Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- RK3188 with 32bit memory bus
- tinymembench v0.3.9 (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 : 469.1 MB/s
- C copy : 883.3 MB/s (1.7%)
- C copy prefetched (32 bytes step) : 683.3 MB/s
- C copy prefetched (64 bytes step) : 684.8 MB/s (1.2%)
- C 2-pass copy : 796.2 MB/s
- C 2-pass copy prefetched (32 bytes step) : 737.6 MB/s (1.5%)
- C 2-pass copy prefetched (64 bytes step) : 735.2 MB/s (1.2%)
- C fill : 3069.5 MB/s
- ---
- standard memcpy : 1013.6 MB/s (2.2%)
- standard memset : 3071.3 MB/s
- ---
- NEON read : 1371.4 MB/s (1.0%)
- NEON read prefetched (32 bytes step) : 1854.8 MB/s
- NEON read prefetched (64 bytes step) : 1903.1 MB/s (1.7%)
- NEON read 2 data streams : 1571.6 MB/s
- NEON read 2 data streams prefetched (32 bytes step) : 2161.9 MB/s (1.9%)
- NEON read 2 data streams prefetched (64 bytes step) : 2353.7 MB/s
- NEON copy : 609.4 MB/s (4.3%)
- NEON copy prefetched (32 bytes step) : 1102.6 MB/s (1.4%)
- NEON copy prefetched (64 bytes step) : 950.9 MB/s
- NEON unrolled copy : 912.5 MB/s (1.4%)
- NEON unrolled copy prefetched (32 bytes step) : 1021.9 MB/s
- NEON unrolled copy prefetched (64 bytes step) : 998.1 MB/s
- NEON copy backwards : 339.4 MB/s (0.2%)
- NEON copy backwards prefetched (32 bytes step) : 746.0 MB/s
- NEON copy backwards prefetched (64 bytes step) : 811.9 MB/s (1.2%)
- NEON 2-pass copy : 769.4 MB/s
- NEON 2-pass copy prefetched (32 bytes step) : 1039.7 MB/s
- NEON 2-pass copy prefetched (64 bytes step) : 1022.0 MB/s (1.3%)
- NEON unrolled 2-pass copy : 775.2 MB/s (1.0%)
- NEON unrolled 2-pass copy prefetched (32 bytes step) : 1029.0 MB/s
- NEON unrolled 2-pass copy prefetched (64 bytes step) : 1067.7 MB/s (1.5%)
- NEON fill : 3070.4 MB/s (1.0%)
- NEON fill backwards : 3071.9 MB/s (1.1%)
- VFP copy : 907.3 MB/s (1.0%)
- VFP 2-pass copy : 845.9 MB/s
- ARM fill (STRD) : 3071.0 MB/s
- ARM fill (STM with 8 registers) : 3070.8 MB/s (1.0%)
- ARM fill (STM with 4 registers) : 3071.6 MB/s (1.8%)
- ARM copy prefetched (incr pld) : 1050.2 MB/s
- ARM copy prefetched (wrap pld) : 1052.0 MB/s
- ARM 2-pass copy prefetched (incr pld) : 1044.6 MB/s (1.6%)
- ARM 2-pass copy prefetched (wrap pld) : 1051.4 MB/s (1.0%)
- ==========================================================================
- == 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
- 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 : 8.2 ns / 12.9 ns
- 131072 : 12.2 ns / 16.6 ns
- 262144 : 16.8 ns / 20.8 ns
- 524288 : 22.5 ns / 27.5 ns
- 1048576 : 86.3 ns / 131.3 ns
- 2097152 : 116.7 ns / 167.6 ns
- 4194304 : 133.2 ns / 184.6 ns
- 8388608 : 144.1 ns / 199.9 ns
- 16777216 : 155.0 ns / 212.3 ns
- 33554432 : 165.1 ns / 226.1 ns
- 67108864 : 178.9 ns / 251.2 ns
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement