Skip to content

FIPS202/x86_64: Don't use native x4 backend on x86_64 when not needed#1032

Open
flynd wants to merge 1 commit intomainfrom
no-unneeded-x64-x4-backend
Open

FIPS202/x86_64: Don't use native x4 backend on x86_64 when not needed#1032
flynd wants to merge 1 commit intomainfrom
no-unneeded-x64-x4-backend

Conversation

@flynd
Copy link
Copy Markdown
Contributor

@flynd flynd commented Apr 8, 2026

Match the AArch64 behavior and skip the native Keccak-f1600x4 backend when MLD_CONFIG_SERIAL_FIPS202_ONLY or MLD_CONFIG_REDUCE_RAM is set.

This change was part of PR #1000 but doesn't need to be part of that stack.

@flynd flynd requested a review from a team as a code owner April 8, 2026 10:12
@oqs-bot
Copy link
Copy Markdown
Contributor

oqs-bot commented Apr 8, 2026 •
edited
Loading

CBMC Results (ML-DSA-87)

⚠️ Attention Required

Proof Status Current Previous Change
mld_attempt_signature_generation ⚠️ 202s 96s +110%
Full Results (184 proofs)
Proof Status Current Previous Change
**TOTAL** ✅ 2442s 2411s +1.3%
poly_pointwise_montgomery_c ✅ 252s 186s +35%
polyvec_matrix_expand ✅ 247s 255s -3%
polyvecl_pointwise_acc_montgomery_c ✅ 206s 218s -6%
mld_attempt_signature_generation ⚠️ 202s 96s +110%
rej_uniform_native ✅ 154s 155s -1%
sign_verify_internal ✅ 108s 147s -27%
mld_invntt_layer ✅ 104s 104s +0%
polyvec_matrix_expand_serial ✅ 86s 88s -2%
mld_ct_memcmp ✅ 81s 89s -9%
mld_ntt_layer ✅ 54s 59s -8%
sign_signature_internal ✅ 40s 41s -2%
polymat_permute_bitrev_to_custom ✅ 26s 27s -4%
mld_compute_t0_t1_tr_from_sk_components ✅ 24s 25s -4%
rej_uniform ✅ 21s 24s -12%
fqmul ✅ 20s 23s -13%
poly_chknorm_c ✅ 19s 23s -17%
poly_uniform_eta_4x ✅ 19s 16s +19%
polyeta_unpack ✅ 19s 19s +0%
polyveck_power2round ✅ 19s 20s -5%
rej_uniform_c ✅ 17s 19s -11%
poly_uniform_4x ✅ 15s 17s -12%
mld_polyvecl_permute_bitrev_to_custom_native ✅ 14s 14s +0%
polyt0_unpack ✅ 14s 20s -30%
keccak_absorb_once_x4 ✅ 13s 12s +8%
keccakf1600x4_permute_native ✅ 12s 14s -14%
mld_check_pct ✅ 12s 17s -29%
mld_ntt_butterfly_block ✅ 12s 16s -25%
polyveck_decompose ✅ 12s 11s +9%
poly_invntt_tomont_c ✅ 11s 7s +57%
polyveck_ntt ✅ 11s 8s +38%
polyveck_use_hint ✅ 11s 12s -8%
poly_add ✅ 10s 11s -9%
polyveck_add ✅ 10s 11s -9%
unpack_sk ✅ 10s 8s +25%
mld_compute_pack_z ✅ 9s 10s -10%
keccak_squeezeblocks_x4 ✅ 8s 9s -11%
keccakf1600_permute_native ✅ 8s 9s -11%
mld_sample_s1_s2 ✅ 8s 5s +60%
poly_caddq_c ✅ 8s 7s +14%
polyvec_matrix_pointwise_montgomery ✅ 8s 10s -20%
polyveck_invntt_tomont ✅ 8s 10s -20%
polyveck_pointwise_poly_montgomery ✅ 8s 6s +33%
polyvecl_ntt ✅ 8s 9s -11%
sign ✅ 8s 7s +14%
sign_pk_from_sk ✅ 8s 7s +14%
keccak_absorb ✅ 7s 7s +0%
keccakf1600_permute ✅ 7s 6s +17%
mld_sample_s1_s2_serial ✅ 7s 7s +0%
poly_uniform_eta ✅ 7s 7s +0%
polyveck_chknorm ✅ 7s 7s +0%
polyveck_reduce ✅ 7s 8s -12%
polyveck_sub ✅ 7s 7s +0%
polyz_unpack_c ✅ 7s 12s -42%
poly_challenge ✅ 6s 5s +20%
poly_decompose_c ✅ 6s 7s -14%
polyveck_caddq ✅ 6s 8s -25%
polyveck_shiftl ✅ 6s 8s -25%
polyveck_unpack_t0 ✅ 6s 5s +20%
polyvecl_chknorm ✅ 6s 8s -25%
rej_eta_native ✅ 6s 4s +50%
sign_keypair ✅ 6s 3s +100%
sign_signature ✅ 6s 5s +20%
intt_native_x86_64 ✅ 5s 1s +400%
mld_value_barrier_u32 ✅ 5s 1s +400%
pack_sk ✅ 5s 6s -17%
pointwise_native_x86_64 ✅ 5s 2s +150%
poly_invntt_tomont_native ✅ 5s 3s +67%
poly_uniform_gamma1_4x ✅ 5s 2s +150%
polyvecl_pack_eta ✅ 5s 3s +67%
shake256_release ✅ 5s 3s +67%
sign_open ✅ 5s 5s +0%
sign_signature_pre_hash_shake256 ✅ 5s 4s +25%
unpack_hints ✅ 5s 6s -17%
keccak_squeeze ✅ 4s 3s +33%
keccakf1600x4_extract_bytes ✅ 4s 2s +100%
mld_prepare_domain_separation_prefix ✅ 4s 4s +0%
pack_pk ✅ 4s 4s +0%
poly_caddq ✅ 4s 4s +0%
poly_caddq_native ✅ 4s 2s +100%
poly_chknorm ✅ 4s 1s +300%
poly_decompose ✅ 4s 4s +0%
poly_ntt ✅ 4s 4s +0%
poly_ntt_c ✅ 4s 6s -33%
poly_ntt_native ✅ 4s 3s +33%
poly_pointwise_montgomery_native ✅ 4s 4s +0%
poly_power2round ✅ 4s 6s -33%
poly_uniform ✅ 4s 5s -20%
poly_use_hint_c ✅ 4s 2s +100%
polyveck_unpack_eta ✅ 4s 6s -33%
polyvecl_permute_bitrev_to_custom ✅ 4s 4s +0%
polyvecl_uniform_gamma1 ✅ 4s 2s +100%
polyvecl_uniform_gamma1_serial ✅ 4s 2s +100%
polyw1_pack ✅ 4s 3s +33%
polyz_unpack_native ✅ 4s 2s +100%
shake128_finalize ✅ 4s 2s +100%
sign_keypair_internal ✅ 4s 6s -33%
sign_verify_extmu ✅ 4s 4s +0%
sign_verify_pre_hash_internal ✅ 4s 2s +100%
decompose ✅ 3s 2s +50%
keccak_f1600_x1_native_aarch64 ✅ 3s 1s +200%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid ✅ 3s 1s +200%
keccakf1600_xor_bytes (big endian) ✅ 3s 3s +0%
keccakf1600x4_permute ✅ 3s 2s +50%
keccakf1600x4_xor_bytes ✅ 3s 3s +0%
mld_ct_cmask_nonzero_u8 ✅ 3s 2s +50%
mld_ct_get_optblocker_i64 ✅ 3s 4s -25%
mld_ct_sel_int32 ✅ 3s 2s +50%
mld_h ✅ 3s 3s +0%
mld_value_barrier_i64 ✅ 3s 2s +50%
mld_value_barrier_u8 ✅ 3s 3s +0%
montgomery_reduce ✅ 3s 3s +0%
ntt_native_aarch64 ✅ 3s 4s -25%
ntt_native_x86_64 ✅ 3s 3s +0%
pack_sig_c ✅ 3s 2s +50%
pack_sig_h_poly ✅ 3s 4s -25%
pack_sig_z ✅ 3s 4s -25%
poly_caddq_native_aarch64 ✅ 3s 5s -40%
poly_chknorm_native ✅ 3s 4s -25%
poly_chknorm_native_aarch64 ✅ 3s 2s +50%
poly_decompose_native ✅ 3s 7s -57%
poly_invntt_tomont ✅ 3s 4s -25%
poly_shiftl ✅ 3s 5s -40%
poly_sub ✅ 3s 3s +0%
poly_uniform_gamma1 ✅ 3s 2s +50%
polyt0_pack ✅ 3s 5s -40%
polyt1_pack ✅ 3s 3s +0%
polyveck_pack_eta ✅ 3s 4s -25%
polyvecl_pointwise_acc_montgomery_native ✅ 3s 3s +0%
polyz_unpack ✅ 3s 4s -25%
power2round ✅ 3s 2s +50%
reduce32 ✅ 3s 2s +50%
rej_eta ✅ 3s 3s +0%
rej_eta_c ✅ 3s 3s +0%
shake256 ✅ 3s 3s +0%
shake256_absorb ✅ 3s 4s -25%
shake256_squeeze ✅ 3s 4s -25%
shake256x4_squeezeblocks ✅ 3s 2s +50%
sign_signature_extmu ✅ 3s 6s -50%
sign_verify_pre_hash_shake256 ✅ 3s 7s -57%
sys_check_capability ✅ 3s 3s +0%
unpack_sig ✅ 3s 3s +0%
fqscale ✅ 2s 3s -33%
keccak_finalize ✅ 2s 2s +0%
keccak_init ✅ 2s 3s -33%
keccakf1600_extract_bytes (big endian) ✅ 2s 2s +0%
make_hint ✅ 2s 4s -50%
mld_ct_abs_i32 ✅ 2s 3s -33%
mld_ct_cmask_neg_i32 ✅ 2s 2s +0%
mld_ct_cmask_nonzero_u32 ✅ 2s 3s -33%
mld_ct_get_optblocker_u32 ✅ 2s 3s -33%
mld_ct_get_optblocker_u8 ✅ 2s 2s +0%
pointwise_native_aarch64 ✅ 2s 4s -50%
poly_make_hint ✅ 2s 6s -67%
poly_reduce ✅ 2s 4s -50%
poly_use_hint ✅ 2s 3s -33%
poly_use_hint_native ✅ 2s 3s -33%
polyt1_unpack ✅ 2s 3s -33%
polyveck_pack_t0 ✅ 2s 4s -50%
polyveck_pack_w1 ✅ 2s 4s -50%
polyvecl_unpack_z ✅ 2s 5s -60%
polyz_pack ✅ 2s 2s +0%
shake128_init ✅ 2s 4s -50%
shake128_release ✅ 2s 3s -33%
shake128_squeeze ✅ 2s 2s +0%
shake128x4_absorb_once ✅ 2s 2s +0%
shake128x4_squeezeblocks ✅ 2s 2s +0%
shake256_finalize ✅ 2s 6s -67%
shake256_init ✅ 2s 1s +100%
shake256x4_absorb_once ✅ 2s 2s +0%
sign_signature_pre_hash_internal ✅ 2s 4s -50%
sign_verify ✅ 2s 6s -67%
use_hint ✅ 2s 2s +0%
caddq ✅ 1s 2s -50%
keccak_f1600_x1_native_aarch64_v84a ✅ 1s 2s -50%
keccak_f1600_x4_native_aarch64_v84a ✅ 1s 1s +0%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid ✅ 1s 1s +0%
keccakf1600_xor_bytes ✅ 1s 2s -50%
mld_keccakf1600_extract_bytes ✅ 1s 2s -50%
poly_pointwise_montgomery ✅ 1s 1s +0%
polyeta_pack ✅ 1s 5s -80%
polyvecl_pointwise_acc_montgomery ✅ 1s 4s -75%
polyvecl_unpack_eta ✅ 1s 5s -80%
shake128_absorb ✅ 1s 3s -67%
unpack_pk ✅ 1s 5s -80%

