Read This First
All performance numbers provided in this document are gathered using following Evaluation Modules unless otherwise specified.
| Name | Description |
|---|---|
| AM64x EVM | AM64x Evaluation Module rev E1 with ARM running at 1GHz, DDR data rate 1600 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/
stress-ng (next-generation) will stress test a embedded platform in various selectable ways. It was designed to exercise various physical subsystems as well as the various operating system kernel interfaces. stress-ng can also measure test throughput rates; this can be useful to observe performance changes across different operating system or types of hardware.
Cyclictest is most commonly used for benchmarking RT systems. It is one of the most frequently used tools for evaluating the relative performance of real-time systems. Some performance tests which use Cyclictest are System benchmarking, Latency debugging with tracing and -approximating application performance.
Test commands used for running stress-ng and cyclictest together
stress-ng --cpu-method=all -c 4 &
cyclictest -m -Sp80 -D6h -h400 -i200 -M -qThe following summarizes the latencies observed using the yocto based default SDK image:
Note
A known issue in this SDK release is affecting this benchmark. Using OP-TEE's PRNG drivers rather than the hardware accelerated TRNG drivers restores the context switch latencies to the values you see here.
More information on switching to the PRNG drivers can be found in the Foundational Components section, here :ref:`building-optee-with-prng`
| Latencies | CPU 0 | CPU 1 |
|---|---|---|
| Minimum (usec) | 6 | 6 |
| Average (usec) | 9 | 9 |
| Maximum (usec) | 58 | 73 |
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 https://lmbench.sourceforge.net/whatis_lmbench.html and https://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
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| af_unix_sock_stream_latency (microsec) | 42.81 (min 37.94, max 47.17) |
| af_unix_socket_stream_bandwidth (mbs) | 541.45 (min 530.54, max 556.53) |
| bw_file_rd-io-1mb (mb/s) | 875.51 (min 850.20, max 890.31) |
| bw_file_rd-o2c-1mb (mb/s) | 491.77 (min 456.00, max 508.82) |
| bw_mem-bcopy-16mb (mb/s) | 969.07 (min 937.70, max 1007.87) |
| bw_mem-bcopy-1mb (mb/s) | 947.26 (min 915.75, max 980.01) |
| bw_mem-bcopy-2mb (mb/s) | 934.26 (min 902.53, max 970.87) |
| bw_mem-bcopy-4mb (mb/s) | 949.68 (min 911.68, max 983.28) |
| bw_mem-bcopy-8mb (mb/s) | 976.43 (min 911.99, max 1010.61) |
| bw_mem-bzero-16mb (mb/s) | 2117.38 (min 2116.68, max 2118.08) |
| bw_mem-bzero-1mb (mb/s) | 1532.87 (min 915.75, max 2119.87) |
| bw_mem-bzero-2mb (mb/s) | 1524.64 (min 902.53, max 2117.90) |
| bw_mem-bzero-4mb (mb/s) | 1532.55 (min 911.68, max 2118.64) |
| bw_mem-bzero-8mb (mb/s) | 1546.42 (min 911.99, max 2118.36) |
| bw_mem-cp-16mb (mb/s) | 592.63 (min 576.54, max 600.85) |
| bw_mem-cp-1mb (mb/s) | 1585.65 (min 543.97, max 2718.45) |
| bw_mem-cp-2mb (mb/s) | 1432.40 (min 524.11, max 2363.14) |
| bw_mem-cp-4mb (mb/s) | 1376.53 (min 523.49, max 2214.43) |
| bw_mem-cp-8mb (mb/s) | 1352.70 (min 523.01, max 2157.79) |
| bw_mem-fcp-16mb (mb/s) | 1051.58 (min 1000.13, max 1099.43) |
| bw_mem-fcp-1mb (mb/s) | 1591.05 (min 1028.10, max 2119.87) |
| bw_mem-fcp-2mb (mb/s) | 1567.15 (min 987.33, max 2117.90) |
| bw_mem-fcp-4mb (mb/s) | 1565.96 (min 982.08, max 2118.64) |
| bw_mem-fcp-8mb (mb/s) | 1571.64 (min 971.46, max 2118.36) |
| bw_mem-frd-16mb (mb/s) | 1317.14 (min 1282.36, max 1364.84) |
| bw_mem-frd-1mb (mb/s) | 1170.35 (min 1028.10, max 1305.24) |
| bw_mem-frd-2mb (mb/s) | 1167.91 (min 987.33, max 1373.86) |
| bw_mem-frd-4mb (mb/s) | 1161.96 (min 982.08, max 1330.23) |
| bw_mem-frd-8mb (mb/s) | 1175.62 (min 971.46, max 1385.76) |
| bw_mem-fwr-16mb (mb/s) | 2122.36 (min 2106.37, max 2131.91) |
| bw_mem-fwr-1mb (mb/s) | 1933.56 (min 1216.12, max 2718.45) |
| bw_mem-fwr-2mb (mb/s) | 1822.22 (min 1261.83, max 2363.14) |
| bw_mem-fwr-4mb (mb/s) | 1754.10 (min 1265.82, max 2214.43) |
| bw_mem-fwr-8mb (mb/s) | 1736.68 (min 1236.09, max 2157.79) |
| bw_mem-rd-16mb (mb/s) | 1366.15 (min 1338.24, max 1382.65) |
| bw_mem-rd-1mb (mb/s) | 1100.78 (min 793.27, max 1403.09) |
| bw_mem-rd-2mb (mb/s) | 1101.55 (min 801.18, max 1400.81) |
| bw_mem-rd-4mb (mb/s) | 1105.82 (min 824.91, max 1399.09) |
| bw_mem-rd-8mb (mb/s) | 1125.63 (min 861.05, max 1393.73) |
| bw_mem-rdwr-16mb (mb/s) | 856.58 (min 844.73, max 870.65) |
| bw_mem-rdwr-1mb (mb/s) | 693.30 (min 543.97, max 838.64) |
| bw_mem-rdwr-2mb (mb/s) | 671.55 (min 524.11, max 834.84) |
| bw_mem-rdwr-4mb (mb/s) | 684.01 (min 523.49, max 844.77) |
| bw_mem-rdwr-8mb (mb/s) | 701.39 (min 523.01, max 857.08) |
| bw_mem-wr-16mb (mb/s) | 893.90 (min 886.48, max 899.79) |
| bw_mem-wr-1mb (mb/s) | 819.90 (min 754.86, max 847.17) |
| bw_mem-wr-2mb (mb/s) | 814.11 (min 777.91, max 836.35) |
| bw_mem-wr-4mb (mb/s) | 828.96 (min 793.41, max 876.33) |
| bw_mem-wr-8mb (mb/s) | 864.35 (min 836.65, max 897.06) |
| bw_mmap_rd-mo-1mb (mb/s) | 1320.67 (min 1272.50, max 1351.35) |
| bw_mmap_rd-o2c-1mb (mb/s) | 474.18 (min 454.27, max 483.64) |
| bw_pipe (mb/s) | 540.02 (min 518.86, max 558.01) |
| bw_unix (mb/s) | 541.45 (min 530.54, max 556.53) |
| lat_connect (us) | 75.76 (min 74.84, max 78.74) |
| lat_ctx-2-128k (us) | 10.34 (min 8.76, max 11.41) |
| lat_ctx-2-256k (us) | 22.65 (min 15.21, max 28.60) |
| lat_ctx-4-128k (us) | 11.46 (min 10.21, max 12.85) |
| lat_ctx-4-256k (us) | 10.78 (min 0.00, max 18.44) |
| lat_fs-0k (num_files) | 198.38 (min 194.00, max 203.00) |
| lat_fs-10k (num_files) | 88.38 (min 82.00, max 103.00) |
| lat_fs-1k (num_files) | 132.88 (min 129.00, max 138.00) |
| lat_fs-4k (num_files) | 118.50 (min 107.00, max 126.00) |
| lat_mem_rd-stride128-sz1000k (ns) | 47.38 (min 46.80, max 48.23) |
