Jump to content

Armbian on MiQi SBC hardware ?


Peba

Recommended Posts

 

Yes, you are correct. It was on powersave, those are now on proper speed:  :)

type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
aes-128 cbc      79215.37k    86050.52k    89357.40k    90211.33k    90456.06k
bf-cbc           50721.71k    57495.64k    59830.53k    60454.91k    60631.72k

 

I've just tested on my MiQi and got about 2% higher performance at 1.8 GHz :

 

    type                  16 bytes        64 bytes       256 bytes     1024 bytes    8192 bytes

    aes-128 cbc      80197.72k    87590.12k    91353.34k    92255.91k    92520.45k

    bf-cbc                52186.86k    59762.60k    62007.89k    62678.36k    62876.33k

 

At 2.0 GHz I get this :

 

    type                  16 bytes        64 bytes      256 bytes      1024 bytes    8192 bytes

    aes-128 cbc      88145.11k    97268.91k   101094.74k   102091.43k  102386.35k

    bf-cbc                57751.12k    66213.18k    68619.26k     69361.66k     69580.12k

 

My openssl is build in thumb mode with the following options under linaro gcc-4.7.4 :

 

   armv7thf-gcc47l_glibc218-linux-gnueabi-gcc -I. -I.. -I../include  -fPIC -DOPENSSL_PIC -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -mcpu=cortex-a9 -mlittle-endian -O3 -fomit-frame-pointer -march=armv7-a -D__ARM_MAX_ARCH__=7 -O3 -Wall -DOPENSSL_BN_ASM_MONT -DOPENSSL_BN_ASM_GF2m -DSHA1_ASM -DSHA256_ASM -DSHA512_ASM -DAES_ASM -DBSAES_ASM -DGHASH_ASM

 

Given that most AES code is written in asm, this 2% performance gain likely comes from the rest of the code and could be caused by the mcpu=cortex-a9 build options, I'm forcing it everywhere as I found it to perform better on the RK3288 than mcpu=cortex-a17 on most programs. I suspect that the A17 core in the RK3288 is more of an A12 (as reported in cpuinfo) hence closer to A9 than what the A17 was supposed to be before ARM decided that both were the same.

Link to comment
Share on other sites

For comparison:

 

1.99 GHz

type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
aes-128 cbc      87295.72k    94739.03k    98363.39k    99325.95k    99562.84k
bf-cbc           55820.25k    63628.57k    65862.06k    66524.84k    66680.15k

2.14 Ghz (maximum possible value for my board)

aes-128 cbc      93640.97k   101616.77k   105520.55k   106535.94k   106785.45k
bf-cbc           59887.06k    68163.84k    70644.39k    71375.19k    71538.01k

 

20170221_123035.jpg

Link to comment
Share on other sites

44 minutes ago, Igor said:

2.14 Ghz (maximum possible value for my board)

 

Did you still power the board through Micro USB or fan header? OpenSSL Distro packages (Xenial?) or self-compiled with Willy's settings?

Link to comment
Share on other sites

8 hours ago, tkaiser said:

Did you still power the board through Micro USB or fan header? OpenSSL Distro packages (Xenial?) or self-compiled with Willy's settings?


Xenial packages and by best micro USB cable.  Not long after tests I experienced instability and power cycling due to excessive heat. Board was hot like hell and when it was cooled down, everything was back to normal.

 

Heat dissipation that I used was clearly not enough when overclocking. 

Link to comment
Share on other sites

6 hours ago, jernej said:

Using latest u-boot-rockchip repo I managed to add MiQi support to U-Boot. Not everything is properly configured yet, but it should be able to boot kernel. @Igor now that SPL is working again, maybe we should go away from that binary blobs?


Sure, why not.

 

I added a patch but had troubles booting ... have to check later what am I doing wrong.

Link to comment
Share on other sites

With both methods you ended up with bricked MiQi ... then use tweezers to short resistor, force maskrom mode and reflash stock u-boot. :rolleyes: Something is wrong.


Edit:

It stuck at displaying: U-Boot SPL 2017.03

Edited by Igor
stuck point
Link to comment
Share on other sites

Hm, I never had to short resistor. First I invalidated eMMC bootloader using built in U-Boot command, so board always boots to special USB mode. Then I just played with flashing U-Boot with SPL to SD card...

 

Which repo are you using for a base? I will provide compiled version later this day.

Link to comment
Share on other sites

Well I flashed to eMMC :)
 

4 minutes ago, jernej said:

Which repo are you using for a base? I will provide compiled version later this day.


Denx 2017.03 + your patch.

Link to comment
Share on other sites

I couldn't make it work with 2017.03-rc3 and according to log, 2017.03 release doesn't change much. Try to use rockchip next repo or alternatively, use miqi-new branch from my u-boot repo. I'm also not sure if eMMC works properly. I have to recheck DT. As I said, early success :)

Link to comment
Share on other sites

Ok, now it should be fully supported (eth, usb, video, emmc, etc.)

https://github.com/jernejsk/u-boot/tree/miqi

 

Binary (14 days available):

https://transfer.sh/fVGr4/u-boot-rockchip-miqi-sd.bin

 

Unfortunately, many patches are not in any stable release, so for testing purposes it would be best just to use my branch and miqi-rk3288_defconfig

Link to comment
Share on other sites

On 3/21/2017 at 5:40 AM, Igor said:

With both methods you ended up with bricked MiQi ... then use tweezers to short resistor, force maskrom mode and reflash stock u-boot. :rolleyes: Something is wrong.


Edit:

It stuck at displaying: U-Boot SPL 2017.03

 

Just a thought. the boot loader offset locations *should* be hardware bound by the SoC, there shouldn't be any difference between MiQi and Tinker Board unless the MiQi is using an SPI device to load a bootloader, but that does not sound like the case.  With the cat command throwing it's data away instead of adding it to the output binary, the Tinker Board did exactly as you describe here, only showed the U-Boot SPL message.

Link to comment
Share on other sites

As @tkaiser mentioned elsewhere, there's currently a great bargain on Chiptrip Q8 at GB. 

 

Libreelec already runs on Ugoos UT3+, and with Linux support bound to improve in the coming months, this looks very attractive.  Great minds think alike, eh @mdel? ;) 

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines