Polygon zkEVM Goldilocks域各项运算性能

本文链接：https://blog.csdn.net/mutourend/article/details/135055435

1. 引言

开源代码见：

https://github.com/0xPolygonHermez/goldilocks（C++）

goldilocks# ./bench
2023-12-18T02:06:43+00:00
Running ./bench
Run on (128 X 2650 MHz CPU s)
CPU Caches:
  L1 Data 32 KiB (x64)
  L1 Instruction 32 KiB (x64)
  L2 Unified 512 KiB (x64)
  L3 Unified 32768 KiB (x8)
Load Average: 0.08, 1.92, 4.67
---------------------------------------------------------------------------------------------------
Benchmark                                         Time             CPU   Iterations UserCounters...
---------------------------------------------------------------------------------------------------
POSEIDON_BENCH_FULL/64/real_time             215672 us       215665 us            3 BytesProcessed=890.241M/s Rate=6.58178us
POSEIDON_BENCH_FULL/128/real_time            117095 us       117093 us            6 BytesProcessed=1.60126G/s Rate=3.57346us
POSEIDON_BENCH_FULL_AVX/64/real_time          77219 us        77216 us            9 BytesProcessed=2.42815G/s Rate=2.35655us
POSEIDON_BENCH_FULL_AVX/128/real_time         42010 us        42007 us           17 BytesProcessed=4.46323G/s Rate=1.28204us
POSEIDON_BENCH/64/real_time                  212826 us       212819 us            3 BytesProcessed=902.145M/s Rate=6.49494us
POSEIDON_BENCH/128/real_time                 109300 us       109296 us            6 BytesProcessed=1.71546G/s Rate=3.33557us
POSEIDON_BENCH_AVX/64/real_time               77543 us        77540 us            9 BytesProcessed=2.418G/s Rate=2.36644us
POSEIDON_BENCH_AVX/128/real_time              69186 us        69132 us           11 BytesProcessed=2.71009G/s Rate=2.11139us
LINEAR_HASH_BENCH/64/real_time             10262542 us     10261880 us            1 BytesProcessed=798.243M/s Rate=4.89356us
LINEAR_HASH_BENCH/128/real_time             7548851 us      7235086 us            1 BytesProcessed=1085.2M/s Rate=3.59957us
LINEAR_HASH_BENCH_AVX/64/real_time          4279616 us      3876050 us            1 BytesProcessed=1.86933G/s Rate=2.04068us
LINEAR_HASH_BENCH_AVX/128/real_time         3023446 us      2963734 us            1 BytesProcessed=2.64599G/s Rate=1.44169us
MERKLETREE_BENCH/64/real_time               9595622 us      7825490 us            1 BytesProcessed=853.723M/s Rate=4.57555us
MERKLETREE_BENCH/128/real_time              8261920 us      7953147 us            1 BytesProcessed=991.537M/s Rate=3.93959us
MERKLETREE_BENCH_AVX/64/real_time           4609280 us      3884992 us            1 BytesProcessed=1.73563G/s Rate=2.19788us
MERKLETREE_BENCH_AVX/128/real_time          3614813 us      3299295 us            1 BytesProcessed=2.21312G/s Rate=1.72368us
MERKLETREE_BATCH_BENCH/64/real_time         9701176 us      9111133 us            1 BytesProcessed=844.434M/s Rate=4.62588us
MERKLETREE_BATCH_BENCH/128/real_time        8798086 us      8709507 us            1 BytesProcessed=931.112M/s Rate=4.19525us
MERKLETREE_BATCH_BENCH_AVX/64/real_time     4650252 us      4424796 us            1 BytesProcessed=1.72034G/s Rate=2.21741us
MERKLETREE_BATCH_BENCH_AVX/128/real_time    3623811 us      3496373 us            1 BytesProcessed=2.20762G/s Rate=1.72797us
NTT_BENCH/64/real_time                         3.82 s          3.27 s             1
NTT_BENCH/128/real_time                        3.54 s          3.09 s             1
NTT_BLOCK_BENCH/64/real_time                   2.18 s          2.18 s             1
NTT_BLOCK_BENCH/128/real_time                  2.87 s          2.85 s             1
LDE_BENCH/64/real_time                         11.5 s          10.5 s             1
LDE_BENCH/128/real_time                        11.6 s          10.5 s             1
LDE_BLOCK_BENCH/64/real_time                   17.6 s          13.7 s             1
LDE_BLOCK_BENCH/128/real_time                  14.2 s          13.5 s             1
EXTENDEDPOL_BENCH/64/real_time                 5.21 s          5.21 s             1
EXTENDEDPOL_BENCH/128/real_time                5.40 s          5.07 s             1