Discussion

Banana Pi & Pro SATA and USB Hard Drive Tests

11 10567
tkaiser  
Edited by tkaiser at Wed Mar 18, 2015 03:34
mikronauts replied at Tue Mar 17, 2015 08:11
We can agree to disagree


Seems so

One final note: In my opinion it's useless to be after "out of the box" experiences. Please have a look at this benchmark from last June:

http://hardware-libre.fr/2014/06/raspberry-vs-banana-hardware-duel/

This is a surprise : in a single-threaded test, the Allwinner A20 chip is slower than the Broadcom BCM2835, even with theorically 1GHz, compared to the Raspberry’s 700MHz


(test 4 "SD card" is also an excellent example how people misuse hdparm/dd and publish totally misleading results)

Back in these days the LeMaker guys just ported Raspbian to the A20, replaced RPi's kernel will Allwinner's Android kernel and didn't adjust relevant stuff unknown on the RPi. For example the CPU governor. It was "fantasy" back then which is the wrong choice if you want the board to perform good and keep performance stable. This and only this explains the bad "out of the box" performance people measured back in these days (you called it "shipped from the factory"). The whole stuff is explained here in detail: http://forum.lemaker.org/forum.php?mod=redirect&goto=findpost&ptid=153&pid=10212

This was "out of the box" or "shipped from the factory" a year ago and people believed the hardware had to be blamed for the benchmark results people published on the net. Then serious benchmark testers discovered the things that are relevant for fast performance on such a slow platform like these ARM SBCs: More importantly appropriate settings. And the stuff you called "NAS tuning" is just identifying relevant parameters, documenting them and hoping that they will be adopted by the "factory staff" in the future (like they did with the CPU governor and many other tunables and maybe soon also regarding kernel config, eg. CONFIG_SCHED_MC=y, CONFIG_SCHED_SMT=y that will help with both network throughput and the stuff you focused on in your review: 2-3 MB/s and almost 10 MB/s more SATA write/read throughput)

I know, it was very evident that you found "out of the box" comparisons useless.

While I still find them useful, you have gotten me to the point of seriously considering writing "tuned" follow up articles (something I've considered in the past, but was not sure if I should given the "noob" nature of the Pi audience)

You are also correct, I do hope that "factory staff" learn from my articles - and yours - which will also hopefully get end users thinking.

Regarding SD cards - I recently found out that the Pi's Broadcom SOC is actually capable of UHS-1 but the pcb does not support the required 1v8 regulator and passives... now that is a crying shame.

You have to log in before you can reply Login | Sign Up

Points Rules