The EspressoBin: Something completely different? // Review


Phoronix Tests

So, on to some Phoronix tests. For this I used a simple copper heatsink. There was no thermal info built into this particular kernel, and as mentioned before the CPU was set to 1GHz.

I also attached a 2.5″ 250G SATA disk. It’s not the world’s fastest disk, but good enough for these tests.

And rigged up a very permanent cooling arrangement. There wasn’t enough space to use my usual Uber heatsink.This time round I ran the Phoronix tests for around two weeks. There were the occasional crashes here and there, but overall it chugged away.

I kept tabs on the temperature regularly using a cheap infrared sensor. I saw an average of around 30C and a peak of almost 36C. This was with a fairly constant ambient temperature of 20C.

  • 27.5
  • 27.9
  • 29.4
  • 30.4
  • 35.8
  • 27.6

Even taking into account for the emissivity of copper, which would add at least 5 degrees, it’s still running quite cool.

As for power I didn’t have my USB power logger with me and so resorted to taking timed photos. Current draw often hit 1A with a peak of 1.7A and the average was around 830mA. Note that this is at 12v.

And what was the end result of all these benchmark tests?

On the network side, local loopback was a little on the high side,and there was a mixed bag of results for SATA performance. For example: sqlite, apache and kernel compiles all lagged behind other boards,

but was around the middle mark within boards of the same price for the dbench, postmark, aio-stress, and redis benchmarks.

When it came to CPU processing power; cachebench, clomp, found the EspressoBin lagging behind most of the other boards.


However, RAMspeed highlighted the fact that the EspressoBin was around the same as the Pine64.

When it came to maths benchmarks; The BORK benchmark which tests encryption capabilities, surprisingly was extremely slow. So much so that I’m discounting this benchmark as there was too much variance between subsequent tests.


However, compression benchmarks showed the board either lagging behind, or sitting in the middle of the bunch.
The same happened with audio encoding, Fast Fourier Transforms and encryption benchmarks.

Running general CPU benchmarks such as dolfyn, fhourstones, timed hmmer, minion, stockfish all saw the EspressoBin lagging behind everything else,

but the SmallPT benchmark placed it, once again in the middle of the pack.

The same could be said with the interpreted language benchmarks.

If you want to see the results from my Phoronix tests you can go here.

Or if you want to see a huge comparison against all my tested SBCs you can go here.


Armbian

Next, I thought I’d try out the beta release of Armbian. It didn’t quite boot up, but it was a beta after all. Seemed to be missing the Device Tree Table file.This involved simply adding a symlink to the correct DTB file. It was there, just the kernel expected it to be in a different place.It then booted up without issue. Alas, there were no GPIOs, I2C or SPI built into the kernel.This is something else I’ll need to look at in my review update.


Summary

So, what do I think of the EspressoBin?

Overall it’s a well designed board with a fair amount of ESD protection, but the USB console powering issue needs to be looked at.

The 1.8v GPIOs will be a real pain for most users and places this board into the ProMaker market.

It would make an excellent router, but the low TCP throughput might be an issue for some and it will be interesting to see if this improves with newer kernel releases.

The temperature stayed well within lower limits and it performed reasonably well across all the benchmarks.

Comparing to other boards; on CPU power it sits amoungst other boards such as NanoPi, Pine64, Orange Pi, original HiKey.

The filesystem performance is fairly decent, but, again, the slow network might be an issue for some.

The company is doing well building up a community around it and have put up several tutorials, but they will need to expand their documentation a bit more.

There’s still a number of tests I need to look at, and I’ll publish a review update once development has progressed a bit more.


If you like this YouTube channel please subscribe by clicking the in-video button, or the little red "Subscribe" button in your browser or app. You can also get updates by subscribing to me on Facebook, Twitter, Google+, Pinterest, and also Tumblr.