@oqs-bot
Copy link
Copy Markdown
Contributor

oqs-bot commented Apr 8, 2026 •
edited
Loading

CBMC Results (ML-DSA-65)

⚠️ Attention Required

Proof Status Current Previous Change
poly_pointwise_montgomery_c ⚠️ 248s 156s +59%
Full Results (184 proofs)
Proof Status Current Previous Change
**TOTAL** ✅ 2502s 2329s +7.4%
sign_verify_internal ✅ 308s 286s +8%
polyvecl_pointwise_acc_montgomery_c ✅ 295s 280s +5%
poly_pointwise_montgomery_c ⚠️ 248s 156s +59%
rej_uniform_native ✅ 148s 148s +0%
mld_attempt_signature_generation ✅ 116s 107s +8%
polyvec_matrix_expand ✅ 110s 107s +3%
mld_invntt_layer ✅ 100s 94s +6%
polyvec_matrix_expand_serial ✅ 83s 81s +2%
mld_ct_memcmp ✅ 76s 79s -4%
mld_ntt_layer ✅ 54s 55s -2%
polymat_permute_bitrev_to_custom ✅ 33s 37s -11%
mld_compute_t0_t1_tr_from_sk_components ✅ 28s 28s +0%
polyveck_use_hint ✅ 26s 22s +18%
sign_signature_internal ✅ 22s 25s -12%
fqmul ✅ 21s 18s +17%
rej_uniform ✅ 21s 22s -5%
poly_chknorm_c ✅ 20s 19s +5%
polyveck_decompose ✅ 19s 16s +19%
poly_uniform_eta_4x ✅ 18s 16s +12%
poly_uniform_4x ✅ 16s 15s +7%
polyt0_unpack ✅ 15s 15s +0%
mld_ntt_butterfly_block ✅ 14s 13s +8%
rej_uniform_c ✅ 14s 15s -7%
keccak_absorb_once_x4 ✅ 13s 10s +30%
polyveck_power2round ✅ 13s 12s +8%
keccakf1600x4_permute_native ✅ 12s 13s -8%
poly_add ✅ 12s 10s +20%
polyvec_matrix_pointwise_montgomery ✅ 12s 15s -20%
mld_check_pct ✅ 11s 13s -15%
mld_polyvecl_permute_bitrev_to_custom_native ✅ 10s 8s +25%
poly_invntt_tomont_c ✅ 10s 7s +43%
poly_challenge ✅ 9s 6s +50%
polyveck_chknorm ✅ 9s 8s +12%
polyveck_pointwise_poly_montgomery ✅ 9s 6s +50%
polyvecl_ntt ✅ 9s 5s +80%
keccak_squeezeblocks_x4 ✅ 8s 6s +33%
keccakf1600_permute ✅ 8s 10s -20%
poly_decompose_c ✅ 8s 7s +14%
polyveck_add ✅ 8s 10s -20%
polyveck_caddq ✅ 8s 11s -27%
sign_pk_from_sk ✅ 8s 9s -11%
unpack_sk ✅ 8s 9s -11%
intt_native_x86_64 ✅ 7s 2s +250%
keccakf1600_permute_native ✅ 7s 8s -12%
mld_h ✅ 7s 5s +40%
polyveck_invntt_tomont ✅ 7s 9s -22%
polyveck_ntt ✅ 7s 7s +0%
polyveck_sub ✅ 7s 8s -12%
sign ✅ 7s 7s +0%
sign_keypair_internal ✅ 7s 6s +17%
keccak_absorb ✅ 6s 5s +20%
pointwise_native_x86_64 ✅ 6s 1s +500%
poly_ntt ✅ 6s 4s +50%
polyz_pack ✅ 6s 2s +200%
shake256x4_squeezeblocks ✅ 6s 3s +100%
sign_signature_extmu ✅ 6s 4s +50%
sign_signature_pre_hash_shake256 ✅ 6s 5s +20%
mld_prepare_domain_separation_prefix ✅ 5s 6s -17%
mld_sample_s1_s2 ✅ 5s 7s -29%
montgomery_reduce ✅ 5s 5s +0%
ntt_native_x86_64 ✅ 5s 5s +0%
pack_sig_z ✅ 5s 2s +150%
pack_sk ✅ 5s 4s +25%
pointwise_native_aarch64 ✅ 5s 3s +67%
poly_pointwise_montgomery ✅ 5s 6s -17%
poly_power2round ✅ 5s 7s -29%
poly_reduce ✅ 5s 2s +150%
poly_use_hint_c ✅ 5s 4s +25%
polyveck_reduce ✅ 5s 5s +0%
polyveck_shiftl ✅ 5s 7s -29%
polyvecl_uniform_gamma1_serial ✅ 5s 2s +150%
polyvecl_unpack_eta ✅ 5s 3s +67%
rej_eta ✅ 5s 4s +25%
sign_open ✅ 5s 4s +25%
sign_verify_pre_hash_internal ✅ 5s 6s -17%
caddq ✅ 4s 2s +100%
fqscale ✅ 4s 4s +0%
keccak_f1600_x1_native_aarch64_v84a ✅ 4s 3s +33%
mld_compute_pack_z ✅ 4s 6s -33%
mld_ct_cmask_neg_i32 ✅ 4s 4s +0%
mld_ct_cmask_nonzero_u8 ✅ 4s 2s +100%
mld_ct_get_optblocker_u8 ✅ 4s 2s +100%
mld_sample_s1_s2_serial ✅ 4s 3s +33%
pack_pk ✅ 4s 3s +33%
pack_sig_c ✅ 4s 2s +100%
poly_caddq_c ✅ 4s 5s -20%
