Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- RK3188 1080p 32bpp
- 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 : 263.2 MB/s (1.5%)
- C copy : 293.7 MB/s (1.9%)
- C copy prefetched (32 bytes step) : 219.9 MB/s (1.4%)
- C copy prefetched (64 bytes step) : 219.7 MB/s (1.5%)
- C 2-pass copy : 240.2 MB/s (1.4%)
- C 2-pass copy prefetched (32 bytes step) : 226.9 MB/s (1.5%)
- C 2-pass copy prefetched (64 bytes step) : 227.0 MB/s (1.4%)
- C fill : 619.9 MB/s (2.2%)
- ---
- standard memcpy : 302.0 MB/s (1.8%)
- standard memset : 619.4 MB/s (2.4%)
- ---
- NEON read : 765.4 MB/s
- NEON read prefetched (32 bytes step) : 805.8 MB/s (3.7%)
- NEON read prefetched (64 bytes step) : 807.8 MB/s
- NEON read 2 data streams : 784.6 MB/s (3.0%)
- NEON read 2 data streams prefetched (32 bytes step) : 789.6 MB/s (2.8%)
- NEON read 2 data streams prefetched (64 bytes step) : 803.7 MB/s (3.9%)
- NEON copy : 290.8 MB/s (1.6%)
- NEON copy prefetched (32 bytes step) : 303.0 MB/s (1.4%)
- NEON copy prefetched (64 bytes step) : 303.4 MB/s (1.7%)
- NEON unrolled copy : 296.7 MB/s (1.7%)
- NEON unrolled copy prefetched (32 bytes step) : 301.8 MB/s (1.6%)
- NEON unrolled copy prefetched (64 bytes step) : 303.3 MB/s (1.6%)
- NEON copy backwards : 232.9 MB/s (7.8%)
- NEON copy backwards prefetched (32 bytes step) : 281.4 MB/s (1.6%)
- NEON copy backwards prefetched (64 bytes step) : 291.8 MB/s (1.8%)
- NEON 2-pass copy : 318.5 MB/s (1.5%)
- NEON 2-pass copy prefetched (32 bytes step) : 326.5 MB/s (1.5%)
- NEON 2-pass copy prefetched (64 bytes step) : 329.6 MB/s (1.9%)
- NEON unrolled 2-pass copy : 316.5 MB/s (1.5%)
- NEON unrolled 2-pass copy prefetched (32 bytes step) : 325.7 MB/s (1.8%)
- NEON unrolled 2-pass copy prefetched (64 bytes step) : 329.5 MB/s (1.9%)
- NEON fill : 619.5 MB/s (1.4%)
- NEON fill backwards : 631.6 MB/s (2.1%)
- VFP copy : 297.1 MB/s (1.7%)
- VFP 2-pass copy : 318.9 MB/s (1.8%)
- ARM fill (STRD) : 619.7 MB/s
- ARM fill (STM with 8 registers) : 619.2 MB/s (1.6%)
- ARM fill (STM with 4 registers) : 619.7 MB/s (2.8%)
- ARM copy prefetched (incr pld) : 302.6 MB/s (1.9%)
- ARM copy prefetched (wrap pld) : 302.7 MB/s (1.9%)
- ARM 2-pass copy prefetched (incr pld) : 328.4 MB/s (1.9%)
- ARM 2-pass copy prefetched (wrap pld) : 327.2 MB/s (2.0%)
- ==========================================================================
- == 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 : 12.1 ns / 16.6 ns
- 262144 : 16.8 ns / 20.8 ns
- 524288 : 23.7 ns / 29.0 ns
- 1048576 : 107.9 ns / 184.4 ns
- 2097152 : 152.3 ns / 259.0 ns
- 4194304 : 176.2 ns / 303.9 ns
- 8388608 : 193.5 ns / 320.5 ns
- 16777216 : 203.3 ns / 340.1 ns
- 33554432 : 217.8 ns / 362.9 ns
- 67108864 : 237.5 ns / 399.7 ns
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement