MISC-TN-010: Using NXP eIQ Machine Learning Development Environment with Mito8M SoM

Info Box Applies to MITO 8M Applies to Machine Learning

This technical note was validated against specific versions of hardware and software. What is described here may not work with other versions.

History

Version	Date	Notes
1.0.0	January 2020	First public release

Introduction

According to NXP documentation, eIQ Machine Learning Software is a collection of software and development tools for NXP microprocessors and microcontrollers to do inference of neural network AI models on embedded systems.

This Technical Note (TN for short) illustrates hot to use eIQ in combination with Mito8M, DAVE Embedded Systems's latest SoM, which is built upon the i.MX8M processor by NXP.

Testbed

With regards to the hardware, the testbed consists of the same platform described MISC-TN-008:_Running_Debian_Buster_(armbian)_on_Mito8M here.

Concerning the software, the following combination was used:

• U-Boot 2018.03 retrieved from the standard Mito8M Yocto-based Board Support Package (BSP)
• Device tree retrieved from the standard Mito8M Yocto-based BSP
• Linux kernel imx8qmmek 4.14.98-imx_4.14.98_2.0.0
• eIQ-enabled Yocot-based root file system.

For more information about the kernel and the root file system, please refer to the following section.

Building NXP eIQ software

NXP document UM11226 describes how to build

This processor is capable of running either at 800 MHz or 1.3 GHz. The tests were performed at either frequencies in order to determine how the it affects the RAM bandwidth.

The following table details the characteristics of the SDRAM bank connected to the SoC.

Subsystem	Feature	Platform
		Mito8M
SoC	SoC	NXP i.MX8M Quad
	ARM core(s)	4 x Cortex A53
	ARM core frequency [MHz]	800 or 1300
	L1 cache (D) [kB]	32
	L1 cache (I) [kB]	32
	L2 cache [MB]	1
SDRAM	Type	LPDDR4
	Frequency [MHz]	1600
	Bus witdth [bit]	32
	Theoretical bandwidth [Gb/s]	102.4
	Theoretical bandwidth [GB/s]	12.8
	Size [MB]	3072

Software configuration

• Linux kernel: 4.14.98
• Root file system: Debian GNU/Linux 10 (buster)
• Architecture: aarch64
• Governor: userspace @ 800 MHz or 1300 MHz

root@Mito8M:~# echo userspace > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
root@Mito8M:~# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
userspace
root@Mito8M:~# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
800000

Some benchmarks were built natively on the platform under test. For the sake of completeness, the version of the GCC compiler is then indicated as well:

armbian@Mito8M:~/devel/lmbench$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/aarch64-linux-gnu/8/lto-wrapper
Target: aarch64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 8.3.0-6' --with-bugurl=file:///usr/share/doc/gcc-8/README.Bugs --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++ --prefix=/usr --with-gcc-major-version-only --program-suffix=-8 --program-prefix=aarch64-linux-gnu- --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --enable-bootstrap --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-gnu-unique-object --disable-libquadmath --disable-libquadmath-support --enable-plugin --enable-default-pie --with-system-zlib --disable-libphobos --enable-multiarch --enable-fix-cortex-a53-843419 --disable-werror --enable-checking=release --build=aarch64-linux-gnu --host=aarch64-linux-gnu --target=aarch64-linux-gnu
Thread model: posix
gcc version 8.3.0 (Debian 8.3.0-6)

Overall results

This section illustrates the overall results achieved by the benchmarks.

STREAM

Overall results (ARM core frequency = 800 MHz)

Function	Mito8M		Axel Lite efficiency [%]
	Best rate [MB/s]	Efficiency [%]
Copy	6770	51.7	14.0
Scale	6093	46.5	13.8
Add	5263	40.1	14.6
Triad	4820	36.8	14.9

Overall results (ARM core frequency = 1300 MHz)

Function	Mito8M		Axel Lite efficiency [%]
	Best rate [MB/s]	Efficiency [%]
Copy	7125	54.3	14.0
Scale	7501	57.2	13.8
Add	6762	51.6	14.6
Triad	6354	48.5	14.9

Apart from the increase over Axel Lite in absolute terms, it is noteworthy that Mito8M exhibits a significant improvement in terms of efficiency too, as shown in the above tables. This is especially true in the case of ARM core frequency set to 1300 MHz.

Another interesting thing to note is how the bandwidth is affected by the ARM core frequency. If it scaled linearly, we should have an improvement of 62.5% from 800 to 1300 MHz. The average bandwidth at 800 MHz is 5761 MB/s. At 1300 MHz, it is 6935 MB/s. Therefore, the increase is 20.4%. With regard to STREAM benchmark, the achieved bandwidth does not scale linearly with ARM core frequency.

Please see this page for more details about STREAM benchmark.

LMbench

For what regards the memory bandwidth, LMbench provides many results organized in different categories. For the sake of simplicity, the following tables details just a couple of categories. The full results are available for download here (ARM core frequency set to 800 MHz) and here (ARM core frequency set to 1300 MHz).

Memory read bandwidth

Buffer size	Bandwitdth [MB/s]
	ARM core frequency = 800 MHz	ARM core frequency = 1300 MHz
512B	1553	2521
1kB	1567	2546
2kB	1575	2560
4kB	1575	2564
8kB	1577	2564
16kB	1577	2567
32kB	1528	2490
64kB	1531	2494
128kB	1547	2530
256kB	1552	2526
512kB	1514	2518
1MB	1318	2181
2MB	1430	2148
4MB	1420	2108
8MB	1423	2038
16MB	1420	2116
32MB	1365	2117
64MB	1393	2035
128MB	1382	2035
256MB	1372	2050
512MB	1367	1998

Memory write bandwidth

Buffer size	Bandwitdth [MB/s]
	ARM core frequency = 800 MHz	ARM core frequency = 1300 MHz
512B	2932	4771
1kB	3048	4956
2kB	3100	5046
4kB	3136	5097
8kB	3135	5101
16kB	3150	5120
32kB	2864	5127
64kB	3033	5071
128kB	3093	4886
256kB	2956	5056
512kB	3024	5054
1MB	3075	5092
2MB	3095	5116
4MB	3121	5118
8MB	3137	5120
16MB	3145	5121
32MB	3146	5120
64MB	3146	5125
128MB	3147	5123
256MB	3150	5124
512MB	3144	5125
1GB	3146	5124

There are some interesting facts to stress:

• Read and write bandwitdth are not effected by the buffer size.
• They are significantly affected by the ARM core frequency. For instance, the improvement of the write bandwidth (about 62% when the buffer is 1GB) is practically the same of the increase in frequency.

For more information regarding LMbench, please see this page.

pmbw

As defined by the author, pmbw is "a set of assembler routines to measure the parallel memory (cache and RAM) bandwidth of modern multi-core machines." It performs a myriad of tests. Luckily, it comes with a handful tool that plots the results—which are stored in a text file—in a series of charts. Again,the benchmark was run at two different ARM core frequencies, 800 and 1300 MHz.

The complete results and the charts are available at the following links:

• http://mirror.dave.eu/mito/Mito8M/pmbw-stats-Mito8M-800MHz.txt
• http://mirror.dave.eu/mito/Mito8M/pmbw-plots-Mito8M-800MHz.pdf
• http://mirror.dave.eu/mito/Mito8M/pmbw-stats-Mito8M-1300MHz.txt
• http://mirror.dave.eu/mito/Mito8M/pmbw-plots-Mito8M-1300MHz.pdf

Generally speaking, the charts exhibit significant declines in the performances when the array size is around the L1 and the L2 cache size.

For more details about pmbw, please refer to this page.

stressapptest

According to the documentation, stressapptest—which was developed at Google—is "a memory interface test. It tries to maximize randomized traffic to memory from processor and I/O, with the intent of creating a realistic high load situation in order to test the existing hardware devices in a computer."

Test	Bandwidth [MB/s]
	ARM core frequency = 800 MHz	ARM core frequency = 1300 MHz
Memory copy	5483	5804

The above table lists the achieved results when the benchmark was run as detailed in this section. In this case, the different when running at different ARM core frequencies is very little.

For more information about stressapptest, please refer to this page.

Useful links

• Joshua Wyatt Smith and Andrew Hamilton, Parallel benchmarks for ARM processors in the highenergy context
• T Wrigley, G Harmsen and B Mellado, Memory performance of ARM processors and itsrelevance to High Energy Physics
• G. T. Wrigley, R. G. Reed, B. Mellado, Memory benchmarking characterisation of ARM-based SoCs

Appendix A: Detailed testing procedures

This section details how the benchmarks were configured and run on the testbed.

STREAM

Building

To build STREAM:

• clone its git repository
• modify the Makefile as shown below
• issue the make command.