|
Boost : |
Subject: Re: [boost] What's wrong with this benchmark
From: Michael Marcin (mike.marcin_at_[hidden])
Date: 2016-02-16 23:20:57
On 2/16/2016 11:56 AM, Benedek Thaler wrote:
> Dear Boost,
>
> In a previous mail, I introduced a container library, `double_ended`.
> This library has benchmarks [0]. However, the benchmarks show
> unexpected results.
>
> This is what I do [1]:
> - compile program using G++ 5.3, -O2
> - remove CPU0 from kernel scheduler pool (isolcpu=0 boot flag)
> - disable CPU scaling
> - pin benchmark thread to CPU0
> - in the test, create a vector
> - reserve memory for 1M unsigned
> - write a byte to each page reserved, to warm up the page table
> - get TSC value (machine has constant_tsc flag)
> - insert 1M unsigned value
> - get TSC value again, subtract previous value
> - repeat test 25 times, average results
>
> I do this with 3 containers: std::vector, boost::container::vector and
> double_ended::devector. While the latter two produces almost
> equivalent results, std::vector seriously falls behind.
>
> - No reallocation happens between the TSC sampling
> - Test order doesn't matter: every order produces the same results
> - Container size or sample count doesn't matter, same results.
> - The deviations of individual measures of each container is small,
> the performance is consistent.
> - I compared the produced ASM of std::vector and
> double_ended::devector, they are very similar, almost the same
> instructions in different order [2].
>
> What do I miss here?
>
> Thanks,
> Benedek
>
>
> [0]: http://erenon.hu/double_ended/double_ended/benchmarks.html
> [1]: https://github.com/erenon/double_ended/blob/master/benchmark/push_back.cpp
> [2]: http://pastebin.com/tUL54i9f
>
FWIW I can't reproduce your results on Visual Studio 2015 Update 1. Here
are my results if you're interested.
X std::vector
0 0
100 592.380000
200 1178.640000
400 2310.270000
800 4517.580000
1600 8893.160000
3200 17600.610000
6400 35895.900000
12800 71004.060000
25600 142081.980000
51200 283691.720000
102400 567921.980000
204800 1138470.790000
409600 2274810.920000
819200 4556220.740000
X devector
0 0
100 711.840000
200 1458.540000
400 2924.550000
800 5862.690000
1600 11653.440000
3200 23825.170000
6400 46841.700000
12800 93916.710000
25600 187148.900000
51200 372430.440000
102400 745044.570000
204800 1488865.490000
409600 2971389.220000
819200 5942622.180000
X boost::container::vector
0 0
100 837.930000
200 1666.050000
400 3259.280000
800 6423.750000
1600 12706.180000
3200 25197.520000
6400 50089.290000
12800 100486.150000
25600 200127.660000
51200 398981.780000
102400 800517.160000
204800 1610967.220000
409600 3250085.410000
819200 6454347.270000
X std::deque
0 0
100 18529.880000
200 25942.200000
400 38937.040000
800 59459.400000
1600 100410.640000
3200 175987.360000
6400 321362.520000
12800 569617.680000
25600 1019693.840000
51200 1930645.720000
102400 3806069.880000
204800 7704495.360000
409600 15629599.200000
819200 31983278.800000
X batch_deque
0 0
100 991.760000
200 1735.560000
400 3127.280000
800 5864.880000
1600 11357.320000
3200 22291.800000
6400 44334.520000
12800 88523.040000
25600 179476.320000
51200 358522.960000
102400 716579.960000
204800 1433212.760000
409600 2868735.640000
819200 5741968.240000
X boost::container::deque
0 0
100 1187.320000
200 4253.640000
400 6874.320000
800 11220.320000
1600 20418.400000
3200 38970.160000
6400 74156.560000
12800 142497.720000
25600 281098.400000
51200 537988.920000
102400 1084152.000000
204800 2232533.560000
409600 4669299.360000
819200 9431670.920000
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk