evmone
evmone copied to clipboard
ethereum
cmake: Enable -mtune=haswell
The -mtune=generic is not default although some sources claim it is. Besides the generic target is too low in current compilers. Use haswell for instruction tuning which should match x86-64-v3 architecture (unfortunately, the -mtune=x86-64-v3 is not supported).
baseline/execute/main/blake2b_huff/empty_mean -0.0203 -0.0203 12 12 12 12
baseline/execute/main/blake2b_huff/2805nulls_mean -0.0192 -0.0192 259 254 259 254
baseline/execute/main/blake2b_huff/5610nulls_mean -0.0195 -0.0195 506 496 506 496
baseline/execute/main/blake2b_huff/8415nulls_mean -0.0219 -0.0219 741 725 741 725
baseline/execute/main/blake2b_huff/65536nulls_mean -0.0182 -0.0182 5751 5647 5751 5647
baseline/execute/main/blake2b_shifts/2805nulls_mean +0.0030 +0.0030 2597 2605 2597 2605
baseline/execute/main/blake2b_shifts/5610nulls_mean +0.0023 +0.0023 5181 5193 5181 5193
baseline/execute/main/blake2b_shifts/8415nulls_mean +0.0024 +0.0024 7763 7781 7763 7781
baseline/execute/main/blake2b_shifts/65536nulls_mean +0.0009 +0.0009 60113 60168 60113 60168
baseline/execute/main/sha1_divs/empty_mean +0.0042 +0.0042 49 49 49 49
baseline/execute/main/sha1_divs/1351_mean +0.0068 +0.0068 1021 1028 1021 1028
baseline/execute/main/sha1_divs/2737_mean +0.0079 +0.0079 1991 2007 1991 2007
baseline/execute/main/sha1_divs/5311_mean +0.0073 +0.0073 3887 3916 3887 3916
baseline/execute/main/sha1_divs/65536_mean +0.0067 +0.0067 47397 47714 47397 47715
baseline/execute/main/sha1_shifts/empty_mean -0.0012 -0.0012 28 28 28 28
baseline/execute/main/sha1_shifts/1351_mean -0.0003 -0.0003 591 591 591 591
baseline/execute/main/sha1_shifts/2737_mean -0.0016 -0.0016 1155 1153 1155 1153
baseline/execute/main/sha1_shifts/5311_mean +0.0008 +0.0008 2250 2251 2250 2251
baseline/execute/main/sha1_shifts/65536_mean +0.0003 +0.0003 27445 27453 27445 27453
baseline/execute/main/weierstrudel/0_mean +0.0185 +0.0185 151 154 151 154
baseline/execute/main/weierstrudel/1_mean +0.0311 +0.0311 331 341 331 341
baseline/execute/main/weierstrudel/3_mean +0.0287 +0.0287 518 533 518 533
baseline/execute/main/weierstrudel/9_mean +0.0279 +0.0279 1072 1102 1072 1102
baseline/execute/main/weierstrudel/14_mean +0.0288 +0.0288 1535 1579 1535 1579
baseline/execute/micro/jump_around/empty_mean -0.0170 -0.0170 24 23 24 23
baseline/execute/micro/jumpdests_0xffff/empty_mean +0.0213 +0.0213 106 108 106 108
baseline/execute/micro/loop_with_many_jumpdests/empty_mean +0.0366 +0.0366 12891 13363 12891 13363
baseline/execute/micro/memory_grow_mload/nogrow_mean -0.0004 -0.0004 46 46 46 46
baseline/execute/micro/memory_grow_mload/by1_mean -0.0023 -0.0023 49 49 49 49
baseline/execute/micro/memory_grow_mload/by16_mean -0.0034 -0.0034 53 53 53 53
baseline/execute/micro/memory_grow_mload/by32_mean +0.0014 +0.0014 61 61 61 61
baseline/execute/micro/memory_grow_mstore/nogrow_mean -0.0007 -0.0007 49 49 49 49
baseline/execute/micro/memory_grow_mstore/by1_mean -0.0007 -0.0007 51 51 51 51
baseline/execute/micro/memory_grow_mstore/by16_mean -0.0009 -0.0009 58 58 58 58
baseline/execute/micro/memory_grow_mstore/by32_mean -0.0035 -0.0035 68 67 68 67
baseline/execute/micro/signextend/zero_mean -0.0259 -0.0259 57 55 57 55
baseline/execute/micro/signextend/one_mean -0.0233 -0.0233 57 55 57 55
OVERALL_GEOMEAN +0.0014 +0.0014 0 0 0 0
Codecov Report
Merging #434 (f853c61) into master (31fd534) will not change coverage. The diff coverage is
n/a.
Additional details and impacted files
@@ Coverage Diff @@
## master #434 +/- ##
=======================================
Coverage 97.29% 97.29%
=======================================
Files 78 78
Lines 7648 7648
=======================================
Hits 7441 7441
Misses 207 207
| Flag | Coverage Δ | |
|---|---|---|
| blockchaintests | 65.16% <ø> (ø) |
|
| statetests | 63.18% <ø> (ø) |
|
| unittests | 94.84% <ø> (ø) |
Flags with carried forward coverage won't be shown. Click here to find out more.
![codecov[bot] avatar](https://avatars.githubusercontent.com/in/254?v=4 "Published by a pub.dev verified publisher"){kind=small}
De numbers are pretty similar to #435, perhaps should rebase this and re-bench?
De numbers are pretty similar to #435, perhaps should rebase this and re-bench?
The difference is close to 0.