Created
February 25, 2020 10:48
-
-
Save imbushuo/20d45676801d88a9c5f81ee9ad5c73e7 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| tinymembench v0.4.9 (simple benchmark for memory throughput and latency) | |
| ========================================================================== | |
| == Memory bandwidth tests == | |
| == == | |
| == Note 1: 1MB = 1000000 bytes == | |
| == Note 2: Results for 'copy' tests show how many bytes can be == | |
| == copied per second (adding together read and writen == | |
| == bytes would have provided twice higher numbers) == | |
| == Note 3: 2-pass copy means that we are using a small temporary buffer == | |
| == to first fetch data into it, and only then write it to the == | |
| == destination (source -> L1 cache, L1 cache -> destination) == | |
| == Note 4: If sample standard deviation exceeds 0.1%, it is shown in == | |
| == brackets == | |
| ========================================================================== | |
| C copy backwards : 12442.3 MB/s (5.8%) | |
| C copy backwards (32 byte blocks) : 12567.4 MB/s (2.9%) | |
| C copy backwards (64 byte blocks) : 12717.2 MB/s (1.8%) | |
| C copy : 13679.0 MB/s (1.4%) | |
| C copy prefetched (32 bytes step) : 14011.8 MB/s (1.2%) | |
| C copy prefetched (64 bytes step) : 13911.5 MB/s (0.5%) | |
| C 2-pass copy : 5345.9 MB/s (10.9%) | |
| C 2-pass copy prefetched (32 bytes step) : 8757.7 MB/s (4.0%) | |
| C 2-pass copy prefetched (64 bytes step) : 8025.0 MB/s (1.7%) | |
| C fill : 32692.9 MB/s (11.3%) | |
| C fill (shuffle within 16 byte blocks) : 32590.8 MB/s (0.4%) | |
| C fill (shuffle within 32 byte blocks) : 32490.2 MB/s (0.3%) | |
| C fill (shuffle within 64 byte blocks) : 32638.4 MB/s (0.6%) | |
| --- | |
| standard memcpy : 13689.7 MB/s (0.6%) | |
| standard memset : 32481.3 MB/s (6.6%) | |
| --- | |
| NEON LDP/STP copy : 13698.7 MB/s (8.7%) | |
| NEON LDP/STP copy pldl2strm (32 bytes step) : 13278.6 MB/s (2.3%) | |
| NEON LDP/STP copy pldl2strm (64 bytes step) : 13252.9 MB/s (0.4%) | |
| NEON LDP/STP copy pldl1keep (32 bytes step) : 13101.1 MB/s (1.9%) | |
| NEON LDP/STP copy pldl1keep (64 bytes step) : 13140.6 MB/s (1.1%) | |
| NEON LD1/ST1 copy : 13496.3 MB/s (0.4%) | |
| NEON STP fill : 32511.8 MB/s (5.9%) | |
| NEON STNP fill : 32770.8 MB/s (8.4%) | |
| ARM LDP/STP copy : 13682.9 MB/s (0.8%) | |
| ARM STP fill : 32467.4 MB/s (2.0%) | |
| ARM STNP fill : 32205.5 MB/s | |
| ========================================================================== | |
| == Memory latency test == | |
| == == | |
| == Average time is measured for random memory accesses in the buffers == | |
| == of different sizes. The larger is the buffer, the more significant == | |
| == are relative contributions of TLB, L1/L2 cache misses and SDRAM == | |
| == accesses. For extremely large buffer sizes we are expecting to see == | |
| == page table walk with several requests to SDRAM for almost every == | |
| == memory access (though 64MiB is not nearly large enough to experience == | |
| == this effect to its fullest). == | |
| == == | |
| == Note 1: All the numbers are representing extra time, which needs to == | |
| == be added to L1 cache latency. The cycle timings for L1 cache == | |
| == latency can be usually found in the processor documentation. == | |
| == Note 2: Dual random read means that we are simultaneously performing == | |
| == two independent memory accesses at a time. In the case if == | |
| == the memory subsystem can't handle multiple outstanding == | |
| == requests, dual random read has the same timings as two == | |
| == single reads performed one after another. == | |
| ========================================================================== | |
| block size : single random read / dual random read, [MADV_NOHUGEPAGE] | |
| 1024 : 0.0 ns / 0.0 ns | |
| 2048 : 0.0 ns / 0.0 ns | |
| 4096 : 0.0 ns / 0.0 ns | |
| 8192 : 0.0 ns / 0.0 ns | |
| 16384 : 0.0 ns / 0.0 ns | |
| 32768 : 0.0 ns / 0.0 ns | |
| 65536 : 0.0 ns / 0.0 ns | |
| 131072 : 0.9 ns / 1.2 ns | |
| 262144 : 1.7 ns / 2.2 ns | |
| 524288 : 3.4 ns / 4.4 ns | |
| 1048576 : 9.1 ns / 12.3 ns | |
| 2097152 : 12.2 ns / 15.4 ns | |
| 4194304 : 16.8 ns / 20.1 ns | |
| 8388608 : 64.3 ns / 86.5 ns | |
| 16777216 : 100.1 ns / 129.8 ns | |
| 33554432 : 127.4 ns / 153.3 ns | |
| 67108864 : 146.2 ns / 163.5 ns | |
| block size : single random read / dual random read, [MADV_HUGEPAGE] | |
| 1024 : 0.0 ns / 0.0 ns | |
| 2048 : 0.0 ns / 0.0 ns | |
| 4096 : 0.0 ns / 0.0 ns | |
| 8192 : 0.0 ns / 0.0 ns | |
| 16384 : 0.0 ns / 0.0 ns | |
| 32768 : 0.0 ns / 0.0 ns | |
| 65536 : 0.0 ns / 0.0 ns | |
| 131072 : 0.8 ns / 1.2 ns | |
| 262144 : 1.6 ns / 2.0 ns | |
| 524288 : 3.2 ns / 4.1 ns | |
| 1048576 : 8.9 ns / 12.3 ns | |
| 2097152 : 12.2 ns / 15.3 ns | |
| 4194304 : 16.7 ns / 20.4 ns | |
| 8388608 : 60.4 ns / 84.4 ns | |
| 16777216 : 101.2 ns / 130.0 ns | |
| 33554432 : 125.6 ns / 148.5 ns | |
| 67108864 : 140.5 ns / 156.2 ns |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment