Jump to content

Orange Pi Zero went to the market


zgoda_j

Recommended Posts

Most likely if they are correct on the actual devicetree (boot.bin on older).
The file is just a wrapper for the python functions to be targetted to the correct GPIO numbers (or port-map).
I did the new mapping table but did not test it just changed the pin I was using to the first free GPIO which is found without any changes.
Edit the file and recompile the python to test it out...

 

I am leading towards not to use Orange Pi devices anything other than plain linux servers and leave the actual sensor reading to more suitable devices. As 1wire don´t require any custom code it should be relatively easy to use those even on zero.

 

But it is possible to correct them by hand? There is no "driver" limitation or something that would prevent from doing so?

Link to comment
Share on other sites

Only the starting pins are correct all the rest are incorrect.

Most of the pins on the zero are on PA registry unlike the "normal" PGXXX

 

https://raw.githubusercontent.com/duxingkei33/orangepi_PC_gpio_pyH3/master/pyA20/gpio/mapping.h

Since I have not test them all, which pins did you found incorrect ?

We can submit a PR if some correction is needed...

Link to comment
Share on other sites

Can you share the new file please? Even if everything is not correct it would be a good start. Do you have it on github already?

 

I did the new mapping table but did not test it just changed the pin I was using to the first free GPIO which is found without any changes.

Link to comment
Share on other sites

Sorry I am away few days and don't have the device with me.

Remove comment rows and recheck against the wiki page for the hardware its not terrible hard.

The base code has the lower header(3v) row pins first and then the upper header (5v) row.

If you are unfamilar with unix editors download the file to desktop and change it there (althouht it may need to be in unix line endings)

 

The table has values, phys pin number, gpio number and some registryprefix. Look the first row (3v pin gets skippef)

 

Can you share the new file please? Even if everything is not correct it would be a good start. Do you have it on github already?

Link to comment
Share on other sites

So after you will have time, I'd appreciate it, since I can't even find in which file in WiringOP I need to change stuff, its not so evident like in py20a.

Sorry I am away few days and don't have the device with me.
Remove comment rows and recheck against the wiki page for the hardware its not terrible hard.
The base code has the lower header(3v) row pins first and then the upper header (5v) row.
If you are unfamilar with unix editors download the file to desktop and change it there (althouht it may need to be in unix line endings)

The table has values, phys pin number, gpio number and some registryprefix. Look the first row (3v pin gets skippef)
 

Link to comment
Share on other sites

Hi everyone,

 

The topic I submitted my first two messages to on this forum is now locked and I've been referred to this thread so the below might appear to be a bit out of context.

 

I was asked a question by hmartin: "Xunlong never advertised the Orange Pi Zero as an AP! Please, show me where they said "use the Orange Pi Zero as a WiFi router". The advertisement can be found here: http://www.orangepi.org/orangepizero, look for "What can I do with Orange Pi Zero". One of the bullet points is "wireless server". I think an AP qualifies as a "wireless server".

 

As to "Why the hell do people buy an SBC if they want an AP instead? Is it really saving the few bucks? Why are people that weird and believe in advertising and then add also (wrong) assumptions to that ("if there's Wi-Fi I expect it to be performant and stable, it has to be useable as access point and of course with monitor mode too")?" by tkaiser, in my case I'm prototyping a solar controller. The OPI0 - if it's possible to get it to work stable - would be an absolute perfect fit. One of it's many functions would be to allow someone to connect to it wirelessly and see a few webpages with some graphs. I don't need any serious throughput, I'd be very happy with "crappy 2.4GHz single antenna" wifi performing at a few Mbps covering a short range (say 10 meters) only - if it's stable. Besides it's really simple to replace the antenna with something with a bit more gain but that's besides the point here. I'm not using it as a wireless AP in the traditional sense, that was just an initial test. I guess one could say I picked the right test as it provided the answer I'm seeking. Apologies for not detailing in many words (as I've done now in this post) how I was intending to use the OPI0 as an AP: not to stream Youtube but to serve a few simple webpages to a few wireless connected clients.

 

Before I however embark on picking the OPI0 (and ordering a box full of them), I need to find out if it actually works. Having paid <$10, I kept in mind that it indeed might not work. There are plenty alternatives. It'd however be great if the board I end up with can run my favorite OS (full fledged Debian) and use Python. Since this is a non-commercial open source project, it would be great if costs can be kept down too.  Heck I even almost made a donation to Armbian in appreciation for all the work the people involved with this project have put into it making it work with the OPI0. I'm glad I didn't (yet). I'm reading a very clear message (please correct me if I'm wrong) here. Apparently, the OPI0 is too cheap to be useable for any purpose (not even a paper weight ;-) ). If that is the case, I don't understand why people are even bothering supporting Armbian on a device like the OPI0. Why would you if you can't really use it for anything requiring stability?

 

Which brings up the following question. If I should not use an OPI0 for my purpose, what device should I use that can run stable? Clearly I can't trust advertising ;-). Anything costing more than $XX (please define XX)? Any (none?) of the other devices supported by Armbian?

 

Or if anyone got the OPI0 to work with "crappy wireless performance" but stable nevertheless, then I'd be very happy to hear from you. I know, I should not expect that much...

 

As to calling people weird and/or whining, I'd like to extend the invitation to keep an open mind to use cases that you might not have thought of before you do so as a matter of keeping conversations pleasant and civilised.

 

best regards

Link to comment
Share on other sites

@pacifica

 

Thank you for posting !

 

I have had the same concerns/comments here on Armbian threads ( many deleted/moved with myself receiving ban threats from tkaiser) and also on cnxsoft ...

 

Judge for yourself- read the comments:

 

http://www.cnx-software.com/2017/01/16/getting-started-with-onion-omega2-lede-wifi-iot-board-and-expansion-dock/

 

http://www.cnx-software.com/2017/01/19/orange-pi-zero-nas-expansion-board-with-sata-usb-and-av-port-sells-for-10-shipped/

 

 

I also almost made a donation and glad now I haven't !

Link to comment
Share on other sites

 If I should not use an OPI0 for my purpose, what device should I use that can run stable?

 

What if you already received best possible answers and you are simply not willing to accept the facts? In any case, do more research, try to understand the background if you wanna discuss it and stop abusing forum for your personal consulting service and we will treat you accordingly. Right @athamarmian?

Link to comment
Share on other sites

The topic I submitted my first two messages to on this forum is now locked and I've been referred to this thread so the below might appear to be a bit out of context.

 

I was asked a question by hmartin: "Xunlong never advertised the Orange Pi Zero as an AP! Please, show me where they said "use the Orange Pi Zero as a WiFi router". The advertisement can be found here: http://www.orangepi.org/orangepizero, look for "What can I do with Orange Pi Zero". One of the bullet points is "wireless server". I think an AP qualifies as a "wireless server".

 

As to "Why the hell do people buy an SBC if they want an AP instead? Is it really saving the few bucks? Why are people that weird and believe in advertising and then add also (wrong) assumptions to that ("if there's Wi-Fi I expect it to be performant and stable, it has to be useable as access point and of course with monitor mode too")?" by tkaiser, in my case I'm prototyping a solar controller. The OPI0 - if it's possible to get it to work stable - would be an absolute perfect fit. One of it's many functions would be to allow someone to connect to it wirelessly and see a few webpages with some graphs. I don't need any serious throughput, I'd be very happy with "crappy 2.4GHz single antenna" wifi performing at a few Mbps covering a short range (say 10 meters) only - if it's stable. Besides it's really simple to replace the antenna with something with a bit more gain but that's besides the point here. I'm not using it as a wireless AP in the traditional sense, that was just an initial test. I guess one could say I picked the right test as it provided the answer I'm seeking. Apologies for not detailing in many words (as I've done now in this post) how I was intending to use the OPI0 as an AP: not to stream Youtube but to serve a few simple webpages to a few wireless connected clients.

I checked the first 10 pages of this subforum and found 11 threads describing different OPi Zero wireless related issues and questions, not counting this one and locked development related one. I think it's a good indication of what you should expect from this wireless module.

The thread that you posted in was related to the ongoing development and improvements of the driver, and after a dozen messages it turned into off-topic discussions, so it was closed for a good reason.

 

Also please keep in mind that Armbian is a non-commercial project too, not affiliated with Xunlong. We didn't make the board manufacturer use this wireless module, we didn't advertise any wireless related specifications or abilities like perfectly working AP mode, and constant unending spam of same questions and reports can make any developer regret that he decided to support this device.

 

I'm reading a very clear message (please correct me if I'm wrong) here. Apparently, the OPI0 is too cheap to be useable for any purpose (not even a paper weight ;-) ). If that is the case, I don't understand why people are even bothering supporting Armbian on a device like the OPI0. Why would you if you can't really use it for anything requiring stability?

Opi Zero it's still a good device for its price, just imagine that it doesn't have onboard wireless module. Want similar device with wireless? I think NanoPi Air may be a good alternative.

Link to comment
Share on other sites

@Igor

 

You mixed up the quote, and misspelled too !

Big hurry you have ;)

 

@zador

 

Good. Finally you clear up that OPi0 today simply doesn't have WiFi.

 

Now wouldn't it be nice to have a FAQ upfront on Armbian homepage to clear up all this confusion, updated monthly?

 

But you guys are happy instead spending 80% of your time on repair jobs. Where is the innovation?

Link to comment
Share on other sites

@Atharmian, no worries. Thanks for suggesting to look at Omega2. Debian/Armbian is a much better fit for my purpose than OpenWRT though.