| lat_mem_rd-stride128-sz125k (ns) | 7.85 (min 7.80, max 7.97) |
| lat_mem_rd-stride128-sz250k (ns) | 12.09 (min 10.07, max 14.63) |
| lat_mem_rd-stride128-sz31k (ns) | 5.25 (min 3.07, max 5.90) |
| lat_mem_rd-stride128-sz50 (ns) | 3.02 |
| lat_mem_rd-stride128-sz500k (ns) | 43.10 (min 41.58, max 44.38) |
| lat_mem_rd-stride128-sz62k (ns) | 7.15 (min 6.43, max 7.42) |
| lat_mmap-1m (us) | 69.00 (min 65.00, max 76.00) |
| lat_ops-double-add (ns) | 4.02 |
| lat_ops-double-div (ns) | 22.13 (min 22.12, max 22.14) |
| lat_ops-double-mul (ns) | 4.02 (min 4.02, max 4.03) |
| lat_ops-float-add (ns) | 4.02 |
| lat_ops-float-div (ns) | 13.07 |
| lat_ops-float-mul (ns) | 4.02 |
| lat_ops-int-add (ns) | 1.01 |
| lat_ops-int-bit (ns) | 0.67 |
| lat_ops-int-div (ns) | 6.03 (min 6.03, max 6.04) |
| lat_ops-int-mod (ns) | 6.37 (min 6.37, max 6.39) |
| lat_ops-int-mul (ns) | 4.34 (min 4.32, max 4.42) |
| lat_ops-int64-add (ns) | 1.01 |
| lat_ops-int64-bit (ns) | 0.67 |
| lat_ops-int64-div (ns) | 9.56 (min 9.55, max 9.57) |
| lat_ops-int64-mod (ns) | 7.38 (min 7.37, max 7.39) |
| lat_ops-int64-mul (ns) | 5.01 (min 4.99, max 5.13) |
| lat_pagefault (us) | 1.81 (min 1.76, max 1.84) |
| lat_pipe (us) | 26.34 (min 25.73, max 27.06) |
| lat_proc-exec (us) | 1176.48 (min 1144.00, max 1221.00) |
| lat_proc-fork (us) | 1019.08 (min 1002.83, max 1054.80) |
| lat_proc-proccall (us) | 0.01 |
| lat_select (us) | 46.72 (min 46.13, max 47.43) |
| lat_sem (us) | 3.45 (min 3.16, max 3.71) |
| lat_sig-catch (us) | 5.78 (min 5.65, max 5.93) |
| lat_sig-install (us) | 0.94 (min 0.93, max 1.00) |
| lat_sig-prot (us) | 0.75 (min 0.69, max 0.81) |
| lat_syscall-fstat (us) | 2.55 (min 2.49, max 2.67) |
| lat_syscall-null (us) | 0.62 (min 0.60, max 0.68) |
| lat_syscall-open (us) | 386.20 (min 345.40, max 419.63) |
| lat_syscall-read (us) | 0.83 (min 0.82, max 0.88) |
| lat_syscall-stat (us) | 6.88 (min 6.48, max 7.35) |
| lat_syscall-write (us) | 0.79 (min 0.77, max 0.83) |
| lat_tcp (us) | 1.12 (min 1.10, max 1.16) |
| lat_unix (us) | 42.81 (min 37.94, max 47.17) |
| latency_for_0.50_mb_block_size (nanosec) | 43.10 (min 41.58, max 44.38) |
| latency_for_1.00_mb_block_size (nanosec) | 23.69 (min 0.00, max 48.23) |
| pipe_bandwidth (mbs) | 540.02 (min 518.86, max 558.01) |
| pipe_latency (microsec) | 26.34 (min 25.73, max 27.06) |
| procedure_call (microsec) | 0.01 |
| select_on_200_tcp_fds (microsec) | 46.72 (min 46.13, max 47.43) |
| semaphore_latency (microsec) | 3.45 (min 3.16, max 3.71) |
| signal_handler_latency (microsec) | 0.94 (min 0.93, max 1.00) |
| signal_handler_overhead (microsec) | 5.78 (min 5.65, max 5.93) |
| tcp_ip_connection_cost_to_localhost (microsec) | 75.76 (min 74.84, max 78.74) |
| tcp_latency_using_localhost (microsec) | 1.12 (min 1.10, max 1.16) |
Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. For standard ARM cores the DMIPS/MHz score will be identical with the same compiler and flags.
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| cpu_clock (mhz) | 1000.00 |
| dhrystone_per_mhz (dmips/mhz) | 2.90 |
| dhrystone_per_second (dhrystonep) | 5128205.00 |
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| whetstone (mips) | 4583.33 (min 3333.30, max 5000.00) |
Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system.
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| linpack (kflops) | 407927.00 (min 406745.00, max 408901.00) |
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.
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| cjpeg-rose7-preset (workloads/) | 29.58 (min 29.41, max 29.67) |
| core (workloads/) | 0.21 |
| coremark-pro () | 588.47 (min 585.75, max 593.10) |
| linear_alg-mid-100x100-sp (workloads/) | 10.41 (min 10.41, max 10.42) |
| loops-all-mid-10k-sp (workloads/) | 0.48 |
| nnet_test (workloads/) | 0.77 (min 0.76, max 0.77) |
| parser-125k (workloads/) | 5.46 (min 5.21, max 5.62) |
| radix2-big-64k (workloads/) | 20.60 (min 19.46, max 21.53) |
| sha-test (workloads/) | 57.72 (min 57.47, max 57.80) |
| zip-test (workloads/) | 15.47 (min 15.38, max 15.63) |
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| cjpeg-rose7-preset (workloads/) | 55.33 (min 38.46, max 59.17) |
| core (workloads/) | 0.43 (min 0.42, max 0.43) |
| coremark-pro () | 1041.66 (min 1003.86, max 1052.62) |
| linear_alg-mid-100x100-sp (workloads/) | 20.81 (min 20.78, max 20.84) |
| loops-all-mid-10k-sp (workloads/) | 0.88 (min 0.87, max 0.89) |
| nnet_test (workloads/) | 1.54 |
| parser-125k (workloads/) | 5.86 (min 5.56, max 6.15) |
| radix2-big-64k (workloads/) | 32.59 (min 31.39, max 33.66) |
| sha-test (workloads/) | 115.13 (min 114.94, max 116.28) |
| zip-test (workloads/) | 28.17 (min 27.78, max 28.57) |
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.
| Benchmarks | am64xx-hsevm: perf |
|---|---|
| 4m-check (workloads/) | 282.33 (min 280.68, max 284.84) |
| 4m-check-reassembly (workloads/) | 61.37 (min 61.12, max 61.50) |
| 4m-check-reassembly-tcp (workloads/) | 36.64 (min 36.39, max 36.93) |
| 4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/) | 14.60 (min 14.57, max 14.63) |
| 4m-check-reassembly-tcp-x264w2 (workloads/) | 0.75 (min 0.74, max 0.75) |
| 4m-cmykw2 (workloads/) | 85.93 (min 85.80, max 86.17) |
| 4m-cmykw2-rotatew2 (workloads/) | 17.54 (min 16.62, max 18.35) |
| 4m-reassembly (workloads/) | 56.76 (min 56.34, max 57.24) |
| 4m-rotatew2 (workloads/) | 20.14 (min 18.86, max 21.70) |
| 4m-tcp-mixed (workloads/) | 84.66 (min 83.77, max 85.56) |
| 4m-x264w2 (workloads/) | 0.77 (min 0.76, max 0.77) |
| empty-wld (workloads/) | 1.00 |
| idct-4m (workloads/) | 13.67 (min 13.65, max 13.68) |
| idct-4mw1 (workloads/) | 13.67 (min 13.65, max 13.68) |
| ippktcheck-4m (workloads/) | 282.00 (min 280.96, max 284.51) |
| ippktcheck-4mw1 (workloads/) | 282.27 (min 281.37, max 284.61) |
| ipres-4m (workloads/) | 72.04 (min 71.70, max 72.39) |
| ipres-4mw1 (workloads/) | 72.29 (min 72.05, max 72.50) |
| md5-4m (workloads/) | 19.98 (min 19.88, max 20.02) |
| md5-4mw1 (workloads/) | 20.01 (min 19.94, max 20.06) |
| rgbcmyk-4m (workloads/) | 44.75 (min 44.72, max 44.77) |
| rgbcmyk-4mw1 (workloads/) | 44.72 (min 44.69, max 44.77) |
| rotate-4ms1 (workloads/) | 16.62 (min 16.49, max 16.78) |
| rotate-4ms1w1 (workloads/) | 16.57 (min 16.30, max 16.74) |
| rotate-4ms64 (workloads/) | 16.83 (min 16.77, max 16.95) |
| rotate-4ms64w1 (workloads/) | 16.82 (min 16.77, max 16.92) |
| x264-4mq (workloads/) | 0.41 |
| x264-4mqw1 (workloads/) | 0.41 |
| Boot Configuration | am64xx-hsevm: Boot time in seconds: avg(min,max) |
|---|---|
| Linux boot time from SD with default rootfs (20 boot cycles) | 22.36 (min 20.60, max 27.42) |
Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles.
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
- CPSW2g: AM65x, J7200, J721e
- CPSW3g: AM64x
TCP Bidirectional Throughput
| Command Used | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: 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 | 1057.82 (min 989.78, max 1130.34) | 97.00 (min 78.90, max 99.90) |
| UDP Datagram Size(bytes) (LOCAL_SEND_SIZE) | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: Packets Per Second (kPPS) | am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL) |
|---|---|---|---|
| 64 | 39.82 | 77.00 | 89.91 |
| 128 | 75.5 | 74.00 | 89.57 |
| 256 | 148.51 | 73.00 | 88.11 |
| 1024 | 575.58 | 70.00 | 91.17 |
| 1472 | 583.88 | 48.00 | 84.69 |
| UDP Datagram Size(bytes) (LOCAL_SEND_SIZE) | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: Packets Per Second (kPPS) | am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL) |
|---|---|---|---|
| 64 | 2.41 | 5.00 | 0.14 |
| 128 | 4.81 | 5.00 | 0.55 |
| 256 | 10.85 | 5.00 | 0.3 |
| 1024 | 43.42 | 5.00 | 0.15 |
| 1472 | 62.41 | 5.00 | 3.28 |
TCP Bidirectional Throughput
| Command Used | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL) |
|---|---|---|
| netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS | 1007.16 | 99.64 |
TCP Bidirectional Throughput Interrupt Pacing
| Command Used | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL) |
|---|---|---|
| netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS | 1108.21 | 98.21 |
UDP Egress Throughput
| Frame Size(bytes) | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: Packets Per Second (kPPS) | am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL) |
|---|---|---|---|
| 64 | 30.77 | 60.13 | 73.74 |
| 128 | 66.04 | 64.67 | 79.53 |
| 256 | 126.02 | 61.50 | 77.46 |
| 1024 | 495.53 | 60.63 | 77.25 |
| 1472 | 789.13 | 67 | 89.64 |
UDP Ingress Throughput
| Frame Size(bytes) | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: Packets Per Second (kPPS) | am64xx-hsevm: CPU Load % |
|---|---|---|---|
| 64 | 2.15 | 4.14 | 1.49 |
| 128 | 4.96 | 4.78 | 3.72 |
| 256 | 10.21 | 5.00 | 2.97 |
| 1024 | 44.54 | 5.25 | 6.07 |
| 1472 | 631.74 | 54 | 68.59 |
Switch Mode
| Mode | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: CPU Load % (SENDER) | am64xx-hsevm: CPU Load % (FORWARDING) | am64xx-hsevm: CPU Load % (RECIEVER) |
|---|---|---|---|---|
| Switch with HW Offload | 938 | 0.3 | 0.12 | 37.74 |
HSR Mode
| Mode | am64xx-hsevm: THROUGHPUT (Mbits/sec) | am64xx-hsevm: CPU Load % (SENDER) | am64xx-hsevm: CPU Load % (FORWARDING) | am64xx-hsevm: CPU Load % (RECIEVER) |
|---|---|---|---|---|
| HSR with HW Offload | 412 | 47.45 | 0.13 | 44.52 |
| HSR with SW Offload | 387 | 39.58 | 27.13 | 48.74 |
AM64XX-EVM
RAW
| File size (Mbytes) | am64xx-hsevm: Raw Read Throughput (Mbytes/sec) |
|---|---|
| 50 | 139.89 (min 113.64, max 147.06) |
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.
| Buffer size (bytes) | am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Write EXT4 CPU Load (%) | am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Read EXT4 CPU Load (%) |
|---|---|---|---|---|
| 1m | 60.75 (min 59.60, max 61.40) | 4.00 (min 3.73, max 4.22) | 175.00 | 7.20 (min 6.77, max 7.82) |
| 4m | 60.83 (min 59.80, max 61.70) | 2.88 (min 2.78, max 2.95) | 174.75 (min 174.00, max 175.00) | 5.76 (min 5.30, max 6.12) |
| 4k | 49.86 (min 49.10, max 50.40) | 52.21 (min 51.71, max 52.53) | 56.18 (min 56.00, max 56.40) | 47.11 (min 46.00, max 49.21) |
| 256k | 60.75 (min 59.70, max 61.50) | 5.94 (min 5.65, max 6.25) | 174.00 | 8.90 (min 8.58, max 9.48) |
| Buffer size (bytes) | am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Write EXT4 CPU Load (%) | am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Read EXT4 CPU Load (%) |
|---|---|---|---|---|
| 102400 | 53.87 (min 49.65, max 55.64) | 11.09 (min 9.38, max 15.25) | 176.52 (min 171.28, max 177.31) | 29.36 (min 28.30, max 29.63) |
| 262144 | 53.52 (min 49.20, max 55.06) | 11.06 (min 9.50, max 15.22) | 181.79 (min 181.43, max 182.11) | 30.36 (min 29.36, max 31.53) |
| 524288 | 53.64 (min 49.72, max 55.58) | 10.94 (min 9.33, max 14.88) | 182.88 (min 182.70, max 183.02) | 27.74 (min 26.13, max 29.20) |
| 1048576 | 53.83 (min 49.26, max 55.68) | 10.72 (min 9.28, max 14.60) | 182.91 (min 182.86, max 183.00) | 26.93 (min 25.45, max 28.57) |
| 5242880 | 53.93 (min 49.61, max 56.34) | 10.61 (min 9.26, max 14.60) | 182.80 (min 182.62, max 182.95) | 27.64 (min 26.61, max 28.57) |
| Buffer size (bytes) | am64xx-hsevm: Write VFAT Throughput (Mbytes/sec) | am64xx-hsevm: Write VFAT CPU Load (%) | am64xx-hsevm: Read VFAT Throughput (Mbytes/sec) | am64xx-hsevm: Read VFAT CPU Load (%) |
|---|---|---|---|---|
| 102400 | 49.34 (min 40.10, max 53.27) | 13.58 (min 11.41, max 20.45) | 166.32 (min 164.99, max 167.10) | 30.88 (min 29.20, max 31.90) |
| 262144 | 50.55 (min 41.76, max 53.49) | 13.71 (min 11.72, max 20.37) | 168.61 (min 161.34, max 171.00) | 35.58 (min 29.57, max 50.82) |
| 524288 | 50.67 (min 41.94, max 53.88) | 14.01 (min 11.56, max 20.24) | 169.98 (min 169.68, max 170.20) | 27.60 (min 27.50, max 28.10) |
| 1048576 | 50.86 (min 41.78, max 54.00) | 13.53 (min 11.59, max 20.00) | 169.78 (min 169.62, max 169.88) | 27.34 (min 26.05, max 28.10) |
| 5242880 | 50.75 (min 41.22, max 54.11) | 13.44 (min 11.66, max 19.75) | 170.06 (min 169.90, max 170.19) | 27.70 (min 27.27, max 28.33) |
| File size (bytes in hex) | am64xx-hsevm: Write Throughput (Kbytes/sec) | am64xx-hsevm: Read Throughput (Kbytes/sec) |
|---|---|---|
| 2000000 | 59674.24 (min 56594.13, max 61134.33) | 168800.07 (min 168041.03, max 169782.38) |
| 4000000 | 60491.03 (min 57387.04, max 61768.14) | 172690.99 (min 172463.16, max 173375.66) |
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.
| Buffer size (bytes) | am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Write EXT4 CPU Load (%) | am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec) | am64xx-hsevm: Read EXT4 CPU Load (%) |
|---|---|---|---|---|
| 1m | 42.00 (min 41.40, max 42.50) | 3.43 (min 3.33, max 3.67) | 86.94 (min 86.70, max 87.30) | 5.02 (min 4.81, max 5.29) |
| 4m | 41.90 (min 41.20, max 42.70) | 2.59 (min 2.51, max 2.67) | 86.33 (min 82.30, max 87.10) | 3.73 (min 3.38, max 3.99) |
| 4k | 2.76 (min 2.70, max 2.87) | 7.29 (min 7.00, max 8.06) | 12.76 (min 12.70, max 12.80) | 14.14 (min 13.37, max 14.99) |
| 256k | 37.41 (min 36.60, max 38.70) | 4.53 (min 4.26, max 4.84) | 83.15 (min 82.80, max 83.80) | 6.69 (min 6.42, max 7.06) |
| Buffer size (bytes) | am64xx-hsevm: Write Raw Throughput (Mbytes/sec) | am64xx-hsevm: Write Raw CPU Load (%) | am64xx-hsevm: Read Raw Throughput (Mbytes/sec) | am64xx-hsevm: Read Raw CPU Load (%) |
|---|---|---|---|---|
| 102400 | 29.65 (min 27.48, max 32.92) | 6.38 (min 5.22, max 9.52) | 38.94 (min 37.34, max 41.04) | 7.41 (min 7.07, max 7.77) |
| 262144 | 28.96 (min 27.17, max 31.59) | 6.52 (min 5.28, max 9.31) | 40.55 (min 39.38, max 41.90) | 6.57 (min 6.20, max 7.11) |
| 524288 | 29.58 (min 27.10, max 32.84) | 6.12 (min 4.95, max 9.82) | 45.12 (min 42.60, max 45.56) | 6.75 (min 6.36, max 6.99) |
| 1048576 | 29.85 (min 28.58, max 32.27) | 6.13 (min 4.97, max 9.06) | 45.39 (min 45.16, max 45.52) | 6.71 (min 6.36, max 7.17) |
| 5242880 | 29.78 (min 28.51, max 33.57) | 5.85 (min 4.81, max 8.70) | 45.22 (min 44.88, max 45.51) | 6.85 (min 6.51, max 7.33) |
The performance numbers were captured using the following:
- SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA)
- Partition was mounted with async option
| Algorithm | Buffer Size (in bytes) | am64xx-hsevm: throughput (KBytes/Sec) |
|---|---|---|
| aes-128-cbc | 1024 | 21201.11 (min 19586.05, max 21629.27) |
| aes-128-cbc | 16 | 305.74 (min 295.07, max 317.97) |
| aes-128-cbc | 16384 | 137300.65 (min 134600.02, max 139203.93) |
| aes-128-cbc | 256 | 5427.83 (min 5317.12, max 5639.77) |
| aes-128-cbc | 64 | 1333.09 (min 1290.43, max 1389.82) |
| aes-128-cbc | 8192 | 98276.35 (min 96264.19, max 99693.91) |
| aes-128-ecb | 1024 | 21623.04 (min 20551.00, max 22215.00) |
| aes-128-ecb | 16 | 312.77 (min 297.66, max 329.37) |
| aes-128-ecb | 16384 | 141950.29 (min 140525.57, max 143638.53) |
| aes-128-ecb | 256 | 5490.12 (min 5154.73, max 5683.37) |
| aes-128-ecb | 64 | 1360.48 (min 1284.46, max 1411.31) |
| aes-128-ecb | 8192 | 101527.21 (min 99691.18, max 102948.86) |
| aes-192-cbc | 1024 | 21028.78 (min 19525.97, max 21523.11) |
| aes-192-cbc | 16 | 311.25 (min 294.13, max 320.98) |
| aes-192-cbc | 16384 | 128866.99 (min 127462.06, max 130378.41) |
| aes-192-cbc | 256 | 5524.78 (min 5259.43, max 5622.10) |
| aes-192-cbc | 64 | 1364.05 (min 1286.27, max 1390.40) |
| aes-192-cbc | 8192 | 94148.27 (min 92752.55, max 95668.91) |
| aes-192-ecb | 1024 | 21608.83 (min 20942.51, max 22065.83) |
| aes-192-ecb | 16 | 314.19 (min 301.85, max 325.51) |
| aes-192-ecb | 16384 | 132728.15 (min 132104.19, max 133327.53) |
| aes-192-ecb | 256 | 5513.00 (min 5258.24, max 5673.05) |
| aes-192-ecb | 64 | 1369.78 (min 1305.26, max 1411.67) |
| aes-192-ecb | 8192 | 97255.42 (min 96520.87, max 98402.30) |
| aes-256-cbc | 1024 | 21003.86 (min 18761.39, max 21494.78) |
| aes-256-cbc | 16 | 314.16 (min 297.62, max 322.65) |
| aes-256-cbc | 16384 | 121110.53 (min 120105.64, max 122344.79) |
| aes-256-cbc | 256 | 5537.48 (min 5443.41, max 5665.71) |
| aes-256-cbc | 64 | 1377.82 (min 1358.29, max 1398.68) |
| aes-256-cbc | 8192 | 90706.26 (min 89451.18, max 91501.91) |
| aes-256-ecb | 1024 | 21289.98 (min 20055.38, max 21793.11) |
| aes-256-ecb | 16 | 308.54 (min 300.03, max 320.75) |
| aes-256-ecb | 16384 | 125571.07 (min 124474.71, max 127385.60) |
| aes-256-ecb | 256 | 5471.54 (min 5277.53, max 5680.98) |
| aes-256-ecb | 64 | 1351.16 (min 1297.41, max 1409.83) |
| aes-256-ecb | 8192 | 93359.79 (min 92383.91, max 94601.22) |
| sha256 | 1024 | 28245.16 (min 27951.45, max 28524.20) |
| sha256 | 16 | 471.00 (min 466.66, max 475.25) |
| sha256 | 16384 | 218273.11 (min 216738.47, max 220250.11) |
| sha256 | 256 | 7394.91 (min 7313.49, max 7465.56) |
| sha256 | 64 | 1871.55 (min 1853.21, max 1889.11) |
| sha256 | 8192 | 149173.25 (min 148398.08, max 149848.06) |
| sha512 | 1024 | 18903.94 (min 18712.58, max 19061.76) |
| sha512 | 16 | 449.39 (min 444.90, max 451.80) |
| sha512 | 16384 | 48433.15 (min 48059.73, max 48562.18) |
| sha512 | 256 | 6372.07 (min 6279.94, max 6436.78) |
| sha512 | 64 | 1803.87 (min 1779.22, max 1823.55) |
| sha512 | 8192 | 43742.21 (min 43578.71, max 43835.39) |
| Algorithm | am64xx-hsevm: CPU Load |
|---|---|
| aes-128-cbc | 43.25 (min 42.00, max 44.00) |
| aes-128-ecb | 45.13 (min 43.00, max 46.00) |
| aes-192-cbc | 43.75 (min 43.00, max 45.00) |
| aes-192-ecb | 44.63 (min 44.00, max 45.00) |
| aes-256-cbc | 43.63 (min 43.00, max 44.00) |
| aes-256-ecb | 43.88 (min 43.00, max 45.00) |
| sha256 | 93.00 |
| sha512 | 92.50 (min 90.00, max 93.00) |
Listed for each algorithm are the code snippets used to run each benchmark test.
time -v openssl speed -elapsed -evp aes-128-cbc
RP Message latency is the delay measured from sending a round trip echo message from a Linux application to a remote processor and back. The following measurements use a RP message length of 1 byte and 490 bytes for comparison.
The Linux user space application rpmsg_char_benchmark captures these latency values.
Test commands used for running IPC latency tests:
rpmsg_char_benchmark -r 2 -n 100000 -m 1 & chrt -f -p 80 $!Latencies reported:
| Remote Processor | Message Size (in bytes) | Average round trip (usecs) | Max round trip (usecs) |
|---|---|---|---|
| R5F0_0 | 1 | 34 | 165 |
| R5F0_0 | 490 | 155 | 279 |
| M4F | 1 | 45 | 138 |
| M4F | 490 | 285 | 408 |
