Skip to content

Linux_Performance_Guide.rst

Latest commit

 

History

History
1036 lines (701 loc) · 33.7 KB

Linux_Performance_Guide.rst

File metadata and controls

1036 lines (701 loc) · 33.7 KB

Linux 11.01 Performance Guide

Read This First

All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.

Name Description
AM62Px SK AM62Px Starter Kit rev E1 with ARM running at 1.4GHz, DDR data rate 3200 MT/S

Table: Evaluation Modules

About This Manual

This document provides performance data for each of the device drivers which are part of the Processor SDK Linux package. This document should be used in conjunction with release notes and user guides provided with the Processor SDK Linux package for information on specific issues present with drivers included in a particular release.

If You Need Assistance

For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/

System Benchmarks

LMBench

LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at http://lmbench.sourceforge.net/whatis_lmbench.html and http://lmbench.sourceforge.net/man/lmbench.8.html

Latency: lat_mem_rd-stride128-szN, where N is equal to or smaller than the cache size at given level measures the cache miss penalty. N that is at least double the size of last level cache is the latency to external memory.

Bandwidth: bw_mem_bcopy-N, where N is equal to or smaller than the cache size at a given level measures the achievable memory bandwidth from software doing a memcpy() type operation. Typical use is for external memory bandwidth calculation. The bandwidth is calculated as byte read and written counts as 1 which should be roughly half of STREAM copy result.

Execute the LMBench with the following:

cd /opt/ltp
./runltp -P j721e-idk-gw -f ddt/lmbench -s LMBENCH_L_PERF_0001
LMBench Benchmarks
Benchmarks am62pxx_sk-fs: perf
af_unix_sock_stream_latency (microsec) 30.16
af_unix_socket_stream_bandwidth (MBs) 1195.29
bw_file_rd-io-1mb (MB/s) 1441.44
bw_file_rd-o2c-1mb (MB/s) 802.95
bw_mem-bcopy-16mb (MB/s) 1962.71
bw_mem-bcopy-1mb (MB/s) 2096.88
bw_mem-bcopy-2mb (MB/s) 1846.15
bw_mem-bcopy-4mb (MB/s) 1671.54
bw_mem-bcopy-8mb (MB/s) 1938.45
bw_mem-bzero-16mb (MB/s) 8114.96
bw_mem-bzero-1mb (MB/s) 5095.88 (min 2096.88, max 8094.88)
bw_mem-bzero-2mb (MB/s) 4970.95 (min 1846.15, max 8095.74)
bw_mem-bzero-4mb (MB/s) 4891.20 (min 1671.54, max 8110.85)
bw_mem-bzero-8mb (MB/s) 5023.97 (min 1938.45, max 8109.48)
bw_mem-cp-16mb (MB/s) 981.17
bw_mem-cp-1mb (MB/s) 4660.98 (min 968.52, max 8353.44)
bw_mem-cp-2mb (MB/s) 4599.50 (min 955.26, max 8243.73)
bw_mem-cp-4mb (MB/s) 4596.81 (min 1006.67, max 8186.94)
bw_mem-cp-8mb (MB/s) 4586.70 (min 1025.38, max 8148.02)
bw_mem-fcp-16mb (MB/s) 1831.08
bw_mem-fcp-1mb (MB/s) 4939.98 (min 1785.08, max 8094.88)
bw_mem-fcp-2mb (MB/s) 4916.98 (min 1738.22, max 8095.74)
bw_mem-fcp-4mb (MB/s) 4926.51 (min 1742.16, max 8110.85)
bw_mem-fcp-8mb (MB/s) 4962.49 (min 1815.50, max 8109.48)
bw_mem-frd-16mb (MB/s) 1847.79
bw_mem-frd-1mb (MB/s) 1909.49 (min 1785.08, max 2033.90)
bw_mem-frd-2mb (MB/s) 1869.31 (min 1738.22, max 2000.40)
bw_mem-frd-4mb (MB/s) 1831.85 (min 1742.16, max 1921.54)
bw_mem-frd-8mb (MB/s) 1872.19 (min 1815.50, max 1928.87)
bw_mem-fwr-16mb (MB/s) 8128.70
bw_mem-fwr-1mb (MB/s) 5193.67 (min 2033.90, max 8353.44)
bw_mem-fwr-2mb (MB/s) 5122.07 (min 2000.40, max 8243.73)
bw_mem-fwr-4mb (MB/s) 5054.24 (min 1921.54, max 8186.94)
bw_mem-fwr-8mb (MB/s) 5038.45 (min 1928.87, max 8148.02)
bw_mem-rd-16mb (MB/s) 2065.05
bw_mem-rd-1mb (MB/s) 2124.95 (min 1916.38, max 2333.51)
bw_mem-rd-2mb (MB/s) 1874.93 (min 1704.16, max 2045.69)
bw_mem-rd-4mb (MB/s) 1944.50 (min 1829.27, max 2059.73)
bw_mem-rd-8mb (MB/s) 2000.01 (min 1930.97, max 2069.05)
bw_mem-rdwr-16mb (MB/s) 1983.88
bw_mem-rdwr-1mb (MB/s) 1404.23 (min 968.52, max 1839.93)
bw_mem-rdwr-2mb (MB/s) 1272.39 (min 955.26, max 1589.51)
bw_mem-rdwr-4mb (MB/s) 1417.41 (min 1006.67, max 1828.15)
bw_mem-rdwr-8mb (MB/s) 1493.08 (min 1025.38, max 1960.78)
bw_mem-wr-16mb (MB/s) 1972.87
bw_mem-wr-1mb (MB/s) 1878.16 (min 1839.93, max 1916.38)
bw_mem-wr-2mb (MB/s) 1646.84 (min 1589.51, max 1704.16)
bw_mem-wr-4mb (MB/s) 1828.71 (min 1828.15, max 1829.27)
bw_mem-wr-8mb (MB/s) 1945.88 (min 1930.97, max 1960.78)
bw_mmap_rd-mo-1mb (MB/s) 2246.03
bw_mmap_rd-o2c-1mb (MB/s) 796.31
bw_pipe (MB/s) 780.02
bw_unix (MB/s) 1195.29
lat_connect (us) 56.93
lat_ctx-2-128k (us) 7.60
lat_ctx-2-256k (us) 7.56
lat_ctx-4-128k (us) 7.18
lat_ctx-4-256k (us) 6.51
lat_fs-0k (num_files) 248.00
lat_fs-10k (num_files) 108.00
lat_fs-1k (num_files) 178.00
lat_fs-4k (num_files) 175.00
lat_mem_rd-stride128-sz1000k (ns) 29.18
lat_mem_rd-stride128-sz125k (ns) 5.62
lat_mem_rd-stride128-sz250k (ns) 5.83
lat_mem_rd-stride128-sz31k (ns) 3.65
lat_mem_rd-stride128-sz50 (ns) 2.15
lat_mem_rd-stride128-sz500k (ns) 12.89
lat_mem_rd-stride128-sz62k (ns) 5.22
lat_mmap-1m (us) 56.00
lat_ops-double-add (ns) 2.86
lat_ops-double-div (ns) 15.74
lat_ops-double-mul (ns) 2.86
lat_ops-float-add (ns) 2.86
lat_ops-float-div (ns) 9.30
lat_ops-float-mul (ns) 2.86
lat_ops-int-add (ns) 0.72
lat_ops-int-bit (ns) 0.48
lat_ops-int-div (ns) 4.29
lat_ops-int-mod (ns) 4.53
lat_ops-int-mul (ns) 3.08
lat_ops-int64-add (ns) 0.72
lat_ops-int64-bit (ns) 0.48
lat_ops-int64-div (ns) 6.79
lat_ops-int64-mod (ns) 5.25
lat_ops-int64-mul (ns) 3.55
lat_pagefault (us) 0.50
lat_pipe (us) 24.46
lat_proc-exec (us) 682.50
lat_proc-fork (us) 590.20
lat_proc-proccall (us) 0.01
lat_select (us) 34.04
lat_sem (us) 3.53
lat_sig-catch (us) 5.30
lat_sig-install (us) 0.65
lat_sig-prot (us) 0.64
lat_syscall-fstat (us) 1.99
lat_syscall-null (us) 0.46
lat_syscall-open (us) 147.43
lat_syscall-read (us) 0.80
lat_syscall-stat (us) 4.85
lat_syscall-write (us) 0.82
lat_tcp (us) 0.93
lat_unix (us) 30.16
latency_for_0.50_mb_block_size (nanosec) 12.89
latency_for_1.00_mb_block_size (nanosec) 14.59 (min 0.00, max 29.18)
pipe_bandwidth (MBs) 780.02
pipe_latency (microsec) 24.46
procedure_call (microsec) 0.01
select_on_200_tcp_fds (microsec) 34.04
semaphore_latency (microsec) 3.53
signal_handler_latency (microsec) 0.65
signal_handler_overhead (microsec) 5.30
tcp_ip_connection_cost_to_localhost (microsec) 56.93
tcp_latency_using_localhost (microsec) 0.93

Dhrystone

Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed.

Please take note, different run may produce different slightly results. This is advised to run this test multiple times in order to get maximum performance numbers.

Execute the benchmark with the following:

runDhrystone
Dhrystone Benchmarks
Benchmarks am62pxx_sk-fs: perf
cpu_clock (MHz) 1400.00
dhrystone_per_mhz (DMIPS/MHz) 2.90
dhrystone_per_second (DhrystoneP) 7142857.00

Whetstone

Whetstone is a benchmark primarily measuring floating-point arithmetic performance.

Execute the benchmark with the following:

runWhetstone
Whetstone Benchmarks
Benchmarks am62pxx_sk-fs: perf
whetstone (MIPS) 10000.00

Linpack

Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system.

Linpack Benchmarks
Benchmarks am62pxx_sk-fs: perf
linpack (Kflops) 329621.00

NBench

NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at https://en.wikipedia.org/wiki/NBench and https://nbench.io/articles/index.html

NBench Benchmarks
Benchmarks am62pxx_sk-fs: perf
assignment (Iterations) 8.26
fourier (Iterations) 13044.00
fp_emulation (Iterations) 123.21
huffman (Iterations) 676.56
idea (Iterations) 1968.20
lu_decomposition (Iterations) 302.85
neural_net (Iterations) 4.95
numeric_sort (Iterations) 358.59
string_sort (Iterations) 93.66

Stream

STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. It uses double precision floating point (64bit) but in most modern processors the memory access will be the bottleneck. The four individual scores are copy, scale as in multiply by constant, add two numbers, and triad for multiply accumulate. For bandwidth, a byte read counts as one and a byte written counts as one, resulting in a score that is double the bandwidth LMBench will show.

Execute the benchmark with the following:

stream_c
Stream Benchmarks
Benchmarks am62pxx_sk-fs: perf
add (MB/s) 2719.00
copy (MB/s) 3791.20
scale (MB/s) 3426.70
triad (MB/s) 2418.50

CoreMarkPro

CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, adding comprehensive support for multicore technology, a combination of integer and floating-point workloads, and data sets for utilizing larger memory subsystems.

CoreMarkPro Benchmarks
Benchmarks am62pxx_sk-fs: perf
cjpeg-rose7-preset (workloads/) 41.84
core (workloads/) 0.30
coremark-pro () 918.00
linear_alg-mid-100x100-sp (workloads/) 14.67
loops-all-mid-10k-sp (workloads/) 0.71
nnet_test (workloads/) 1.09
parser-125k (workloads/) 8.77
radix2-big-64k (workloads/) 59.36
sha-test (workloads/) 81.30
zip-test (workloads/) 21.74
CoreMarkProTwoCore Benchmarks
Benchmarks am62pxx_sk-fs: perf
cjpeg-rose7-preset (workloads/) 84.03
core (workloads/) 0.60
coremark-pro () 1676.82
linear_alg-mid-100x100-sp (workloads/) 29.36
loops-all-mid-10k-sp (workloads/) 1.30
nnet_test (workloads/) 2.17
parser-125k (workloads/) 14.08
radix2-big-64k (workloads/) 71.98
sha-test (workloads/) 161.29
zip-test (workloads/) 43.48

MultiBench

MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and demonstrating a processor’s support for fine grain parallelism. Processing multiple data streams: uses common code running over multiple threads and demonstrating how well a processor scales over scalable data inputs. Multiple workload processing: shows the scalability of general-purpose processing, demonstrating concurrency over both code and data. MultiBench combines a wide variety of application-specific workloads with the EEMBC Multi-Instance-Test Harness (MITH), compatible and portable with most any multicore processors and operating systems. MITH uses a thread-based API (POSIX-compliant) to establish a common programming model that communicates with the benchmark through an abstraction layer and provides a flexible interface to allow a wide variety of thread-enabled workloads to be tested.

Multibench Benchmarks
Benchmarks am62pxx_sk-fs: perf
4m-check (workloads/) 428.45
4m-check-reassembly (workloads/) 126.90
4m-check-reassembly-tcp (workloads/) 61.12
4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/) 34.48
4m-check-reassembly-tcp-x264w2 (workloads/) 1.91
4m-cmykw2 (workloads/) 243.90
4m-cmykw2-rotatew2 (workloads/) 51.64
4m-reassembly (workloads/) 87.11
4m-rotatew2 (workloads/) 54.23
4m-tcp-mixed (workloads/) 120.30
4m-x264w2 (workloads/) 1.94
empty-wld (workloads/) 1.00
idct-4m (workloads/) 19.28
idct-4mw1 (workloads/) 19.29
ippktcheck-4m (workloads/) 431.93
ippktcheck-4mw1 (workloads/) 430.66
ipres-4m (workloads/) 118.02
ipres-4mw1 (workloads/) 117.46
md5-4m (workloads/) 27.98
md5-4mw1 (workloads/) 28.02
rgbcmyk-4m (workloads/) 62.75
rgbcmyk-4mw1 (workloads/) 62.72
rotate-4ms1 (workloads/) 24.08
rotate-4ms1w1 (workloads/) 24.15
rotate-4ms64 (workloads/) 24.35
rotate-4ms64w1 (workloads/) 24.32
x264-4mq (workloads/) 0.58
x264-4mqw1 (workloads/) 0.58

Boot-time Measurement

Boot media: MMCSD

Linux boot time MMCSD
Boot Configuration am62pxx_sk-fs: Boot time in seconds: avg(min,max)
Linux boot time from SD with default rootfs (20 boot cycles) 13.63 (min 13.22, max 14.01)

Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles.


ALSA SoC Audio Driver

  1. Access type - RW_INTERLEAVED
  2. Channels - 2
  3. Format - S16_LE
  4. Period size - 64
Audio Capture
Sampling Rate (Hz) am62pxx_sk-fs: Throughput (bits/sec) am62pxx_sk-fs: CPU Load (%)
11025 352798.00 0.10
16000 511998.00 0.09
22050 705597.00 0.13
24000 705597.00 0.13
32000 1023995.00 0.10
44100 1411193.00 0.21
48000 1535992.00 0.16
88200 2822382.00 0.35
96000 3071979.00 0.18
Audio Playback
Sampling Rate (Hz) am62pxx_sk-fs: Throughput (bits/sec) am62pxx_sk-fs: CPU Load (%)
11025 352944.00 0.14
16000 512210.00 0.26
22050 705885.00 0.17
24000 705887.00 0.19
32000 1024418.00 0.16
44100 1411775.00 0.28
48000 1536626.00 0.23
88200 2823541.00 0.46
96000 3073248.00 0.31

Graphics SGX/RGX Driver

GFXBench

Run GFXBench and capture performance reported (Score and Display rate in fps). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests

GFXBench Performance
Benchmark am62pxx_sk-fs: Score am62pxx_sk-fs: Fps
GFXBench 3.x gl_manhattan_off
 907.99 14.65
GFXBench 3.x gl_trex_off
 1595.80 28.50
GFXBench 4.x gl_4_off
 260.27 4.40
GFXBench 5.x gl_5_high_off
 113.55 1.77

Glmark2

Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests

Glmark2 Performance
Benchmark am62pxx_sk-fs: Score
Glmark2-DRM 355.00
Glmark2-Wayland 734.00

Ethernet

Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, it is recommended to carry out performance tests in a private network and to avoid running NFS on the same interface used in the test. In these results, CPU utilization was captured as the total percentage used across all cores on the device, while running the performance test over one external interface.

UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram:

burst_size = <bandwidth (bits/sec)> / 8 (bits -> bytes) / <UDP datagram size> / 100 (seconds -> 10 ms)
burst_size = 500000000 / 8 / 1472 / 100 = 425

wait_time = 10 milliseconds (minimum supported by Linux PC used for testing)

UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when running the netperf test with no bandwidth limit (remove -b/-w options).

In order to start a netperf client on one device, the other device must have netserver running. To start netserver:

netserver [-p <port_number>] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)]

Running the following shell script from the DUT will trigger netperf clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in client commands to summarize selected statistics on their own line and -j is used to gain additional timing measurements during the test.

#!/bin/bash
for i in 1
do
   netperf -H <tester ip> -j -c -l 60 -t TCP_STREAM --
      -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE &

   netperf -H <tester ip> -j -c -l 60 -t TCP_MAERTS --
      -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE &
done

Running the following commands will trigger netperf clients to measure UDP burst performance for 60 seconds at various burst/datagram sizes and report CPU utilization.

  • For UDP egress tests, run netperf client from DUT and start netserver on tester.
netperf -H <tester ip> -j -c -l 60 -t UDP_STREAM -b <burst_size> -w <wait_time> -- -m <UDP datagram size>
   -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE
  • For UDP ingress tests, run netperf client from tester and start netserver on DUT.
netperf -H <DUT ip> -j -C -l 60 -t UDP_STREAM -b <burst_size> -w <wait_time> -- -m <UDP datagram size>
   -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE

CPSW/CPSW2g/CPSW3g Ethernet Driver

  • CPSW3g: AM64x, AM62x, AM62ax, AM62px

TCP Bidirectional Throughput

CPSW2g TCP Bidirectional Throughput
Command Used am62pxx_sk-fs: THROUGHPUT (Mbits/sec) am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)
netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS 1818.42 64.41

TCP Bidirectional Throughput Interrupt Pacing

CPSW2g TCP Bidirectional Throughput Interrupt Pacing
Command Used am62pxx_sk-fs: THROUGHPUT (Mbits/sec) am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)
netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS 1812.61 29.43

UDP Throughput

CPSW2g UDP Egress Throughput 0 loss
Frame Size(bytes) am62pxx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE) am62pxx_sk-fs: THROUGHPUT (Mbits/sec) am62pxx_sk-fs: Packets Per Second (kPPS) am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)
64 18.00 55.07 108.00 39.97
128 82.00 108.69 106.00 39.97
256 210.00 218.80 107.00 39.98
1024 978.00 839.58 102.00 39.48
1518 1472.00 856.67 71.00 37.44

Linux OSPI Flash Driver

AM62PXX-SK

UBIFS

OSPI Flash Driver
Buffer size (bytes) am62pxx_sk-fs: Write UBIFS Throughput (Mbytes/sec) am62pxx_sk-fs: Write UBIFS CPU Load (%) am62pxx_sk-fs: Read UBIFS Throughput (Mbytes/sec) am62pxx_sk-fs: Read UBIFS CPU Load (%)
102400 0.18 (min 0.13, max 0.29) 27.43 (min 24.77, max 30.04) 28.58 12.90
262144 0.15 (min 0.11, max 0.19) 28.31 (min 26.75, max 30.10) 28.55 6.90
524288 0.15 (min 0.11, max 0.19) 28.51 (min 27.57, max 30.32) 28.26 12.90
1048576 0.15 (min 0.11, max 0.19) 28.08 (min 25.48, max 30.41) 28.21 9.68

RAW

OSPI Raw Flash Driver
File size (Mbytes) am62pxx_sk-fs: Raw Read Throughput (Mbytes/sec)
50 37.88

EMMC Driver

Warning

IMPORTANT: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode.

EMMC EXT4 FIO 1G

EMMC EXT4 FIO 1G
Buffer size (bytes) am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Write EXT4 CPU Load (%) am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Read EXT4 CPU Load (%)
1m 90.50 1.61 171.00 1.83
4m 95.60 0.99 172.00 1.41
4k 64.30 21.67 93.40 22.26
256k 90.50 2.01 171.00 2.62

EMMC RAW FIO 1G

EMMC RAW FIO 1G
Buffer size (bytes) am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Write Raw CPU Load (%) am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Read Raw CPU Load (%)
1m 90.10 1.38 174.00 1.76
4m 97.00 1.01 138.00 1.20
4k 64.70 16.10 92.40 20.30
256k 90.20 1.85 173.00 2.45

EMMC EXT4

EMMC EXT4
Buffer size (bytes) am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Write EXT4 CPU Load (%) am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Read EXT4 CPU Load (%)
102400 83.74 (min 81.64, max 85.23) 5.78 (min 4.97, max 7.62) 178.99 8.97
262144 82.29 (min 61.99, max 90.58) 5.84 (min 3.99, max 8.08) 181.07 9.48
524288 79.81 (min 51.54, max 89.29) 5.46 (min 3.20, max 8.02) 181.88 9.52
1048576 73.44 (min 49.22, max 89.90) 5.08 (min 2.94, max 8.17) 181.74 8.66
5242880 74.59 (min 50.87, max 87.83) 5.30 (min 3.27, max 8.33) 181.91 8.30

EMMC EXT2

EMMC EXT2
Buffer size (bytes) am62pxx_sk-fs: Write EXT2 Throughput (Mbytes/sec) am62pxx_sk-fs: Write EXT2 CPU Load (%) am62pxx_sk-fs: Read EXT2 Throughput (Mbytes/sec) am62pxx_sk-fs: Read EXT2 CPU Load (%)
102400 36.82 (min 7.02, max 46.19) 4.08 (min 2.94, max 4.65) 140.53 8.00
262144 50.11 (min 20.89, max 59.79) 5.67 (min 2.87, max 11.52) 177.12 14.41
524288 51.44 (min 7.44, max 67.81) 4.78 (min 2.24, max 6.37) 137.86 7.79
1048576 52.32 (min 7.48, max 68.81) 4.22 (min 2.35, max 5.42) 152.16 7.97
5242880 54.76 (min 7.50, max 70.98) 4.21 (min 2.39, max 4.92) 152.90 6.32

EMMC VFAT

EMMC VFAT
Buffer size (bytes) am62pxx_sk-fs: Write VFAT Throughput (Mbytes/sec) am62pxx_sk-fs: Write VFAT CPU Load (%) am62pxx_sk-fs: Read VFAT Throughput (Mbytes/sec) am62pxx_sk-fs: Read VFAT CPU Load (%)
102400 35.84 (min 11.13, max 50.42) 4.93 (min 3.85, max 6.36) 173.89 8.40
262144 45.10 (min 11.46, max 58.09) 6.41 (min 4.90, max 7.20) 175.12 14.23
524288 50.40 (min 12.70, max 72.57) 4.73 (min 3.83, max 5.93) 176.19 8.82
1048576 57.99 (min 12.88, max 75.35) 5.09 (min 4.03, max 5.97) 176.20 9.21
5242880 56.58 (min 13.07, max 82.67) 5.09 (min 3.91, max 6.71) 176.26 9.62

UBoot EMMC Driver

UBOOT EMMC RAW
File size (bytes in hex) am62pxx_sk-fs: Write Throughput (Kbytes/sec) am62pxx_sk-fs: Read Throughput (Kbytes/sec)
2000000 98402.40 172463.16
4000000 96946.75 174762.67

MMCSD

Warning

IMPORTANT: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode.

MMC EXT4 FIO 1G

MMC EXT4 FIO 1G
Buffer size (bytes) am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Write EXT4 CPU Load (%) am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec) am62pxx_sk-fs: Read EXT4 CPU Load (%)
1m 42.10 1.04 87.20 1.28
4m 42.90 0.79 87.30 0.97
4k 2.78 1.69 12.80 4.25
256k 38.10 1.22 83.70 1.55

MMC RAW FIO 1G

MMC RAW FIO 1G
Buffer size (bytes) am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Write Raw CPU Load (%) am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Read Raw CPU Load (%)
1m 43.10 1.40 88.00 1.87
4m 43.60 1.02 88.00 1.40
4k 2.77 2.21 13.00 6.68
256k 36.80 1.56 84.00 2.38

MMC EXT4

MMC EXT4
Buffer size (bytes) am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Write Raw CPU Load (%) am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Read Raw CPU Load (%)
102400 10.37 (min 10.30, max 10.63) 0.80 (min 0.69, max 1.12) 10.68 0.76
262144 10.36 (min 10.28, max 10.49) 0.85 (min 0.63, max 1.32) 11.22 0.67
524288 10.40 (min 10.33, max 10.47) 0.83 (min 0.72, max 1.05) 11.50 0.63
1048576 10.45 (min 10.40, max 10.55) 0.86 (min 0.80, max 1.03) 11.05 0.66
5242880 10.65 (min 10.31, max 10.87) 0.83 (min 0.70, max 1.08) 12.03 0.66

MMC EXT3

MMC EXT3
Buffer size (bytes) am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Write Raw CPU Load (%) am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Read Raw CPU Load (%)
102400 40.74 (min 35.43, max 45.16) 3.67 (min 2.97, max 6.09) 80.08 4.23
262144 40.14 (min 35.43, max 41.78) 3.57 (min 2.82, max 5.91) 82.32 5.14
524288 36.55 (min 32.31, max 38.14) 3.16 (min 2.55, max 5.39) 90.35 4.75
1048576 36.11 (min 31.85, max 37.76) 3.26 (min 2.62, max 5.39) 90.20 4.95
5242880 39.66 (min 34.22, max 42.15) 3.30 (min 2.58, max 5.32) 90.47 4.98

MMC EXT2

MMC EXT2
Buffer size (bytes) am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Write Raw CPU Load (%) am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec) am62pxx_sk-fs: Read Raw CPU Load (%)
102400 17.47 (min 3.49, max 21.09) 2.33 (min 2.14, max 2.47) 76.97 4.41
262144 31.21 (min 3.88, max 38.66) 3.17 (min 2.06, max 3.75) 81.53 4.50
524288 34.84 (min 3.99, max 42.82) 3.34 (min 2.05, max 4.00) 90.76 4.76
1048576 33.57 (min 4.02, max 41.12) 2.83 (min 2.09, max 3.23) 90.54 4.35
5242880 32.54 (min 4.03, max 39.87) 2.62 (min 2.03, max 2.86) 90.56 4.55

The performance numbers were captured using the following:

  • SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA)
  • Partition was mounted with async option

UBoot MMCSD

UBOOT MMCSD FAT

UBOOT MMCSD FAT
File size (bytes in hex) am62pxx_sk-fs: Write Throughput (Kbytes/sec) am62pxx_sk-fs: Read Throughput (Kbytes/sec)
400000 38280.37 81920.00
800000 40554.46 87148.94
1000000 42666.67 90021.98

The performance numbers were captured using the following:

  • SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA)

USB Driver

USB Device Controller

USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT
Number of Blocks am62pxx_sk-fs: Throughput (MB/sec)
150 44.00
USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT
Number of Blocks am62pxx_sk-fs: Throughput (MB/sec)
150 37.30

CRYPTO Driver

OpenSSL Performance

OpenSSL Performance
Algorithm Buffer Size (in bytes) am62pxx_sk-fs: throughput (KBytes/Sec)
aes-128-cbc 1024 23455.06
aes-128-cbc 16 421.01
aes-128-cbc 16384 85224.11
aes-128-cbc 256 6907.22
aes-128-cbc 64 1825.54
aes-128-cbc 8192 71923.03
aes-128-ecb 1024 23801.51
aes-128-ecb 16 430.91
aes-128-ecb 16384 87780.01
aes-128-ecb 256 7053.99
aes-128-ecb 64 1870.70
aes-128-ecb 8192 73659.73
aes-192-cbc 1024 22887.42
aes-192-cbc 16 423.09
aes-192-cbc 16384 76808.19
aes-192-cbc 256 6910.29
aes-192-cbc 64 1825.79
aes-192-cbc 8192 65912.83
aes-192-ecb 1024 23310.34
aes-192-ecb 16 429.07
aes-192-ecb 16384 79123.80
aes-192-ecb 256 6940.50
aes-192-ecb 64 1837.50
aes-192-ecb 8192 67548.50
aes-256-cbc 1024 22148.10
aes-256-cbc 16 421.00
aes-256-cbc 16384 69435.39
aes-256-cbc 256 6811.73
aes-256-cbc 64 1828.39
aes-256-cbc 8192 60792.83
aes-256-ecb 1024 22760.45
aes-256-ecb 16 432.77
aes-256-ecb 16384 72286.21
aes-256-ecb 256 6953.05
aes-256-ecb 64 1869.91
aes-256-ecb 8192 62401.19
sha256 1024 34637.82
sha256 16 574.17
sha256 16384 284317.01
sha256 256 8920.66
sha256 64 2263.98
sha256 8192 190515.88
sha512 1024 24553.81
sha512 16 560.17
sha512 16384 67567.62
sha512 256 8053.59
sha512 64 2252.35
sha512 8192 60293.12
OpenSSL CPU Load
Algorithm am62pxx_sk-fs: CPU Load
aes-128-cbc 31.00
aes-128-ecb 32.00
aes-192-cbc 31.00
aes-192-ecb 32.00
aes-256-cbc 30.00
aes-256-ecb 31.00
sha256 96.00
sha512 96.00
Listed for each algorithm are the code snippets used to run each
benchmark test.
time -v openssl speed -elapsed -evp aes-128-cbc

Low Power Performance

Power Performance

Deep Sleep Power Performance
Rail name Rail voltage(V) Power (mW)
vdd_core 0.85 8.95
vddr_core 0.85 0.84
soc_dvdd_3v3 3.30 7.16
soc_dvdd_1v8 1.80 2.59
vdda_1v8 1.80 68.66
vdd_lpddr4 1.10 0.46
Total   88.66
MCU Only Power Performance
Rail name Rail voltage(V) Power (mW)
vdd_core 0.85 211.34
vddr_core 0.85 2.51
soc_dvdd_3v3 3.30 6.81
soc_dvdd_1v8 1.80 2.63
vdda_1v8 1.80 77.43
vdd_lpddr4 1.10 0.42
Total   301.14

Partial I/O Data - All voltage rails were measured to be near 0V

Note

The power consumption on the vdda_1v8 rail is not indicitive of the SoC's power consumption due to an oscillator on the rail that has significant current consumption.

Note

The measurements shown are from an AM62Px SK rev E1-1. Results may vary based off of the board revision being used.

Further optimizations are possible for these low power modes. Please refer to the AM62x Power Consumption App Note (https://www.ti.com/lit/pdf/spradg1)

Resume Latency Performance

LPM Resume Latency Performance
Low Power Mode Total Resume Latency (ms)
I/O Only + DDR 715.09
Deep Sleep 143.33
MCU Only 94.17

The performance numbers are measured without the Linux printk logs. To remove the Linux printk logs, run the following commands in the terminal:

# Detach kernel serial console
consoles=$(find /sys/bus/platform/devices/*.serial/ -name console)
for console in ${consoles}; do
     echo -n N > ${console}
done

Note

The measurements shown are from using the default SDK with no extra optimizations.