld_keccakf1600_permute_c`	✅	7s	7s	+0%
`poly_invntt_tomont_c`	✅	7s	11s	-36%
`sign`	✅	7s	7s	+0%
`sign_pk_from_sk`	✅	7s	5s	+40%
`sign_signature_pre_hash_internal`	✅	7s	5s	+40%
`compute_pack_t0_t1`	✅	6s	6s	+0%
`mld_sample_s1_s2_serial`	✅	6s	4s	+50%
`pointwise_acc_native_aarch64`	✅	6s	8s	-25%
`poly_reduce`	✅	6s	3s	+100%
`poly_uniform`	✅	6s	3s	+100%
`poly_use_hint_c`	✅	6s	3s	+100%
`polyveck_reduce`	✅	6s	4s	+50%
`polyvecl_pointwise_acc_montgomery_native`	✅	6s	3s	+100%
`polyz_unpack_c`	✅	6s	5s	+20%
`sign_keypair_internal`	✅	6s	2s	+200%
`sign_open`	✅	6s	5s	+20%
`mld_compute_pack_z`	✅	5s	6s	-17%
`mld_sample_s1_s2`	✅	5s	6s	-17%
`ntt_native_x86_64`	✅	5s	6s	-17%
`pointwise_native_x86_64`	✅	5s	2s	+150%
`poly_chknorm`	✅	5s	3s	+67%
`poly_power2round`	✅	5s	7s	-29%
`poly_uniform_gamma1`	✅	5s	3s	+67%
`polyveck_caddq`	✅	5s	7s	-29%
`polyveck_unpack_eta`	✅	5s	2s	+150%
`polyvecl_uniform_gamma1_serial`	✅	5s	1s	+400%
`polyz_unpack_19_native_aarch64`	✅	5s	2s	+150%
`rej_eta_c`	✅	5s	3s	+67%
`rej_uniform_eta_native_aarch64`	✅	5s	3s	+67%
`shake128_finalize`	✅	5s	3s	+67%
`sign_verify`	✅	5s	3s	+67%
`caddq`	✅	4s	4s	+0%
`intt_native_aarch64`	✅	4s	2s	+100%
`intt_native_x86_64`	✅	4s	3s	+33%
`keccak_f1600_x1_native_aarch64_v84a`	✅	4s	2s	+100%
`keccak_finalize`	✅	4s	2s	+100%
`keccakf1600_permute_native`	✅	4s	1s	+300%
`keccakf1600x4_xor_bytes_native`	✅	4s	3s	+33%
`mld_polymat_expand_entry`	✅	4s	3s	+33%
`montgomery_reduce`	✅	4s	3s	+33%
`poly_decompose`	✅	4s	2s	+100%
`poly_decompose_32_native_aarch64`	✅	4s	3s	+33%
`poly_decompose_88_native_aarch64`	✅	4s	1s	+300%
`poly_decompose_c`	✅	4s	6s	-33%
`poly_decompose_native_x86_64`	✅	4s	-	new
`poly_ntt`	✅	4s	4s	+0%
`poly_ntt_c`	✅	4s	3s	+33%
`poly_pointwise_montgomery_native`	✅	4s	5s	-20%
`polyveck_invntt_tomont`	✅	4s	8s	-50%
`polyvecl_uniform_gamma1`	✅	4s	3s	+33%
`polyw1_pack_32`	✅	4s	1s	+300%
`polyw1_pack_88`	✅	4s	3s	+33%
`rej_eta_native`	✅	4s	5s	-20%
`rej_uniform_native_aarch64`	✅	4s	2s	+100%
`shake256`	✅	4s	2s	+100%
`sig_unpack_hints`	✅	4s	3s	+33%
`sign_signature_extmu`	✅	4s	3s	+33%
`sign_signature_internal`	✅	4s	5s	-20%
`sign_signature_pre_hash_shake256`	✅	4s	3s	+33%
`sign_verify_extmu`	✅	4s	4s	+0%
`sign_verify_pre_hash_internal`	✅	4s	5s	-20%
`sk_s2hat_get_poly`	✅	4s	4s	+0%
`unpack_sk_s2hat`	✅	4s	2s	+100%
`use_hint`	✅	4s	3s	+33%
`yvec_init`	✅	4s	1s	+300%
`decompose`	✅	3s	1s	+200%
`keccak_f1600_x1_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	5s	-40%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`keccak_init`	✅	3s	3s	+0%
`keccak_squeezeblocks_x4`	✅	3s	5s	-40%
`keccakf1600_extract_bytes (big endian)`	✅	3s	4s	-25%
`keccakf1600_xor_bytes (big endian)`	✅	3s	4s	-25%
`keccakf1600x4_permute`	✅	3s	3s	+0%
`mld_ct_abs_i32`	✅	3s	3s	+0%
`mld_ct_get_optblocker_i64`	✅	3s	3s	+0%
`mld_ct_sel_int32`	✅	3s	1s	+200%
`mld_prepare_domain_separation_prefix`	✅	3s	3s	+0%
`ntt_native_aarch64`	✅	3s	3s	+0%
`pack_sig_c`	✅	3s	4s	-25%
`pack_sig_h`	✅	3s	5s	-40%
`pack_sk_rho_key_tr_s2`	✅	3s	2s	+50%
`poly_caddq`	✅	3s	4s	-25%
`poly_caddq_c`	✅	3s	6s	-50%
`poly_caddq_native`	✅	3s	6s	-50%
`poly_caddq_native_aarch64`	✅	3s	3s	+0%
`poly_challenge`	✅	3s	3s	+0%
`poly_chknorm_native`	✅	3s	4s	-25%
`poly_chknorm_native_aarch64`	✅	3s	3s	+0%
`poly_chknorm_native_x86_64`	✅	3s	3s	+0%
`poly_decompose_native`	✅	3s	4s	-25%
`poly_invntt_tomont_native`	✅	3s	3s	+0%
`poly_permute_bitrev_to_custom_optional`	✅	3s	3s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	3s	4s	-25%
`poly_pointwise_montgomery`	✅	3s	2s	+50%
`poly_shiftl`	✅	3s	5s	-40%
`poly_sub`	✅	3s	6s	-50%
`poly_uniform_4x`	✅	3s	3s	+0%
`poly_uniform_eta`	✅	3s	4s	-25%
`poly_use_hint_native`	✅	3s	3s	+0%
`polyt1_pack`	✅	3s	3s	+0%
`polyveck_ntt`	✅	3s	2s	+50%
`polyvecl_pointwise_acc_montgomery`	✅	3s	4s	-25%
`polyvecl_pointwise_acc_montgomery_c`	✅	3s	3s	+0%
`polyvecl_unpack_eta`	✅	3s	2s	+50%
`polyvecl_unpack_z`	✅	3s	3s	+0%
`shake128_absorb`	✅	3s	3s	+0%
`shake128_release`	✅	3s	3s	+0%
`shake128x4_absorb_once`	✅	3s	1s	+200%
`shake256_absorb`	✅	3s	2s	+50%
`shake256_init`	✅	3s	4s	-25%
`sign_keypair`	✅	3s	10s	-70%
`sign_signature`	✅	3s	2s	+50%
`sk_s1hat_get_poly`	✅	3s	3s	+0%
`sk_t0hat_get_poly`	✅	3s	2s	+50%
`unpack_pk_t1`	✅	3s	4s	-25%
`unpack_sk`	✅	3s	2s	+50%
`unpack_sk_s1hat`	✅	3s	3s	+0%
`yvec_get_poly`	✅	3s	3s	+0%
`fqscale`	✅	2s	4s	-50%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`keccakf1600_permute`	✅	2s	2s	+0%
`keccakf1600x4_extract_bytes_native`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes`	✅	2s	3s	-33%
`mld_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u8`	✅	2s	1s	+100%
`mld_keccakf1600_extract_bytes`	✅	2s	4s	-50%
`mld_value_barrier_i64`	✅	2s	1s	+100%
`mld_value_barrier_u32`	✅	2s	3s	-33%
`mld_value_barrier_u8`	✅	2s	4s	-50%
`nttunpack_native_x86_64`	✅	2s	2s	+0%
`pack_sig_z`	✅	2s	4s	-50%
`pack_sk_s1`	✅	2s	2s	+0%
`pointwise_native_aarch64`	✅	2s	3s	-33%
`poly_caddq_native_x86_64`	✅	2s	4s	-50%
`poly_uniform_gamma1_4x`	✅	2s	2s	+0%
`poly_use_hint`	✅	2s	4s	-50%
`poly_use_hint_native_aarch64`	✅	2s	3s	-33%
`polyeta_pack`	✅	2s	3s	-33%
`polyt0_pack`	✅	2s	4s	-50%
`polyvec_matrix_expand_serial`	✅	2s	1s	+100%
`polyveck_pack_eta`	✅	2s	4s	-50%
`polyveck_pack_w1`	✅	2s	2s	+0%
`polyvecl_pack_eta`	✅	2s	1s	+100%
`polyw1_pack`	✅	2s	3s	-33%
`polyz_pack`	✅	2s	3s	-33%
`polyz_unpack_17_native_aarch64`	✅	2s	4s	-50%
`power2round`	✅	2s	3s	-33%
`reduce32`	✅	2s	3s	-33%
`rej_eta`	✅	2s	3s	-33%
`shake128_squeeze`	✅	2s	2s	+0%
`shake128x4_squeezeblocks`	✅	2s	3s	-33%
`shake256_release`	✅	2s	3s	-33%
`shake256x4_absorb_once`	✅	2s	1s	+100%
`shake256x4_squeezeblocks`	✅	2s	3s	-33%
`sign_verify_pre_hash_shake256`	✅	2s	8s	-75%
`sys_check_capability`	✅	2s	4s	-50%
`unpack_sk_t0hat`	✅	2s	3s	-33%
`keccak_squeeze`	✅	1s	1s	+0%
`keccakf1600_xor_bytes`	✅	1s	3s	-67%
`keccakf1600x4_extract_bytes`	✅	1s	2s	-50%
`make_hint`	✅	1s	2s	-50%
`mld_ct_cmask_neg_i32`	✅	1s	1s	+0%
`mld_ct_cmask_nonzero_u32`	✅	1s	3s	-67%
`mld_keccakf1600x4_extract_bytes_c`	✅	1s	5s	-80%
`mld_keccakf1600x4_xor_bytes_c`	✅	1s	2s	-50%
`poly_invntt_tomont`	✅	1s	5s	-80%
`poly_ntt_native`	✅	1s	2s	-50%
`polyt1_unpack`	✅	1s	3s	-67%
`polyvec_matrix_expand`	✅	1s	5s	-80%
`polyz_unpack`	✅	1s	3s	-67%
`polyz_unpack_native`	✅	1s	2s	-50%
`shake128_init`	✅	1s	2s	-50%
`shake256_finalize`	✅	1s	2s	-50%
`shake256_squeeze`	✅	1s	2s	-50%

oqs-bot · 2026-06-10T05:58:40Z

CBMC Results (ML-DSA-65)

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	2039s	2067s	-1.4%
`mld_invntt_layer`	✅	270s	296s	-9%
`polyvecl_pointwise_acc_montgomery_c`	✅	187s	205s	-9%
`rej_uniform_native`	✅	147s	145s	+1%
`polyvec_matrix_expand`	✅	130s	127s	+2%
`poly_pointwise_montgomery_c`	✅	95s	100s	-5%
`mld_ct_memcmp`	✅	67s	67s	+0%
`mld_attempt_signature_generation`	✅	63s	66s	-5%
`sign_verify_internal`	✅	63s	61s	+3%
`sign_signature_internal`	✅	46s	45s	+2%
`mld_ntt_layer`	✅	45s	44s	+2%
`fqmul`	✅	42s	41s	+2%
`polyvec_matrix_expand_serial`	✅	26s	26s	+0%
`keccakf1600x4_permute_native`	✅	24s	21s	+14%
`mld_ntt_butterfly_block`	✅	23s	20s	+15%
`rej_uniform`	✅	21s	21s	+0%
`poly_chknorm_c`	✅	18s	20s	-10%
`polyt0_unpack`	✅	16s	15s	+7%
`polyveck_decompose`	✅	16s	15s	+7%
`polyvecl_chknorm`	✅	15s	18s	-17%
`compute_pack_t0_t1`	✅	14s	15s	-7%
`poly_uniform_4x`	✅	13s	10s	+30%
`rej_uniform_c`	✅	13s	14s	-7%
`mld_check_pct`	✅	11s	11s	+0%
`poly_add`	✅	11s	10s	+10%
`poly_uniform_eta_4x`	✅	11s	11s	+0%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	11s	8s	+38%
`mld_compute_pack_z`	✅	9s	12s	-25%
`polyveck_chknorm`	✅	9s	10s	-10%
`polyveck_ntt`	✅	9s	8s	+12%
`polyvecl_ntt`	✅	9s	6s	+50%
`keccak_absorb_once_x4`	✅	8s	7s	+14%
`poly_invntt_tomont_c`	✅	8s	7s	+14%
`polyveck_invntt_tomont`	✅	8s	7s	+14%
`polyvecl_pointwise_acc_montgomery_native`	✅	8s	4s	+100%
`keccak_squeezeblocks_x4`	✅	7s	3s	+133%
`mld_h`	✅	7s	2s	+250%
`mld_keccakf1600_permute_c`	✅	7s	5s	+40%
`polyz_unpack_c`	✅	7s	5s	+40%
`sign_open`	✅	7s	3s	+133%
`sign_verify_pre_hash_shake256`	✅	7s	3s	+133%
`keccak_absorb`	✅	6s	5s	+20%
`mld_prepare_domain_separation_prefix`	✅	6s	5s	+20%
`pointwise_acc_native_aarch64`	✅	6s	6s	+0%
`pointwise_acc_native_x86_64`	✅	6s	6s	+0%
`poly_caddq_native_aarch64`	✅	6s	2s	+200%
`poly_challenge`	✅	6s	4s	+50%
`polyveck_caddq`	✅	6s	4s	+50%
`unpack_sk_t0hat`	✅	6s	4s	+50%
`yvec_get_poly`	✅	6s	4s	+50%
`intt_native_x86_64`	✅	5s	2s	+150%
`keccakf1600_permute`	✅	5s	2s	+150%
`mld_sample_s1_s2_serial`	✅	5s	3s	+67%
`poly_chknorm_native`	✅	5s	2s	+150%
`poly_ntt_c`	✅	5s	3s	+67%
`poly_permute_bitrev_to_custom_optional_native`	✅	5s	2s	+150%
`poly_uniform_eta`	✅	5s	4s	+25%
`poly_uniform_gamma1`	✅	5s	4s	+25%
`poly_uniform_gamma1_4x`	✅	5s	4s	+25%
`polyt0_pack`	✅	5s	3s	+67%
`polyt1_pack`	✅	5s	3s	+67%
`polyvecl_pointwise_acc_montgomery`	✅	5s	2s	+150%
`rej_eta`	✅	5s	3s	+67%
`rej_eta_native`	✅	5s	5s	+0%
`shake128_absorb`	✅	5s	3s	+67%
`shake256_init`	✅	5s	1s	+400%
`sign`	✅	5s	10s	-50%
`sign_keypair`	✅	5s	4s	+25%
`sign_pk_from_sk`	✅	5s	7s	-29%
`sign_verify`	✅	5s	4s	+25%
`sign_verify_pre_hash_internal`	✅	5s	4s	+25%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	4s	2s	+100%
`keccak_init`	✅	4s	3s	+33%
`keccakf1600_extract_bytes (big endian)`	✅	4s	4s	+0%
`keccakf1600_xor_bytes`	✅	4s	2s	+100%
`keccakf1600_xor_bytes (big endian)`	✅	4s	3s	+33%
`keccakf1600x4_extract_bytes_native`	✅	4s	4s	+0%
`keccakf1600x4_permute`	✅	4s	3s	+33%
`mld_sample_s1_s2`	✅	4s	7s	-43%
`montgomery_reduce`	✅	4s	3s	+33%
`ntt_native_aarch64`	✅	4s	3s	+33%
`pack_sig_h`	✅	4s	4s	+0%
`poly_chknorm_native_x86_64`	✅	4s	2s	+100%
`poly_decompose`	✅	4s	2s	+100%
`poly_decompose_32_native_aarch64`	✅	4s	6s	-33%
`poly_permute_bitrev_to_custom_optional`	✅	4s	3s	+33%
`poly_pointwise_montgomery`	✅	4s	2s	+100%
`poly_power2round`	✅	4s	3s	+33%
`poly_reduce`	✅	4s	2s	+100%
`poly_use_hint_native`	✅	4s	2s	+100%
`polyeta_unpack`	✅	4s	3s	+33%
`polyveck_unpack_eta`	✅	4s	3s	+33%
`polyvecl_uniform_gamma1_serial`	✅	4s	4s	+0%
`polyw1_pack`	✅	4s	3s	+33%
`polyz_unpack_17_native_aarch64`	✅	4s	3s	+33%
`polyz_unpack_native`	✅	4s	2s	+100%
`rej_uniform_native_aarch64`	✅	4s	5s	-20%
`shake128_release`	✅	4s	4s	+0%
`shake128x4_absorb_once`	✅	4s	5s	-20%
`sign_signature`	✅	4s	7s	-43%
`sign_signature_extmu`	✅	4s	2s	+100%
`decompose`	✅	3s	4s	-25%
`keccak_f1600_x1_native_aarch64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	3s	+0%
`keccak_f1600_x4_native_avx2`	✅	3s	5s	-40%
`keccakf1600_permute_native`	✅	3s	3s	+0%
`mld_ct_cmask_neg_i32`	✅	3s	2s	+50%
`mld_ct_cmask_nonzero_u32`	✅	3s	5s	-40%
`mld_ct_cmask_nonzero_u8`	✅	3s	4s	-25%
`mld_ct_sel_int32`	✅	3s	2s	+50%
`mld_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mld_polymat_expand_entry`	✅	3s	3s	+0%
`pack_sig_z`	✅	3s	3s	+0%
`pack_sk_rho_key_tr_s2`	✅	3s	2s	+50%
`pack_sk_s1`	✅	3s	3s	+0%
`poly_caddq`	✅	3s	4s	-25%
`poly_caddq_c`	✅	3s	8s	-62%
`poly_caddq_native`	✅	3s	2s	+50%
`poly_caddq_native_x86_64`	✅	3s	3s	+0%
`poly_chknorm`	✅	3s	4s	-25%
`poly_chknorm_native_aarch64`	✅	3s	5s	-40%
`poly_decompose_88_native_aarch64`	✅	3s	2s	+50%
`poly_decompose_c`	✅	3s	5s	-40%
`poly_decompose_native`	✅	3s	2s	+50%
`poly_invntt_tomont_native`	✅	3s	6s	-50%
`poly_ntt`	✅	3s	3s	+0%
`poly_shiftl`	✅	3s	3s	+0%
`polyt1_unpack`	✅	3s	2s	+50%
`polyvec_matrix_pointwise_montgomery_row`	✅	3s	2s	+50%
`polyveck_pack_eta`	✅	3s	4s	-25%
`polyveck_reduce`	✅	3s	2s	+50%
`polyw1_pack_32`	✅	3s	4s	-25%
`polyw1_pack_88`	✅	3s	4s	-25%
`polyz_pack`	✅	3s	3s	+0%
`polyz_unpack`	✅	3s	2s	+50%
`polyz_unpack_19_native_aarch64`	✅	3s	3s	+0%
`power2round`	✅	3s	4s	-25%
`reduce32`	✅	3s	4s	-25%
`rej_eta_c`	✅	3s	4s	-25%
`rej_uniform_eta_native_aarch64`	✅	3s	5s	-40%
`shake128_finalize`	✅	3s	5s	-40%
`shake256`	✅	3s	2s	+50%
`shake256_absorb`	✅	3s	2s	+50%
`shake256_squeeze`	✅	3s	3s	+0%
`sign_keypair_internal`	✅	3s	5s	-40%
`sign_signature_pre_hash_shake256`	✅	3s	2s	+50%
`sign_verify_extmu`	✅	3s	4s	-25%
`sk_s1hat_get_poly`	✅	3s	2s	+50%
`sk_t0hat_get_poly`	✅	3s	6s	-50%
`sys_check_capability`	✅	3s	3s	+0%
`unpack_pk_t1`	✅	3s	2s	+50%
`unpack_sk`	✅	3s	3s	+0%
`use_hint`	✅	3s	2s	+50%
`yvec_init`	✅	3s	5s	-40%
`caddq`	✅	2s	2s	+0%
`fqscale`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	3s	-33%
`keccak_finalize`	✅	2s	5s	-60%
`keccak_squeeze`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes`	✅	2s	5s	-60%
`keccakf1600x4_xor_bytes_native`	✅	2s	2s	+0%
`make_hint`	✅	2s	3s	-33%
`mld_ct_get_optblocker_i64`	✅	2s	2s	+0%
`mld_ct_get_optblocker_u8`	✅	2s	3s	-33%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mld_keccakf1600x4_xor_bytes_c`	✅	2s	3s	-33%
`mld_value_barrier_i64`	✅	2s	2s	+0%
`mld_value_barrier_u32`	✅	2s	2s	+0%
`mld_value_barrier_u8`	✅	2s	3s	-33%
`ntt_native_x86_64`	✅	2s	3s	-33%
`nttunpack_native_x86_64`	✅	2s	4s	-50%
`pack_sig_c`	✅	2s	2s	+0%
`pointwise_native_aarch64`	✅	2s	3s	-33%
`pointwise_native_x86_64`	✅	2s	2s	+0%
`poly_decompose_native_x86_64`	✅	2s	-	new
`poly_invntt_tomont`	✅	2s	4s	-50%
`poly_ntt_native`	✅	2s	3s	-33%
`poly_pointwise_montgomery_native`	✅	2s	4s	-50%
`poly_sub`	✅	2s	3s	-33%
`poly_uniform`	✅	2s	3s	-33%
`poly_use_hint`	✅	2s	4s	-50%
`poly_use_hint_c`	✅	2s	3s	-33%
`poly_use_hint_native_aarch64`	✅	2s	4s	-50%
`polyeta_pack`	✅	2s	2s	+0%
`polyvecl_pack_eta`	✅	2s	2s	+0%
`polyvecl_uniform_gamma1`	✅	2s	6s	-67%
`polyvecl_unpack_eta`	✅	2s	4s	-50%
`shake128_init`	✅	2s	2s	+0%
`shake128_squeeze`	✅	2s	4s	-50%
`shake256_release`	✅	2s	3s	-33%
`shake256x4_squeezeblocks`	✅	2s	2s	+0%
`sign_signature_pre_hash_internal`	✅	2s	5s	-60%
`sk_s2hat_get_poly`	✅	2s	3s	-33%
`unpack_sk_s1hat`	✅	2s	2s	+0%
`unpack_sk_s2hat`	✅	2s	3s	-33%
`intt_native_aarch64`	✅	1s	5s	-80%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	2s	-50%
`keccakf1600x4_extract_bytes`	✅	1s	2s	-50%
`mld_ct_abs_i32`	✅	1s	4s	-75%
`mld_ct_get_optblocker_u32`	✅	1s	1s	+0%
`polyveck_pack_w1`	✅	1s	3s	-67%
`polyvecl_unpack_z`	✅	1s	5s	-80%
`shake128x4_squeezeblocks`	✅	1s	2s	-50%
`shake256_finalize`	✅	1s	2s	-50%
`shake256x4_absorb_once`	✅	1s	4s	-75%
`sig_unpack_hints`	✅	1s	4s	-75%

oqs-bot · 2026-06-10T05:59:28Z

CBMC Results (ML-DSA-87)

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	2641s	2351s	+12.3%
`polyvecl_pointwise_acc_montgomery_c`	✅	377s	321s	+17%
`mld_invntt_layer`	✅	328s	276s	+19%
`polyvec_matrix_expand`	✅	251s	218s	+15%
`rej_uniform_native`	✅	161s	152s	+6%
`poly_pointwise_montgomery_c`	✅	114s	86s	+33%
`mld_attempt_signature_generation`	✅	109s	105s	+4%
`mld_ct_memcmp`	✅	78s	66s	+18%
`sign_signature_internal`	✅	68s	63s	+8%
`sign_verify_internal`	✅	62s	59s	+5%
`polyvec_matrix_expand_serial`	✅	50s	47s	+6%
`fqmul`	✅	47s	40s	+18%
`mld_ntt_layer`	✅	46s	42s	+10%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	33s	32s	+3%
`compute_pack_t0_t1`	✅	32s	34s	-6%
`rej_uniform`	✅	27s	22s	+23%
`poly_chknorm_c`	✅	26s	19s	+37%
`mld_ntt_butterfly_block`	✅	24s	21s	+14%
`keccakf1600x4_permute_native`	✅	23s	23s	+0%
`polyt0_unpack`	✅	18s	16s	+12%
`polyeta_unpack`	✅	17s	14s	+21%
`mld_check_pct`	✅	16s	13s	+23%
`rej_uniform_c`	✅	16s	11s	+45%
`poly_uniform_eta_4x`	✅	13s	11s	+18%
`poly_add`	✅	12s	12s	+0%
`poly_uniform_4x`	✅	11s	14s	-21%
`polyveck_invntt_tomont`	✅	11s	9s	+22%
`polyvecl_ntt`	✅	11s	10s	+10%
`poly_invntt_tomont_c`	✅	10s	8s	+25%
`polyveck_decompose`	✅	10s	11s	-9%
`keccak_absorb_once_x4`	✅	9s	10s	-10%
`sign`	✅	9s	8s	+12%
`mld_compute_pack_z`	✅	8s	8s	+0%
`mld_sample_s1_s2`	✅	8s	3s	+167%
`mld_sample_s1_s2_serial`	✅	8s	6s	+33%
`poly_decompose_c`	✅	8s	5s	+60%
`polyveck_caddq`	✅	8s	8s	+0%
`keccak_absorb`	✅	7s	6s	+17%
`mld_keccakf1600_permute_c`	✅	7s	8s	-12%
`mld_value_barrier_i64`	✅	7s	1s	+600%
`pointwise_acc_native_aarch64`	✅	7s	7s	+0%
`pointwise_acc_native_x86_64`	✅	7s	7s	+0%
`sign_pk_from_sk`	✅	7s	4s	+75%
`sign_verify_extmu`	✅	7s	3s	+133%
`sign_verify_pre_hash_internal`	✅	7s	7s	+0%
`sign_verify_pre_hash_shake256`	✅	7s	3s	+133%
`keccak_squeezeblocks_x4`	✅	6s	4s	+50%
`poly_decompose`	✅	6s	3s	+100%
`poly_pointwise_montgomery`	✅	6s	4s	+50%
`polyveck_chknorm`	✅	6s	5s	+20%
`polyveck_ntt`	✅	6s	9s	-33%
`polyw1_pack_32`	✅	6s	3s	+100%
`polyz_unpack_c`	✅	6s	6s	+0%
`keccakf1600x4_permute`	✅	5s	4s	+25%
`mld_ct_cmask_nonzero_u32`	✅	5s	5s	+0%
`ntt_native_x86_64`	✅	5s	3s	+67%
`pack_sig_c`	✅	5s	2s	+150%
`poly_caddq_c`	✅	5s	5s	+0%
`poly_shiftl`	✅	5s	2s	+150%
`poly_use_hint_c`	✅	5s	2s	+150%
`polyt1_unpack`	✅	5s	1s	+400%
`polyveck_pack_eta`	✅	5s	2s	+150%
`polyveck_pack_w1`	✅	5s	3s	+67%
`polyvecl_pointwise_acc_montgomery_native`	✅	5s	5s	+0%
`rej_eta_native`	✅	5s	3s	+67%
`rej_uniform_native_aarch64`	✅	5s	2s	+150%
`shake256`	✅	5s	2s	+150%
`shake256_init`	✅	5s	2s	+150%
`sign_keypair_internal`	✅	5s	4s	+25%
`sign_open`	✅	5s	5s	+0%
`sign_signature_extmu`	✅	5s	5s	+0%
`unpack_sk`	✅	5s	3s	+67%
`decompose`	✅	4s	1s	+300%
`make_hint`	✅	4s	3s	+33%
`mld_ct_cmask_nonzero_u8`	✅	4s	1s	+300%
`mld_h`	✅	4s	3s	+33%
`mld_keccakf1600x4_extract_bytes_c`	✅	4s	2s	+100%
`pack_sk_rho_key_tr_s2`	✅	4s	1s	+300%
`pointwise_native_x86_64`	✅	4s	3s	+33%
`poly_caddq`	✅	4s	4s	+0%
`poly_caddq_native`	✅	4s	2s	+100%
`poly_caddq_native_aarch64`	✅	4s	3s	+33%
`poly_caddq_native_x86_64`	✅	4s	3s	+33%
`poly_challenge`	✅	4s	6s	-33%
`poly_chknorm`	✅	4s	3s	+33%
`poly_decompose_88_native_aarch64`	✅	4s	2s	+100%
`poly_invntt_tomont`	✅	4s	1s	+300%
`poly_invntt_tomont_native`	✅	4s	4s	+0%
`poly_ntt`	✅	4s	6s	-33%
`poly_sub`	✅	4s	1s	+300%
`poly_uniform`	✅	4s	5s	-20%
`poly_uniform_gamma1_4x`	✅	4s	4s	+0%
`polyeta_pack`	✅	4s	2s	+100%
`polyt0_pack`	✅	4s	2s	+100%
`polyt1_pack`	✅	4s	2s	+100%
`polyvecl_pointwise_acc_montgomery`	✅	4s	5s	-20%
`polyvecl_uniform_gamma1_serial`	✅	4s	2s	+100%
`polyvecl_unpack_z`	✅	4s	2s	+100%
`polyz_unpack`	✅	4s	4s	+0%
`rej_eta`	✅	4s	2s	+100%
`shake256_absorb`	✅	4s	4s	+0%
`sign_keypair`	✅	4s	3s	+33%
`sk_s1hat_get_poly`	✅	4s	3s	+33%
`sk_s2hat_get_poly`	✅	4s	2s	+100%
`sk_t0hat_get_poly`	✅	4s	4s	+0%
`unpack_pk_t1`	✅	4s	3s	+33%
`unpack_sk_s1hat`	✅	4s	5s	-20%
`yvec_init`	✅	4s	2s	+100%
`fqscale`	✅	3s	2s	+50%
`intt_native_x86_64`	✅	3s	6s	-50%
`keccak_f1600_x1_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	3s	+0%
`keccak_f1600_x4_native_avx2`	✅	3s	3s	+0%
`keccak_finalize`	✅	3s	3s	+0%
`keccakf1600_extract_bytes (big endian)`	✅	3s	4s	-25%
`keccakf1600_permute_native`	✅	3s	1s	+200%
`keccakf1600_xor_bytes (big endian)`	✅	3s	3s	+0%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes`	✅	3s	1s	+200%
`mld_ct_abs_i32`	✅	3s	3s	+0%
`mld_ct_cmask_neg_i32`	✅	3s	3s	+0%
`mld_ct_get_optblocker_u8`	✅	3s	2s	+50%
`mld_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mld_prepare_domain_separation_prefix`	✅	3s	3s	+0%
`mld_value_barrier_u32`	✅	3s	2s	+50%
`montgomery_reduce`	✅	3s	4s	-25%
`ntt_native_aarch64`	✅	3s	1s	+200%
`pack_sig_z`	✅	3s	3s	+0%
`pack_sk_s1`	✅	3s	2s	+50%
`poly_decompose_native_x86_64`	✅	3s	-	new
`poly_ntt_c`	✅	3s	3s	+0%
`poly_permute_bitrev_to_custom_optional_native`	✅	3s	5s	-40%
`poly_power2round`	✅	3s	5s	-40%
`poly_reduce`	✅	3s	5s	-40%
`poly_use_hint_native`	✅	3s	4s	-25%
`poly_use_hint_native_aarch64`	✅	3s	3s	+0%
`polyveck_unpack_eta`	✅	3s	3s	+0%
`polyvecl_chknorm`	✅	3s	3s	+0%
`polyvecl_pack_eta`	✅	3s	2s	+50%
`polyvecl_uniform_gamma1`	✅	3s	2s	+50%
`polyvecl_unpack_eta`	✅	3s	5s	-40%
`polyw1_pack`	✅	3s	3s	+0%
`polyw1_pack_88`	✅	3s	5s	-40%
`power2round`	✅	3s	3s	+0%
`reduce32`	✅	3s	3s	+0%
`rej_eta_c`	✅	3s	2s	+50%
`rej_uniform_eta_native_aarch64`	✅	3s	3s	+0%
`shake128_finalize`	✅	3s	2s	+50%
`shake128_init`	✅	3s	2s	+50%
`shake128_release`	✅	3s	3s	+0%
`shake256_finalize`	✅	3s	2s	+50%
`shake256_release`	✅	3s	2s	+50%
`shake256_squeeze`	✅	3s	5s	-40%
`shake256x4_absorb_once`	✅	3s	2s	+50%
`sig_unpack_hints`	✅	3s	4s	-25%
`sign_signature`	✅	3s	4s	-25%
`sign_signature_pre_hash_shake256`	✅	3s	4s	-25%
`sys_check_capability`	✅	3s	2s	+50%
`unpack_sk_s2hat`	✅	3s	3s	+0%
`unpack_sk_t0hat`	✅	3s	8s	-62%
`use_hint`	✅	3s	3s	+0%
`caddq`	✅	2s	3s	-33%
`intt_native_aarch64`	✅	2s	6s	-67%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	4s	-50%
`keccak_init`	✅	2s	2s	+0%
`keccakf1600_permute`	✅	2s	2s	+0%
`keccakf1600_xor_bytes`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes_native`	✅	2s	4s	-50%
`mld_ct_get_optblocker_i64`	✅	2s	1s	+100%
`mld_ct_get_optblocker_u32`	✅	2s	3s	-33%
`mld_ct_sel_int32`	✅	2s	3s	-33%
`mld_keccakf1600x4_xor_bytes_c`	✅	2s	3s	-33%
`mld_polymat_expand_entry`	✅	2s	3s	-33%
`mld_value_barrier_u8`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	4s	-50%
`pack_sig_h`	✅	2s	5s	-60%
`pointwise_native_aarch64`	✅	2s	3s	-33%
`poly_chknorm_native`	✅	2s	4s	-50%
`poly_chknorm_native_aarch64`	✅	2s	3s	-33%
`poly_chknorm_native_x86_64`	✅	2s	3s	-33%
`poly_decompose_32_native_aarch64`	✅	2s	4s	-50%
`poly_decompose_native`	✅	2s	3s	-33%
`poly_ntt_native`	✅	2s	3s	-33%
`poly_pointwise_montgomery_native`	✅	2s	4s	-50%
`poly_uniform_eta`	✅	2s	4s	-50%
`poly_uniform_gamma1`	✅	2s	2s	+0%
`poly_use_hint`	✅	2s	2s	+0%
`polyveck_reduce`	✅	2s	2s	+0%
`polyz_unpack_17_native_aarch64`	✅	2s	3s	-33%
`polyz_unpack_19_native_aarch64`	✅	2s	4s	-50%
`polyz_unpack_native`	✅	2s	2s	+0%
`shake128x4_absorb_once`	✅	2s	2s	+0%
`shake256x4_squeezeblocks`	✅	2s	2s	+0%
`sign_signature_pre_hash_internal`	✅	2s	4s	-50%
`sign_verify`	✅	2s	5s	-60%
`keccak_f1600_x4_native_aarch64_v84a`	✅	1s	2s	-50%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	3s	-67%
`keccak_squeeze`	✅	1s	3s	-67%
`keccakf1600x4_extract_bytes`	✅	1s	3s	-67%
`poly_permute_bitrev_to_custom_optional`	✅	1s	4s	-75%
`polyvec_matrix_pointwise_montgomery_row`	✅	1s	4s	-75%
`polyz_pack`	✅	1s	6s	-83%
`shake128_absorb`	✅	1s	1s	+0%
`shake128_squeeze`	✅	1s	3s	-67%
`shake128x4_squeezeblocks`	✅	1s	3s	-67%
`yvec_get_poly`	✅	1s	4s	-75%

oqs-bot · 2026-06-10T05:59:47Z

CBMC Results (ML-DSA-44)

Full Results (205 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1885s	1769s	+6.6%
`mld_invntt_layer`	✅	321s	281s	+14%
`rej_uniform_native`	✅	161s	144s	+12%
`polyvecl_pointwise_acc_montgomery_c`	✅	135s	119s	+13%
`poly_pointwise_montgomery_c`	✅	109s	92s	+18%
`mld_ct_memcmp`	✅	75s	63s	+19%
`mld_attempt_signature_generation`	✅	66s	63s	+5%
`mld_ntt_layer`	✅	46s	43s	+7%
`fqmul`	✅	44s	40s	+10%
`polyvec_matrix_expand`	✅	27s	25s	+8%
`sign_verify_internal`	✅	27s	26s	+4%
`mld_ntt_butterfly_block`	✅	24s	20s	+20%
`keccakf1600x4_permute_native`	✅	23s	22s	+5%
`rej_uniform`	✅	23s	23s	+0%
`poly_chknorm_c`	✅	21s	19s	+11%
`polyeta_unpack`	✅	18s	16s	+12%
`polyt0_unpack`	✅	18s	19s	-5%
`sign_signature_internal`	✅	18s	20s	-10%
`compute_pack_t0_t1`	✅	13s	14s	-7%
`mld_check_pct`	✅	13s	11s	+18%
`poly_uniform_eta_4x`	✅	13s	13s	+0%
`polyz_unpack_c`	✅	13s	13s	+0%
`rej_uniform_c`	✅	13s	14s	-7%
`poly_add`	✅	12s	10s	+20%
`polyveck_chknorm`	✅	12s	11s	+9%
`poly_use_hint_c`	✅	11s	6s	+83%
`poly_uniform_4x`	✅	10s	12s	-17%
`keccak_absorb_once_x4`	✅	9s	9s	+0%
`polyvec_matrix_pointwise_montgomery_yvec`	✅	9s	10s	-10%
`mld_compute_pack_z`	✅	8s	9s	-11%
`poly_invntt_tomont_c`	✅	8s	8s	+0%
`polyvec_matrix_expand_serial`	✅	8s	8s	+0%
`polyveck_decompose`	✅	8s	8s	+0%
`keccak_absorb`	✅	7s	7s	+0%
`poly_caddq_c`	✅	7s	7s	+0%
`poly_chknorm_native`	✅	7s	5s	+40%
`keccak_squeeze`	✅	6s	4s	+50%
`mld_keccakf1600_permute_c`	✅	6s	7s	-14%
`pointwise_acc_native_aarch64`	✅	6s	6s	+0%
`poly_decompose`	✅	6s	3s	+100%
`poly_uniform_gamma1_4x`	✅	6s	5s	+20%
`polyvecl_chknorm`	✅	6s	3s	+100%
`rej_eta_c`	✅	6s	3s	+100%
`sign`	✅	6s	6s	+0%
`sign_open`	✅	6s	4s	+50%
`sign_pk_from_sk`	✅	6s	7s	-14%
`sign_signature_pre_hash_shake256`	✅	6s	6s	+0%
`unpack_pk_t1`	✅	6s	3s	+100%
`mld_ct_cmask_nonzero_u8`	✅	5s	1s	+400%
`mld_sample_s1_s2`	✅	5s	4s	+25%
`pack_sk_s1`	✅	5s	2s	+150%
`pointwise_acc_native_x86_64`	✅	5s	5s	+0%
`poly_decompose_c`	✅	5s	5s	+0%
`poly_decompose_native`	✅	5s	5s	+0%
`poly_ntt`	✅	5s	4s	+25%
`poly_power2round`	✅	5s	6s	-17%
`poly_uniform`	✅	5s	3s	+67%
`poly_uniform_eta`	✅	5s	3s	+67%
`poly_use_hint`	✅	5s	2s	+150%
`poly_use_hint_native`	✅	5s	5s	+0%
`polyeta_pack`	✅	5s	4s	+25%
`polyvec_matrix_pointwise_montgomery_row`	✅	5s	3s	+67%
`polyveck_invntt_tomont`	✅	5s	4s	+25%
`polyz_pack`	✅	5s	5s	+0%
`rej_eta_native`	✅	5s	10s	-50%
`shake128_squeeze`	✅	5s	3s	+67%
`sign_signature_pre_hash_internal`	✅	5s	5s	+0%
`caddq`	✅	4s	3s	+33%
`decompose`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	4s	4s	+0%
`keccak_init`	✅	4s	2s	+100%
`keccakf1600_permute_native`	✅	4s	2s	+100%
`keccakf1600x4_extract_bytes_native`	✅	4s	6s	-33%
`mld_h`	✅	4s	4s	+0%
`ntt_native_aarch64`	✅	4s	3s	+33%
`pack_sig_c`	✅	4s	3s	+33%
`pack_sig_z`	✅	4s	4s	+0%
`poly_challenge`	✅	4s	5s	-20%
`poly_chknorm_native_aarch64`	✅	4s	3s	+33%
`poly_ntt_c`	✅	4s	4s	+0%
`poly_permute_bitrev_to_custom_optional`	✅	4s	3s	+33%
`poly_pointwise_montgomery_native`	✅	4s	4s	+0%
`poly_reduce`	✅	4s	3s	+33%
`poly_shiftl`	✅	4s	6s	-33%
`polyt0_pack`	✅	4s	4s	+0%
`polyveck_caddq`	✅	4s	3s	+33%
`polyveck_ntt`	✅	4s	3s	+33%
`polyvecl_ntt`	✅	4s	6s	-33%
`polyvecl_pointwise_acc_montgomery`	✅	4s	2s	+100%
`polyvecl_unpack_z`	✅	4s	4s	+0%
`power2round`	✅	4s	3s	+33%
`rej_eta`	✅	4s	5s	-20%
`rej_uniform_native_aarch64`	✅	4s	4s	+0%
`shake256`	✅	4s	2s	+100%
`shake256_init`	✅	4s	2s	+100%
`sign_keypair`	✅	4s	4s	+0%
`sign_keypair_internal`	✅	4s	5s	-20%
`sign_signature_extmu`	✅	4s	4s	+0%
`sign_verify_pre_hash_shake256`	✅	4s	5s	-20%
`sys_check_capability`	✅	4s	1s	+300%
`use_hint`	✅	4s	2s	+100%
`yvec_init`	✅	4s	2s	+100%
`intt_native_x86_64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	3s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	3s	3s	+0%
`keccak_squeezeblocks_x4`	✅	3s	4s	-25%
`keccakf1600_extract_bytes (big endian)`	✅	3s	2s	+50%
`keccakf1600_permute`	✅	3s	2s	+50%
`keccakf1600_xor_bytes`	✅	3s	3s	+0%
`keccakf1600_xor_bytes (big endian)`	✅	3s	3s	+0%
`keccakf1600x4_extract_bytes`	✅	3s	3s	+0%
`mld_ct_sel_int32`	✅	3s	2s	+50%
`mld_polymat_expand_entry`	✅	3s	3s	+0%
`mld_sample_s1_s2_serial`	✅	3s	2s	+50%
`mld_value_barrier_u8`	✅	3s	1s	+200%
`montgomery_reduce`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	4s	-25%
`pack_sig_h`	✅	3s	3s	+0%
`pack_sk_rho_key_tr_s2`	✅	3s	2s	+50%
`pointwise_native_aarch64`	✅	3s	3s	+0%
`pointwise_native_x86_64`	✅	3s	2s	+50%
`poly_caddq`	✅	3s	3s	+0%
`poly_caddq_native`	✅	3s	4s	-25%
`poly_caddq_native_x86_64`	✅	3s	2s	+50%
`poly_chknorm_native_x86_64`	✅	3s	4s	-25%
`poly_decompose_32_native_aarch64`	✅	3s	7s	-57%
`poly_decompose_88_native_aarch64`	✅	3s	2s	+50%
`poly_decompose_native_x86_64`	✅	3s	-	new
`poly_invntt_tomont`	✅	3s	3s	+0%
`poly_invntt_tomont_native`	✅	3s	4s	-25%
`poly_use_hint_native_aarch64`	✅	3s	1s	+200%
`polyveck_unpack_eta`	✅	3s	3s	+0%
`polyvecl_pointwise_acc_montgomery_native`	✅	3s	2s	+50%
`polyvecl_uniform_gamma1_serial`	✅	3s	5s	-40%
`polyvecl_unpack_eta`	✅	3s	3s	+0%
`polyw1_pack`	✅	3s	4s	-25%
`shake128_release`	✅	3s	2s	+50%
`shake128x4_absorb_once`	✅	3s	1s	+200%
`shake256_absorb`	✅	3s	2s	+50%
`shake256_release`	✅	3s	3s	+0%
`shake256_squeeze`	✅	3s	1s	+200%
`sign_signature`	✅	3s	4s	-25%
`sign_verify_extmu`	✅	3s	3s	+0%
`sk_s2hat_get_poly`	✅	3s	3s	+0%
`unpack_sk`	✅	3s	4s	-25%
`unpack_sk_s1hat`	✅	3s	2s	+50%
`fqscale`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	2s	+0%
`keccak_f1600_x4_native_avx2`	✅	2s	3s	-33%
`keccakf1600x4_permute`	✅	2s	2s	+0%
`keccakf1600x4_xor_bytes`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes_native`	✅	2s	4s	-50%
`make_hint`	✅	2s	3s	-33%
`mld_ct_get_optblocker_u32`	✅	2s	4s	-50%
`mld_keccakf1600_extract_bytes`	✅	2s	3s	-33%
`mld_keccakf1600x4_extract_bytes_c`	✅	2s	4s	-50%
`mld_keccakf1600x4_xor_bytes_c`	✅	2s	3s	-33%
`mld_prepare_domain_separation_prefix`	✅	2s	2s	+0%
`mld_value_barrier_i64`	✅	2s	1s	+100%
`mld_value_barrier_u32`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	4s	-50%
`poly_caddq_native_aarch64`	✅	2s	3s	-33%
`poly_chknorm`	✅	2s	4s	-50%
`poly_ntt_native`	✅	2s	3s	-33%
`poly_permute_bitrev_to_custom_optional_native`	✅	2s	3s	-33%
`poly_pointwise_montgomery`	✅	2s	2s	+0%
`poly_sub`	✅	2s	3s	-33%
`poly_uniform_gamma1`	✅	2s	3s	-33%
`polyt1_pack`	✅	2s	5s	-60%
`polyt1_unpack`	✅	2s	2s	+0%
`polyveck_pack_w1`	✅	2s	4s	-50%
`polyw1_pack_32`	✅	2s	2s	+0%
`polyw1_pack_88`	✅	2s	4s	-50%
`polyz_unpack`	✅	2s	4s	-50%
`polyz_unpack_17_native_aarch64`	✅	2s	3s	-33%
`polyz_unpack_19_native_aarch64`	✅	2s	4s	-50%
`polyz_unpack_native`	✅	2s	3s	-33%
`reduce32`	✅	2s	2s	+0%
`rej_uniform_eta_native_aarch64`	✅	2s	5s	-60%
`shake128_absorb`	✅	2s	3s	-33%
`shake128_finalize`	✅	2s	2s	+0%
`shake128_init`	✅	2s	1s	+100%
`shake256_finalize`	✅	2s	2s	+0%
`shake256x4_squeezeblocks`	✅	2s	3s	-33%
`sig_unpack_hints`	✅	2s	2s	+0%
`sign_verify`	✅	2s	5s	-60%
`sign_verify_pre_hash_internal`	✅	2s	3s	-33%
`sk_s1hat_get_poly`	✅	2s	2s	+0%
`sk_t0hat_get_poly`	✅	2s	2s	+0%
`unpack_sk_s2hat`	✅	2s	6s	-67%
`unpack_sk_t0hat`	✅	2s	5s	-60%
`yvec_get_poly`	✅	2s	2s	+0%
`keccak_finalize`	✅	1s	3s	-67%
`mld_ct_abs_i32`	✅	1s	1s	+0%
`mld_ct_cmask_neg_i32`	✅	1s	2s	-50%
`mld_ct_cmask_nonzero_u32`	✅	1s	3s	-67%
`mld_ct_get_optblocker_i64`	✅	1s	3s	-67%
`mld_ct_get_optblocker_u8`	✅	1s	3s	-67%
`polyveck_pack_eta`	✅	1s	3s	-67%
`polyveck_reduce`	✅	1s	4s	-75%
`polyvecl_pack_eta`	✅	1s	4s	-75%
`polyvecl_uniform_gamma1`	✅	1s	2s	-50%
`shake128x4_squeezeblocks`	✅	1s	3s	-67%
`shake256x4_absorb_once`	✅	1s	2s	-50%

github-actions

Mac Mini (M1, 2020) benchmarks (opt)

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`46539` cycles	`46537` cycles	`1.00`
`ML-DSA-44 sign`	`131082` cycles	`131061` cycles	`1.00`
`ML-DSA-44 verify`	`47348` cycles	`47345` cycles	`1.00`
`ML-DSA-65 keypair`	`81686` cycles	`81683` cycles	`1.00`
`ML-DSA-65 sign`	`215322` cycles	`215331` cycles	`1.00`
`ML-DSA-65 verify`	`79305` cycles	`79302` cycles	`1.00`
`ML-DSA-87 keypair`	`132401` cycles	`132400` cycles	`1.00`
`ML-DSA-87 sign`	`277532` cycles	`277357` cycles	`1.00`
`ML-DSA-87 verify`	`134051` cycles	`134055` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`112752` cycles	`112752` cycles	`1`
`ML-DSA-44 sign`	`400901` cycles	`400863` cycles	`1.00`
`ML-DSA-44 verify`	`119445` cycles	`119445` cycles	`1`
`ML-DSA-65 keypair`	`192978` cycles	`192931` cycles	`1.00`
`ML-DSA-65 sign`	`649977` cycles	`649957` cycles	`1.00`
`ML-DSA-65 verify`	`192863` cycles	`192871` cycles	`1.00`
`ML-DSA-87 keypair`	`318842` cycles	`318724` cycles	`1.00`
`ML-DSA-87 sign`	`828816` cycles	`828761` cycles	`1.00`
`ML-DSA-87 verify`	`326790` cycles	`326654` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`231326` cycles	`226291` cycles	`1.02`
`ML-DSA-44 sign`	`636953` cycles	`613701` cycles	`1.04`
`ML-DSA-44 verify`	`218748` cycles	`223287` cycles	`0.98`
`ML-DSA-65 keypair`	`401703` cycles	`401301` cycles	`1.00`
`ML-DSA-65 sign`	`1038618` cycles	`1019858` cycles	`1.02`
`ML-DSA-65 verify`	`382587` cycles	`377404` cycles	`1.01`
`ML-DSA-87 keypair`	`663545` cycles	`662182` cycles	`1.00`
`ML-DSA-87 sign`	`1385036` cycles	`1364271` cycles	`1.02`
`ML-DSA-87 verify`	`639267` cycles	`646030` cycles	`0.99`

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 (opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 sign`	`636953` cycles	`613701` cycles	`1.04`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`43326` cycles	`43362` cycles	`1.00`
`ML-DSA-44 sign`	`130586` cycles	`131357` cycles	`0.99`
`ML-DSA-44 verify`	`45079` cycles	`45329` cycles	`0.99`
`ML-DSA-65 keypair`	`75623` cycles	`75529` cycles	`1.00`
`ML-DSA-65 sign`	`214943` cycles	`215494` cycles	`1.00`
`ML-DSA-65 verify`	`74300` cycles	`74395` cycles	`1.00`
`ML-DSA-87 keypair`	`123196` cycles	`123052` cycles	`1.00`
`ML-DSA-87 sign`	`271568` cycles	`270650` cycles	`1.00`
`ML-DSA-87 verify`	`120595` cycles	`120614` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`91451` cycles	`91711` cycles	`1.00`
`ML-DSA-44 sign`	`351826` cycles	`353153` cycles	`1.00`
`ML-DSA-44 verify`	`99751` cycles	`100089` cycles	`1.00`
`ML-DSA-65 keypair`	`153919` cycles	`153963` cycles	`1.00`
`ML-DSA-65 sign`	`571974` cycles	`570637` cycles	`1.00`
`ML-DSA-65 verify`	`160011` cycles	`159765` cycles	`1.00`
`ML-DSA-87 keypair`	`255432` cycles	`256233` cycles	`1.00`
`ML-DSA-87 sign`	`726067` cycles	`727004` cycles	`1.00`
`ML-DSA-87 verify`	`264170` cycles	`264081` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`55425` cycles	`55232` cycles	`1.00`
`ML-DSA-44 sign`	`159735` cycles	`159604` cycles	`1.00`
`ML-DSA-44 verify`	`57634` cycles	`57852` cycles	`1.00`
`ML-DSA-65 keypair`	`96240` cycles	`95882` cycles	`1.00`
`ML-DSA-65 sign`	`263039` cycles	`264042` cycles	`1.00`
`ML-DSA-65 verify`	`96049` cycles	`96233` cycles	`1.00`
`ML-DSA-87 keypair`	`154779` cycles	`154587` cycles	`1.00`
`ML-DSA-87 sign`	`320859` cycles	`322317` cycles	`1.00`
`ML-DSA-87 verify`	`151689` cycles	`151310` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`133300` cycles	`133370` cycles	`1.00`
`ML-DSA-44 sign`	`519465` cycles	`519316` cycles	`1.00`
`ML-DSA-44 verify`	`146603` cycles	`146733` cycles	`1.00`
`ML-DSA-65 keypair`	`224063` cycles	`224213` cycles	`1.00`
`ML-DSA-65 sign`	`843482` cycles	`843252` cycles	`1.00`
`ML-DSA-65 verify`	`234302` cycles	`234223` cycles	`1.00`
`ML-DSA-87 keypair`	`367801` cycles	`367144` cycles	`1.00`
`ML-DSA-87 sign`	`1060988` cycles	`1060336` cycles	`1.00`
`ML-DSA-87 verify`	`381200` cycles	`380930` cycles	`1.00`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`112335` cycles	`112525` cycles	`1.00`
`ML-DSA-44 sign`	`354217` cycles	`354017` cycles	`1.00`
`ML-DSA-44 verify`	`117414` cycles	`117394` cycles	`1.00`
`ML-DSA-65 keypair`	`194466` cycles	`194697` cycles	`1.00`
`ML-DSA-65 sign`	`584370` cycles	`584501` cycles	`1.00`
`ML-DSA-65 verify`	`193464` cycles	`193283` cycles	`1.00`
`ML-DSA-87 keypair`	`320864` cycles	`320987` cycles	`1.00`
`ML-DSA-87 sign`	`747862` cycles	`746658` cycles	`1.00`
`ML-DSA-87 verify`	`318010` cycles	`318698` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`46713` cycles	`47001` cycles	`0.99`
`ML-DSA-44 sign`	`140686` cycles	`139110` cycles	`1.01`
`ML-DSA-44 verify`	`49619` cycles	`49262` cycles	`1.01`
`ML-DSA-65 keypair`	`82371` cycles	`82519` cycles	`1.00`
`ML-DSA-65 sign`	`228022` cycles	`227885` cycles	`1.00`
`ML-DSA-65 verify`	`82195` cycles	`82013` cycles	`1.00`
`ML-DSA-87 keypair`	`130193` cycles	`129228` cycles	`1.01`
`ML-DSA-87 sign`	`280026` cycles	`279733` cycles	`1.00`
`ML-DSA-87 verify`	`128511` cycles	`128347` cycles	`1.00`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`67357` cycles	`67336` cycles	`1.00`
`ML-DSA-44 sign`	`198377` cycles	`198343` cycles	`1.00`
`ML-DSA-44 verify`	`70274` cycles	`70240` cycles	`1.00`
`ML-DSA-65 keypair`	`119485` cycles	`119389` cycles	`1.00`
`ML-DSA-65 sign`	`326268` cycles	`325890` cycles	`1.00`
`ML-DSA-65 verify`	`117012` cycles	`116943` cycles	`1.00`
`ML-DSA-87 keypair`	`196441` cycles	`196722` cycles	`1.00`
`ML-DSA-87 sign`	`421298` cycles	`421918` cycles	`1.00`
`ML-DSA-87 verify`	`193282` cycles	`193428` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`299428` cycles	`301288` cycles	`0.99`
`ML-DSA-44 sign`	`1141001` cycles	`1145421` cycles	`1.00`
`ML-DSA-44 verify`	`330622` cycles	`332207` cycles	`1.00`
`ML-DSA-65 keypair`	`563729` cycles	`547720` cycles	`1.03`
`ML-DSA-65 sign`	`1920855` cycles	`1895789` cycles	`1.01`
`ML-DSA-65 verify`	`539918` cycles	`531514` cycles	`1.02`
`ML-DSA-87 keypair`	`861758` cycles	`855749` cycles	`1.01`
`ML-DSA-87 sign`	`2426173` cycles	`2378667` cycles	`1.02`
`ML-DSA-87 verify`	`906246` cycles	`883654` cycles	`1.03`

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

oqs-bot

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

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`118433` cycles	`118101` cycles	`1.00`
`ML-DSA-44 sign`	`458646` cycles	`458153` cycles	`1.00`
`ML-DSA-44 verify`	`130869` cycles	`130875` cycles	`1.00`
`ML-DSA-65 keypair`	`200886` cycles	`200971` cycles	`1.00`
`ML-DSA-65 sign`	`745083` cycles	`742473` cycles	`1.00`
`ML-DSA-65 verify`	`209164` cycles	`209101` cycles	`1.00`
`ML-DSA-87 keypair`	`331397` cycles	`332976` cycles	`1.00`
`ML-DSA-87 sign`	`938261` cycles	`938796` cycles	`1.00`
`ML-DSA-87 verify`	`343275` cycles	`342887` cycles	`1.00`

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

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`127600` cycles	`127655` cycles	`1.00`
`ML-DSA-44 sign`	`441155` cycles	`441153` cycles	`1.00`
`ML-DSA-44 verify`	`136410` cycles	`136366` cycles	`1.00`
`ML-DSA-65 keypair`	`220532` cycles	`220720` cycles	`1.00`
`ML-DSA-65 sign`	`714274` cycles	`713831` cycles	`1.00`
`ML-DSA-65 verify`	`221102` cycles	`220771` cycles	`1.00`
`ML-DSA-87 keypair`	`364562` cycles	`365122` cycles	`1.00`
`ML-DSA-87 sign`	`915619` cycles	`921347` cycles	`0.99`
`ML-DSA-87 verify`	`370883` cycles	`370803` cycles	`1.00`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`212223` cycles	`211830` cycles	`1.00`
`ML-DSA-44 sign`	`761057` cycles	`759846` cycles	`1.00`
`ML-DSA-44 verify`	`229916` cycles	`229343` cycles	`1.00`
`ML-DSA-65 keypair`	`378717` cycles	`377180` cycles	`1.00`
`ML-DSA-65 sign`	`1247998` cycles	`1247170` cycles	`1.00`
`ML-DSA-65 verify`	`373347` cycles	`371571` cycles	`1.00`
`ML-DSA-87 keypair`	`602516` cycles	`600508` cycles	`1.00`
`ML-DSA-87 sign`	`1584868` cycles	`1584292` cycles	`1.00`
`ML-DSA-87 verify`	`618340` cycles	`616048` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`61876` cycles	`61490` cycles	`1.01`
`ML-DSA-44 sign`	`189410` cycles	`189504` cycles	`1.00`
`ML-DSA-44 verify`	`66354` cycles	`66611` cycles	`1.00`
`ML-DSA-65 keypair`	`112014` cycles	`109562` cycles	`1.02`
`ML-DSA-65 sign`	`320464` cycles	`315340` cycles	`1.02`
`ML-DSA-65 verify`	`109875` cycles	`109818` cycles	`1.00`
`ML-DSA-87 keypair`	`171027` cycles	`171377` cycles	`1.00`
`ML-DSA-87 sign`	`379568` cycles	`379373` cycles	`1.00`
`ML-DSA-87 verify`	`170309` cycles	`170479` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`154733` cycles	`154452` cycles	`1.00`
`ML-DSA-44 sign`	`590183` cycles	`589881` cycles	`1.00`
`ML-DSA-44 verify`	`169693` cycles	`170191` cycles	`1.00`
`ML-DSA-65 keypair`	`262845` cycles	`263626` cycles	`1.00`
`ML-DSA-65 sign`	`965740` cycles	`966009` cycles	`1.00`
`ML-DSA-65 verify`	`272539` cycles	`273308` cycles	`1.00`
`ML-DSA-87 keypair`	`432704` cycles	`431959` cycles	`1.00`
`ML-DSA-87 sign`	`1211442` cycles	`1210781` cycles	`1.00`
`ML-DSA-87 verify`	`448136` cycles	`447098` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`112168` cycles	`112136` cycles	`1.00`
`ML-DSA-44 sign`	`353497` cycles	`353809` cycles	`1.00`
`ML-DSA-44 verify`	`117008` cycles	`117213` cycles	`1.00`
`ML-DSA-65 keypair`	`194787` cycles	`194348` cycles	`1.00`
`ML-DSA-65 sign`	`583930` cycles	`583675` cycles	`1.00`
`ML-DSA-65 verify`	`192719` cycles	`193087` cycles	`1.00`
`ML-DSA-87 keypair`	`320908` cycles	`320087` cycles	`1.00`
`ML-DSA-87 sign`	`747304` cycles	`747202` cycles	`1.00`
`ML-DSA-87 verify`	`318766` cycles	`317903` cycles	`1.00`

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

oqs-bot

Graviton3

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`71374` cycles	`71546` cycles	`1.00`
`ML-DSA-44 sign`	`208951` cycles	`208995` cycles	`1.00`
`ML-DSA-44 verify`	`74784` cycles	`74742` cycles	`1.00`
`ML-DSA-65 keypair`	`125930` cycles	`125949` cycles	`1.00`
`ML-DSA-65 sign`	`345622` cycles	`345451` cycles	`1.00`
`ML-DSA-65 verify`	`124102` cycles	`124199` cycles	`1.00`
`ML-DSA-87 keypair`	`207053` cycles	`206632` cycles	`1.00`
`ML-DSA-87 sign`	`443985` cycles	`439852` cycles	`1.01`
`ML-DSA-87 verify`	`204072` cycles	`204472` cycles	`1.00`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`137939` cycles	`138023` cycles	`1.00`
`ML-DSA-44 sign`	`486208` cycles	`486039` cycles	`1.00`
`ML-DSA-44 verify`	`149048` cycles	`149068` cycles	`1.00`
`ML-DSA-65 keypair`	`241520` cycles	`241810` cycles	`1.00`
`ML-DSA-65 sign`	`792069` cycles	`791663` cycles	`1.00`
`ML-DSA-65 verify`	`242200` cycles	`241314` cycles	`1.00`
`ML-DSA-87 keypair`	`395771` cycles	`396304` cycles	`1.00`
`ML-DSA-87 sign`	`1013680` cycles	`1019188` cycles	`0.99`
`ML-DSA-87 verify`	`403660` cycles	`403745` 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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`211517` cycles	`211659` cycles	`1.00`
`ML-DSA-44 sign`	`758603` cycles	`760082` cycles	`1.00`
`ML-DSA-44 verify`	`228966` cycles	`229485` cycles	`1.00`
`ML-DSA-65 keypair`	`377434` cycles	`377835` cycles	`1.00`
`ML-DSA-65 sign`	`1247925` cycles	`1246585` cycles	`1.00`
`ML-DSA-65 verify`	`371551` cycles	`371729` cycles	`1.00`
`ML-DSA-87 keypair`	`600603` cycles	`601814` cycles	`1.00`
`ML-DSA-87 sign`	`1582984` cycles	`1582429` cycles	`1.00`
`ML-DSA-87 verify`	`616265` cycles	`617716` cycles	`1.00`

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

github-actions

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

Details

Benchmark suite	Current: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`267515` cycles	`268565` cycles	`1.00`
`ML-DSA-44 sign`	`807339` cycles	`806886` cycles	`1.00`
`ML-DSA-44 verify`	`269280` cycles	`269851` cycles	`1.00`
`ML-DSA-65 keypair`	`461007` cycles	`459957` cycles	`1.00`
`ML-DSA-65 sign`	`1320496` cycles	`1317354` cycles	`1.00`
`ML-DSA-65 verify`	`447023` cycles	`445463` cycles	`1.00`
`ML-DSA-87 keypair`	`788862` cycles	`788523` cycles	`1.00`
`ML-DSA-87 sign`	`1808107` cycles	`1804026` cycles	`1.00`
`ML-DSA-87 verify`	`770607` cycles	`772022` cycles	`1.00`

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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`463001` cycles	`462832` cycles	`1.00`
`ML-DSA-44 sign`	`2131558` cycles	`2130663` cycles	`1.00`
`ML-DSA-44 verify`	`554513` cycles	`554950` cycles	`1.00`
`ML-DSA-65 keypair`	`780676` cycles	`781350` cycles	`1.00`
`ML-DSA-65 sign`	`3483972` cycles	`3478846` cycles	`1.00`
`ML-DSA-65 verify`	`863731` cycles	`864624` cycles	`1.00`
`ML-DSA-87 keypair`	`1265658` cycles	`1261131` cycles	`1.00`
`ML-DSA-87 sign`	`4297254` cycles	`4307837` cycles	`1.00`
`ML-DSA-87 verify`	`1390213` cycles	`1384611` cycles	`1.00`

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: `460553d`	Previous: `08d40f9`	Ratio
`ML-DSA-44 keypair`	`760293` cycles	`759832` cycles	`1.00`
`ML-DSA-44 sign`	`3140737` cycles	`3139623` cycles	`1.00`
`ML-DSA-44 verify`	`859554` cycles	`859077` cycles	`1.00`
`ML-DSA-65 keypair`	`1286158` cycles	`1285222` cycles	`1.00`
`ML-DSA-65 sign`	`5077195` cycles	`5072020` cycles	`1.00`
`ML-DSA-65 verify`	`1364303` cycles	`1363676` cycles	`1.00`
`ML-DSA-87 keypair`	`2112223` cycles	`2110495` cycles	`1.00`
`ML-DSA-87 sign`	`6355356` cycles	`6366388` cycles	`1.00`
`ML-DSA-87 verify`	`2228739` cycles	`2230493` cycles	`1.00`

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

mkannwischer · 2026-06-14T10:30:19Z

+      mldsa/poly_decompose_32_avx2_asm.o \
+      mldsa/poly_decompose_88_avx2_asm.o \


Please also add them to the README.

…mbly Mirror the AArch64 conversion in poly_decompose_{32,88}_aarch64_asm: replace the C intrinsics with fully-unrolled AVX2 routines, add HOL-Light correctness and memory-safety proofs, and CBMC contracts. Helper lemmas common to both variants are shared via the x86-only mldsa_utils.ml. - Resolves #420 - Resolves #914 Signed-off-by: Jake Massimo <jakemas@amazon.com>

jakemas · 2026-06-15T18:52:32Z

Added to readme

jakemas · 2026-06-15T19:05:47Z

This PR's branch was accidentally force-pushed while the PR was closed (I was trying to add the update my ammending the commit, but failed), which left it unreopenable on GitHub's side. Reopened the work as #1181 with identical content (rebased on latest main, 31 files). The README addition you asked for (@mkannwischer) is included there. Apologies for the churn.

jakemas force-pushed the jakemas/poly-decompose-asm branch 4 times, most recently from 753befe to 460553d Compare June 10, 2026 05:28

jakemas added the benchmark label Jun 10, 2026