Platform/GFX/MemoryBandwidth
Jump to navigation
Jump to search
Description
========================================================================== == 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. == ========================================================================== ========================================================================== == 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 == ==========================================================================
MacBookPro
==========================================================================
== Memory bandwidth tests ==
==========================================================================
C copy backwards : 9805.2 MB/s (2.5%)
C copy : 9832.3 MB/s (6.9%)
C copy prefetched (32 bytes step) : 9556.3 MB/s (2.5%)
C copy prefetched (64 bytes step) : 9532.4 MB/s (2.1%)
C 2-pass copy : 9445.8 MB/s (1.6%)
C 2-pass copy prefetched (32 bytes step) : 9120.2 MB/s (2.7%)
C 2-pass copy prefetched (64 bytes step) : 9102.7 MB/s (2.2%)
C fill : 16685.5 MB/s (4.3%)
---
standard memcpy : 11950.6 MB/s (1.8%)
standard memset : 23193.5 MB/s (1.9%)
==========================================================================
== Memory latency test ==
==========================================================================
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 : 1.0 ns / 1.2 ns
131072 : 1.5 ns / 1.7 ns
262144 : 1.8 ns / 2.0 ns
524288 : 6.4 ns / 8.0 ns
1048576 : 8.9 ns / 10.2 ns
2097152 : 10.7 ns / 11.3 ns
4194304 : 13.7 ns / 14.0 ns
8388608 : 36.8 ns / 45.9 ns
16777216 : 42.8 ns / 47.1 ns
33554432 : 54.9 ns / 49.5 ns
67108864 : 81.9 ns / 94.9 ns
Flame
==========================================================================
== Memory bandwidth tests ==
==========================================================================
C copy backwards : 337.7 MB/s (0.9%)
C copy : 999.7 MB/s (0.6%)
C copy prefetched (32 bytes step) : 866.7 MB/s (0.4%)
C copy prefetched (64 bytes step) : 867.4 MB/s (0.7%)
C 2-pass copy : 660.4 MB/s (0.3%)
C 2-pass copy prefetched (32 bytes step) : 613.3 MB/s
C 2-pass copy prefetched (64 bytes step) : 613.8 MB/s (0.1%)
C fill : 2501.7 MB/s (0.2%)
---
standard memcpy : 1017.0 MB/s (0.4%)
standard memset : 2494.4 MB/s (0.3%)
---
NEON read : 1817.7 MB/s (0.4%)
NEON read prefetched (32 bytes step) : 1899.6 MB/s (0.3%)
NEON read prefetched (64 bytes step) : 1915.5 MB/s (0.2%)
NEON read 2 data streams : 550.9 MB/s (0.1%)
NEON read 2 data streams prefetched (32 bytes step) : 1015.1 MB/s (0.1%)
NEON read 2 data streams prefetched (64 bytes step) : 1087.3 MB/s (0.1%)
NEON copy : 1016.5 MB/s (0.5%)
NEON copy prefetched (32 bytes step) : 1042.1 MB/s (0.3%)
NEON copy prefetched (64 bytes step) : 1063.6 MB/s (0.4%)
NEON unrolled copy : 993.6 MB/s (0.3%)
NEON unrolled copy prefetched (32 bytes step) : 961.7 MB/s (0.3%)
NEON unrolled copy prefetched (64 bytes step) : 1007.5 MB/s
NEON copy backwards : 1016.1 MB/s (0.4%)
NEON copy backwards prefetched (32 bytes step) : 1041.8 MB/s (0.4%)
NEON copy backwards prefetched (64 bytes step) : 1068.9 MB/s (0.4%)
NEON 2-pass copy : 984.4 MB/s (0.4%)
NEON 2-pass copy prefetched (32 bytes step) : 1022.3 MB/s (0.3%)
NEON 2-pass copy prefetched (64 bytes step) : 1037.6 MB/s (0.5%)
NEON unrolled 2-pass copy : 881.1 MB/s (0.2%)
NEON unrolled 2-pass copy prefetched (32 bytes step) : 823.4 MB/s (0.3%)
NEON unrolled 2-pass copy prefetched (64 bytes step) : 885.6 MB/s (0.3%)
NEON fill : 2501.0 MB/s (0.3%)
NEON fill backwards : 2498.9 MB/s (0.4%)
VFP copy : 996.2 MB/s (0.4%)
VFP 2-pass copy : 903.9 MB/s (0.3%)
ARM fill (STRD) : 2498.0 MB/s (0.5%)
ARM fill (STM with 8 registers) : 2499.9 MB/s (0.2%)
ARM fill (STM with 4 registers) : 2498.9 MB/s (0.4%)
ARM copy prefetched (incr pld) : 1060.5 MB/s (0.4%)
ARM copy prefetched (wrap pld) : 1006.0 MB/s
ARM 2-pass copy prefetched (incr pld) : 1005.7 MB/s (0.3%)
ARM 2-pass copy prefetched (wrap pld) : 984.0 MB/s (0.4%)
==========================================================================
== Memory latency test ==
==========================================================================
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 : 5.3 ns / 9.1 ns
131072 : 8.2 ns / 12.8 ns
262144 : 9.6 ns / 14.3 ns
524288 : 12.8 ns / 17.4 ns
1048576 : 69.1 ns / 109.5 ns
2097152 : 104.4 ns / 147.4 ns
4194304 : 122.6 ns / 162.9 ns
8388608 : 133.1 ns / 171.6 ns
16777216 : 140.9 ns / 180.5 ns
33554432 : 148.5 ns / 193.0 ns
67108864 : 167.0 ns / 221.8 ns
Tarako
==========================================================================
== Memory bandwidth tests ==
==========================================================================
C copy backwards : 121.2 MB/s (0.2%)
C copy : 370.5 MB/s
C copy prefetched (32 bytes step) : 381.3 MB/s (0.4%)
C copy prefetched (64 bytes step) : 381.6 MB/s (0.7%)
C 2-pass copy : 425.6 MB/s (0.5%)
C 2-pass copy prefetched (32 bytes step) : 407.2 MB/s (0.5%)
C 2-pass copy prefetched (64 bytes step) : 406.9 MB/s (0.4%)
C fill : 1523.0 MB/s (0.6%)
---
standard memcpy : 363.8 MB/s (0.4%)
standard memset : 1523.4 MB/s (0.7%)
---
NEON read : 1817.7 MB/s (0.4%)
NEON read prefetched (32 bytes step) : 1899.6 MB/s (0.3%)
NEON read prefetched (64 bytes step) : 1915.5 MB/s (0.2%)
NEON read 2 data streams : 550.9 MB/s (0.1%)
NEON read 2 data streams prefetched (32 bytes step) : 1015.1 MB/s (0.1%)
NEON read 2 data streams prefetched (64 bytes step) : 1087.3 MB/s (0.1%)
NEON copy : 1016.5 MB/s (0.5%)
NEON copy prefetched (32 bytes step) : 1042.1 MB/s (0.3%)
NEON copy prefetched (64 bytes step) : 1063.6 MB/s (0.4%)
NEON unrolled copy : 993.6 MB/s (0.3%)
NEON unrolled copy prefetched (32 bytes step) : 961.7 MB/s (0.3%)
NEON unrolled copy prefetched (64 bytes step) : 1007.5 MB/s
NEON copy backwards : 1016.1 MB/s (0.4%)
NEON copy backwards prefetched (32 bytes step) : 1041.8 MB/s (0.4%)
NEON copy backwards prefetched (64 bytes step) : 1068.9 MB/s (0.4%)
NEON 2-pass copy : 984.4 MB/s (0.4%)
NEON 2-pass copy prefetched (32 bytes step) : 1022.3 MB/s (0.3%)
NEON 2-pass copy prefetched (64 bytes step) : 1037.6 MB/s (0.5%)
NEON unrolled 2-pass copy : 881.1 MB/s (0.2%)
NEON unrolled 2-pass copy prefetched (32 bytes step) : 823.4 MB/s (0.3%)
NEON unrolled 2-pass copy prefetched (64 bytes step) : 885.6 MB/s (0.3%)
NEON fill : 2501.0 MB/s (0.3%)
NEON fill backwards : 2498.9 MB/s (0.4%)
VFP copy : 996.2 MB/s (0.4%)
VFP 2-pass copy : 903.9 MB/s (0.3%)
ARM fill (STRD) : 2498.0 MB/s (0.5%)
ARM fill (STM with 8 registers) : 2499.9 MB/s (0.2%)
ARM fill (STM with 4 registers) : 2498.9 MB/s (0.4%)
ARM copy prefetched (incr pld) : 1060.5 MB/s (0.4%)
ARM copy prefetched (wrap pld) : 1006.0 MB/s
ARM 2-pass copy prefetched (incr pld) : 1005.7 MB/s (0.3%)
ARM 2-pass copy prefetched (wrap pld) : 984.0 MB/s (0.4%)
==========================================================================
== Memory latency test ==
==========================================================================
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 : 5.3 ns / 9.1 ns
131072 : 8.2 ns / 12.8 ns
262144 : 9.6 ns / 14.3 ns
524288 : 12.8 ns / 17.4 ns
1048576 : 69.1 ns / 109.5 ns
2097152 : 104.4 ns / 147.4 ns
4194304 : 122.6 ns / 162.9 ns
8388608 : 133.1 ns / 171.6 ns
16777216 : 140.9 ns / 180.5 ns
33554432 : 148.5 ns / 193.0 ns
67108864 : 167.0 ns / 221.8 ns