poly_caddq_native ✅ 4s 2s +100%
poly_chknorm_native ✅ 4s 3s +33%
poly_chknorm_native_aarch64 ✅ 4s 4s +0%
poly_shiftl ✅ 4s 2s +100%
polyeta_unpack ✅ 4s 6s -33%
polyt0_pack ✅ 4s 3s +33%
polyveck_unpack_eta ✅ 4s 2s +100%
polyvecl_permute_bitrev_to_custom ✅ 4s 4s +0%
polyvecl_pointwise_acc_montgomery ✅ 4s 4s +0%
polyvecl_pointwise_acc_montgomery_native ✅ 4s 3s +33%
polyvecl_unpack_z ✅ 4s 2s +100%
shake128_init ✅ 4s 2s +100%
shake128x4_absorb_once ✅ 4s 3s +33%
shake128x4_squeezeblocks ✅ 4s 3s +33%
shake256_squeeze ✅ 4s 5s -20%
sign_signature ✅ 4s 4s +0%
sign_signature_pre_hash_internal ✅ 4s 7s -43%
unpack_hints ✅ 4s 5s -20%
unpack_pk ✅ 4s 4s +0%
unpack_sig ✅ 4s 2s +100%
decompose ✅ 3s 4s -25%
keccak_f1600_x1_native_aarch64 ✅ 3s 2s +50%
keccak_f1600_x4_native_aarch64_v84a ✅ 3s 3s +0%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid ✅ 3s 4s -25%
keccak_init ✅ 3s 2s +50%
keccakf1600_xor_bytes ✅ 3s 2s +50%
keccakf1600x4_extract_bytes ✅ 3s 1s +200%
mld_ct_abs_i32 ✅ 3s 2s +50%
mld_ct_sel_int32 ✅ 3s 4s -25%
mld_value_barrier_i64 ✅ 3s 4s -25%
ntt_native_aarch64 ✅ 3s 3s +0%
pack_sig_h_poly ✅ 3s 4s -25%
poly_caddq ✅ 3s 2s +50%
poly_caddq_native_aarch64 ✅ 3s 5s -40%
poly_chknorm ✅ 3s 2s +50%
poly_decompose_native ✅ 3s 4s -25%
poly_invntt_tomont ✅ 3s 6s -50%
poly_invntt_tomont_native ✅ 3s 3s +0%
poly_ntt_native ✅ 3s 3s +0%
poly_sub ✅ 3s 3s +0%
poly_uniform ✅ 3s 5s -40%
poly_uniform_eta ✅ 3s 3s +0%
poly_use_hint ✅ 3s 1s +200%
polyt1_pack ✅ 3s 3s +0%
polyt1_unpack ✅ 3s 5s -40%
polyveck_pack_eta ✅ 3s 3s +0%
polyveck_pack_t0 ✅ 3s 4s -25%
polyveck_pack_w1 ✅ 3s 3s +0%
polyveck_unpack_t0 ✅ 3s 2s +50%
polyvecl_chknorm ✅ 3s 5s -40%
polyvecl_pack_eta ✅ 3s 3s +0%
polyvecl_uniform_gamma1 ✅ 3s 3s +0%
polyw1_pack ✅ 3s 3s +0%
polyz_unpack ✅ 3s 4s -25%
polyz_unpack_c ✅ 3s 2s +50%
polyz_unpack_native ✅ 3s 5s -40%
reduce32 ✅ 3s 3s +0%
rej_eta_c ✅ 3s 3s +0%
rej_eta_native ✅ 3s 5s -40%
shake128_absorb ✅ 3s 4s -25%
shake128_release ✅ 3s 3s +0%
shake128_squeeze ✅ 3s 2s +50%
shake256 ✅ 3s 2s +50%
shake256_finalize ✅ 3s 2s +50%
sign_keypair ✅ 3s 2s +50%
sign_verify ✅ 3s 2s +50%
sign_verify_extmu ✅ 3s 2s +50%
sign_verify_pre_hash_shake256 ✅ 3s 5s -40%
sys_check_capability ✅ 3s 3s +0%
use_hint ✅ 3s 3s +0%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid ✅ 2s 1s +100%
keccak_finalize ✅ 2s 1s +100%
keccakf1600_extract_bytes (big endian) ✅ 2s 4s -50%
keccakf1600x4_permute ✅ 2s 5s -60%
keccakf1600x4_xor_bytes ✅ 2s 3s -33%
make_hint ✅ 2s 2s +0%
mld_ct_cmask_nonzero_u32 ✅ 2s 3s -33%
mld_ct_get_optblocker_i64 ✅ 2s 2s +0%
mld_ct_get_optblocker_u32 ✅ 2s 2s +0%
mld_keccakf1600_extract_bytes ✅ 2s 1s +100%
mld_value_barrier_u8 ✅ 2s 4s -50%
poly_decompose ✅ 2s 3s -33%
poly_make_hint ✅ 2s 3s -33%
poly_pointwise_montgomery_native ✅ 2s 2s +0%
poly_uniform_gamma1 ✅ 2s 5s -60%
poly_uniform_gamma1_4x ✅ 2s 5s -60%
poly_use_hint_native ✅ 2s 4s -50%
polyeta_pack ✅ 2s 4s -50%
power2round ✅ 2s 2s +0%
shake128_finalize ✅ 2s 1s +100%
shake256_init ✅ 2s 2s +0%
shake256_release ✅ 2s 3s -33%
keccak_squeeze ✅ 1s 6s -83%
keccakf1600_xor_bytes (big endian) ✅ 1s 3s -67%
mld_value_barrier_u32 ✅ 1s 1s +0%
poly_ntt_c ✅ 1s 2s -50%
shake256_absorb ✅ 1s 3s -67%
shake256x4_absorb_once ✅ 1s 1s +0%

@oqs-bot
Copy link
Copy Markdown
Contributor

oqs-bot commented Apr 8, 2026 •
edited
Loading

CBMC Results (ML-DSA-44)

⚠️ Attention Required

Proof Status Current Previous Change
mld_attempt_signature_generation ⚠️ 295s 103s +186%
poly_invntt_tomont_c ⚠️ 26s 6s +333%
Full Results (184 proofs)
Proof Status Current Previous Change
**TOTAL** ✅ 2344s 2039s +15.0%
polyvecl_pointwise_acc_montgomery_c ✅ 377s 336s +12%
mld_attempt_signature_generation ⚠️ 295s 103s +186%
sign_verify_internal ✅ 168s 213s -21%
poly_pointwise_montgomery_c ✅ 159s 148s +7%
rej_uniform_native ✅ 151s 139s +9%
mld_invntt_layer ✅ 112s 85s +32%
mld_ct_memcmp ✅ 74s 72s +3%
mld_ntt_layer ✅ 57s 52s +10%
polymat_permute_bitrev_to_custom ✅ 30s 26s +15%
poly_invntt_tomont_c ⚠️ 26s 6s +333%
polyvec_matrix_expand_serial ✅ 24s 21s +14%
rej_uniform ✅ 24s 20s +20%
poly_chknorm_c ✅ 22s 20s +10%
polyvec_matrix_expand ✅ 21s 23s -9%
fqmul ✅ 19s 20s -5%
polyt0_unpack ✅ 19s 14s +36%
poly_uniform_eta_4x ✅ 18s 17s +6%
sign_signature_internal ✅ 18s 19s -5%
polyeta_unpack ✅ 17s 16s +6%
rej_uniform_c ✅ 15s 15s +0%
keccakf1600x4_permute_native ✅ 14s 13s +8%
poly_uniform_4x ✅ 14s 16s -12%
polyz_unpack_c ✅ 14s 12s +17%
mld_compute_t0_t1_tr_from_sk_components ✅ 13s 11s +18%
keccak_absorb_once_x4 ✅ 12s 9s +33%
mld_ntt_butterfly_block ✅ 12s 11s +9%
keccakf1600_permute_native ✅ 11s 9s +22%
mld_check_pct ✅ 11s 12s -8%
poly_add ✅ 10s 10s +0%
keccak_absorb ✅ 9s 5s +80%
mld_polyvecl_permute_bitrev_to_custom_native ✅ 9s 8s +12%
polyvec_matrix_pointwise_montgomery ✅ 8s 6s +33%
polyveck_use_hint ✅ 8s 7s +14%
polyvecl_ntt ✅ 8s 6s +33%
keccak_squeezeblocks_x4 ✅ 7s 6s +17%
keccakf1600_permute ✅ 7s 8s -12%
poly_power2round ✅ 7s 2s +250%
poly_uniform ✅ 7s 4s +75%
polyveck_add ✅ 7s 8s -12%
polyveck_decompose ✅ 7s 6s +17%
sign ✅ 7s 8s -12%
sign_pk_from_sk ✅ 7s 8s -12%
unpack_sk ✅ 7s 7s +0%
mld_prepare_domain_separation_prefix ✅ 6s 6s +0%
mld_sample_s1_s2_serial ✅ 6s 4s +50%
poly_caddq_c ✅ 6s 5s +20%
poly_chknorm_native_aarch64 ✅ 6s 4s +50%
poly_decompose_c ✅ 6s 3s +100%
poly_use_hint_c ✅ 6s 4s +50%
polyveck_chknorm ✅ 6s 3s +100%
polyveck_pointwise_poly_montgomery ✅ 6s 3s +100%
polyveck_sub ✅ 6s 4s +50%
rej_eta_native ✅ 6s 6s +0%
sign_verify_pre_hash_internal ✅ 6s 6s +0%
fqscale ✅ 5s 3s +67%
keccakf1600x4_extract_bytes ✅ 5s 3s +67%
mld_compute_pack_z ✅ 5s 7s -29%
mld_h ✅ 5s 4s +25%
mld_sample_s1_s2 ✅ 5s 6s -17%
pointwise_native_aarch64 ✅ 5s 2s +150%
poly_challenge ✅ 5s 7s -29%
poly_sub ✅ 5s 3s +67%
poly_uniform_eta ✅ 5s 5s +0%
poly_uniform_gamma1_4x ✅ 5s 4s +25%
polyt1_pack ✅ 5s 4s +25%
polyveck_caddq ✅ 5s 3s +67%
polyveck_power2round ✅ 5s 6s -17%
polyveck_reduce ✅ 5s 4s +25%
polyveck_shiftl ✅ 5s 4s +25%
polyvecl_chknorm ✅ 5s 5s +0%
shake128_init ✅ 5s 5s +0%
sign_keypair ✅ 5s 6s -17%
sign_keypair_internal ✅ 5s 4s +25%
sign_verify ✅ 5s 6s -17%
sign_verify_pre_hash_shake256 ✅ 5s 3s +67%
unpack_pk ✅ 5s 2s +150%
decompose ✅ 4s 2s +100%
keccak_f1600_x1_native_aarch64_v84a ✅ 4s 3s +33%
keccakf1600x4_xor_bytes ✅ 4s 3s +33%
make_hint ✅ 4s 1s +300%
mld_ct_cmask_neg_i32 ✅ 4s 1s +300%
mld_ct_cmask_nonzero_u8 ✅ 4s 3s +33%
pack_sk ✅ 4s 3s +33%
poly_ntt_native ✅ 4s 1s +300%
poly_pointwise_montgomery_native ✅ 4s 3s +33%
poly_reduce ✅ 4s 3s +33%
poly_shiftl ✅ 4s 3s +33%
polyt0_pack ✅ 4s 7s -43%
polyveck_invntt_tomont ✅ 4s 3s +33%
polyveck_ntt ✅ 4s 6s -33%
polyveck_unpack_eta ✅ 4s 4s +0%
polyvecl_pointwise_acc_montgomery_native ✅ 4s 4s +0%
polyvecl_uniform_gamma1 ✅ 4s 3s +33%
polyvecl_unpack_eta ✅ 4s 3s +33%
polyz_unpack ✅ 4s 2s +100%
polyz_unpack_native ✅ 4s 3s +33%
rej_eta ✅ 4s 4s +0%
rej_eta_c ✅ 4s 5s -20%
shake128x4_squeezeblocks ✅ 4s 2s +100%
shake256_release ✅ 4s 3s +33%
sign_signature ✅ 4s 5s -20%
sign_signature_extmu ✅ 4s 6s -33%
sign_signature_pre_hash_shake256 ✅ 4s 4s +0%
sign_verify_extmu ✅ 4s 4s +0%
sys_check_capability ✅ 4s 4s +0%
unpack_hints ✅ 4s 6s -33%
caddq ✅ 3s 2s +50%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid ✅ 3s 1s +200%
keccak_init ✅ 3s 2s +50%
keccak_squeeze ✅ 3s 6s -50%
keccakf1600_xor_bytes (big endian) ✅ 3s 2s +50%
mld_ct_abs_i32 ✅ 3s 3s +0%
mld_ct_cmask_nonzero_u32 ✅ 3s 2s +50%
mld_value_barrier_u32 ✅ 3s 1s +200%
mld_value_barrier_u8 ✅ 3s 2s +50%
ntt_native_x86_64 ✅ 3s 4s -25%
pack_sig_c ✅ 3s 3s +0%
pack_sig_h_poly ✅ 3s 5s -40%
pointwise_native_x86_64 ✅ 3s 2s +50%
poly_caddq ✅ 3s 3s +0%
poly_caddq_native_aarch64 ✅ 3s 4s -25%
poly_chknorm ✅ 3s 2s +50%
poly_chknorm_native ✅ 3s 2s +50%
poly_decompose_native ✅ 3s 3s +0%
poly_ntt ✅ 3s 2s +50%
poly_pointwise_montgomery ✅ 3s 3s +0%
poly_use_hint ✅ 3s 3s +0%
poly_use_hint_native ✅ 3s 6s -50%
polyeta_pack ✅ 3s 2s +50%
polyveck_pack_eta ✅ 3s 4s -25%
polyveck_pack_t0 ✅ 3s 3s +0%
polyveck_pack_w1 ✅ 3s 3s +0%
polyveck_unpack_t0 ✅ 3s 3s +0%
polyvecl_permute_bitrev_to_custom ✅ 3s 5s -40%
polyvecl_unpack_z ✅ 3s 3s +0%
shake128_release ✅ 3s 5s -40%
shake128_squeeze ✅ 3s 3s +0%
shake128x4_absorb_once ✅ 3s 4s -25%
shake256_absorb ✅ 3s 2s +50%
shake256_init ✅ 3s 2s +50%
shake256x4_squeezeblocks ✅ 3s 1s +200%
sign_open ✅ 3s 3s +0%
sign_signature_pre_hash_internal ✅ 3s 2s +50%
unpack_sig ✅ 3s 3s +0%
use_hint ✅ 3s 2s +50%
intt_native_x86_64 ✅ 2s 3s -33%
keccak_f1600_x1_native_aarch64 ✅ 2s 2s +0%
keccak_f1600_x4_native_aarch64_v84a ✅ 2s 2s +0%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid ✅ 2s 4s -50%
keccak_finalize ✅ 2s 4s -50%
keccakf1600_extract_bytes (big endian) ✅ 2s 5s -60%
keccakf1600_xor_bytes ✅ 2s 2s +0%
keccakf1600x4_permute ✅ 2s 2s +0%
mld_ct_get_optblocker_i64 ✅ 2s 2s +0%
mld_ct_get_optblocker_u32 ✅ 2s 2s +0%
mld_ct_get_optblocker_u8 ✅ 2s 1s +100%
mld_ct_sel_int32 ✅ 2s 2s +0%
mld_keccakf1600_extract_bytes ✅ 2s 3s -33%
ntt_native_aarch64 ✅ 2s 5s -60%
pack_sig_z ✅ 2s 5s -60%
poly_caddq_native ✅ 2s 3s -33%
poly_invntt_tomont ✅ 2s 4s -50%
poly_invntt_tomont_native ✅ 2s 2s +0%
poly_make_hint ✅ 2s 5s -60%
poly_ntt_c ✅ 2s 2s +0%
polyt1_unpack ✅ 2s 5s -60%
polyvecl_pointwise_acc_montgomery ✅ 2s 3s -33%
polyvecl_uniform_gamma1_serial ✅ 2s 3s -33%
polyw1_pack ✅ 2s 2s +0%
polyz_pack ✅ 2s 2s +0%
power2round ✅ 2s 5s -60%
shake128_absorb ✅ 2s 3s -33%
shake256 ✅ 2s 2s +0%
shake256_finalize ✅ 2s 3s -33%
shake256_squeeze ✅ 2s 3s -33%
mld_value_barrier_i64 ✅ 1s 2s -50%
montgomery_reduce ✅ 1s 2s -50%
pack_pk ✅ 1s 4s -75%
poly_decompose ✅ 1s 6s -83%
poly_uniform_gamma1 ✅ 1s 3s -67%
polyvecl_pack_eta ✅ 1s 4s -75%
reduce32 ✅ 1s 2s -50%
shake128_finalize ✅ 1s 3s -67%
shake256x4_absorb_once ✅ 1s 4s -75%

@flynd flynd mentioned this pull request Apr 8, 2026
Open
@mkannwischer
Copy link
Copy Markdown
Contributor

mkannwischer commented Apr 8, 2026

Thanks @flynd. Can you elaborate how this matches the AArch64 behavior? I don't see a matching guard in the AArch64 backend.

@flynd
Copy link
Copy Markdown
Contributor Author

flynd commented Apr 8, 2026

Thanks @flynd. Can you elaborate how this matches the AArch64 behavior? I don't see a matching guard in the AArch64 backend.

This was a commit you made (added to my stack in PR #1000) that I noticed didn't have any dependencies to that stack. I see now that you're referring to one of my commits so the commit message is now referring to something that doesn't exist on master yet.

I will grab the relevant parts of my commit and add to this PR to make it consistent.

@flynd
fips202: Don't build Keccak-f1600x2/x4 unless used
c29f36d 
When building with MLD_CONFIG_SERIAL_FIPS202_ONLY, Keccak-f1600x2/x4
is not used and can be skipped.

Signed-off-by: Anders Sonmark <Anders.Sonmark@axis.com>
@flynd flynd force-pushed the no-unneeded-x64-x4-backend branch from 0b8f3f2 to c29f36d Compare April 8, 2026 13:03
@mkannwischer mkannwischer self-assigned this Apr 9, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

@mkannwischer mkannwischer

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@flynd @oqs-bot @mkannwischer