@Igor, it seems you missed that "If I should not use an OPI0 for my purpose, what device should I use that can run stable?" is a rhetorical question. I for one don't believe the OPI0 can't be used for anything requiring stability but will stand corrected if being told so by the people who actually have in-depth knowledge of this SBC and the OS (drivers). If it is true that Armbian-on-OPI0 is a total waste of time from all perspectives, as tkaiser seems to suggest, then that might be true for any of the other boards supported by Armbian too, right? Or would you suggest I buy one of each, try to get Armbian running on it, ask a question here if something fails, possibly get abused verbally for asking and then .... eh, leave? And please, if you can, stay friendly and amicable.

 

@zador.blood.stained, thanks for answering and suggesting to have a look at the NanoPi Air. Are you saying I can expect stability from the OPI0 running Armbian, except for the wireless? Some here (tkaiser for example) seem to suggest that I can't really expect anything with a price tag of <$10 to work at all and that I'm actually rather stupid thinking I could! I am well aware that Armbian is a non-commercial open source project, hence my thoughts to give something back (in this case $$) in appreciation for the work done. I must say though that being the recipient of abusive language (not from you) simply for asking a few questions after reading through the many posts (granted: not all) on this subject is quite effective as a disincentive.

Link to comment
Share on other sites

@pacifica

 

I actually meant looking at comments on both articles at cnxsoft than the article itself .... It tells you how much respect certain Armbians get trying to answer questions or attempting to abuse others, at more open forums.

 

That is the fun part here. Folks here call everything crap, but keep working on it too !

 

Why?

Link to comment
Share on other sites

@martinayotte

 

The main thing some other folks and myself were looking for was clarity, aka honesty, but it was like pulling teeth.

 

If you read my remarks you will have to acknowledge the many grains of truth.

 

Armbian wouldn't need many contributors and certainly would have much less arguments/confusion if you guys don't spend 80% of your time on repair jobs and insulting folks as opposed to having a decent FAQ and asking, say OPi, to do eMMC on OPi0 as well, a better WiFi job etc.

 

Not all Armbian devs seem to waste their time as above but there are some who claim to be big time "designers" of SATA over USB2 stuff, or GigE over H2+ !

 

At least acknowledge the facts: is above any innovation, or even a positive attitude?

 

As I keep saying, even volunteers have responsibilities- or they are just people running amok.

Link to comment
Share on other sites

I don't understand you !!!

Who has "positive attitude" or "negative attitude" here, outside of YOU ...

The hardware design choice of Xunlong about eMMC on OPi0 is NOT an Armbian design choice !!! Armbian has NO relationship with Xunlong except to get some sample board to help community !

 

As I keep saying, even volunteers have responsibilities- or they are just people running amok

 

(I'm usually a quiet gentle guy, but here, please @igor, I'm getting out, if he fight, please BAN if forever)

What the FUCK ! As Volunteer, I'm should simply say to you :  GET OUT OF OUR FORUM !!! YOU'RE A SICK GUY !!!

 

EDIT : as I said eailer, STOP complain and become YOURSELF a CONTRIBUTOR !

Link to comment
Share on other sites

is a rhetorical question.

 

Exactly. Your questions doesn't need more answers and that's the whole point. You have to decide or do something else in your life. If you want to have the same level of understanding and a big picture, you need to invest months into learning. One 2 one learning service is nothing what you can get from us. Not for money and not for free.

Link to comment
Share on other sites

@zador.blood.stained, thanks for answering and suggesting to have a look at the NanoPi Air. Are you saying I can expect stability from the OPI0 running Armbian, except for the wireless? Some here (tkaiser for example) seem to suggest that I can't really expect anything with a price tag of <$10 to work at all and that I'm actually rather stupid thinking I could! I am well aware that Armbian is a non-commercial open source project, hence my thoughts to give something back (in this case $$) in appreciation for the work done. I must say though that being the recipient of abusive language (not from you) simply for asking a few questions after reading through the many posts (granted: not all) on this subject is quite effective as a disincentive.

This depends on your definition of "stability". These boards (most of them) will run perfectly stable on your desk (especially when using mainline kernel), but please keep in mind that these are the cheap development boards and not the industrial-grade PLCs. These are not designed to work in wide temperature range (especially below 0 Â°C), have minimal or close to zero power supply voltage filtering capability, minimal ESD, overvoltage, overcurrent, wrong polarity, strong EMI, dust, dew, moisture, etc. protection, and reliability may be affected by unclean shutdowns, so if you design any kind of hardware project, you have to know what these boards are capable of and treat them accordingly.

Link to comment
Share on other sites

The advertisement can be found here: http://www.orangepi.org/orangepizero, look for "What can I do with Orange Pi Zero". One of the bullet points is "wireless server". I think an AP qualifies as a "wireless server".

 

Show me where on that page is says "AP"

 

A wireless server can be any computer connected via WiFi which exchanges data. e.g. it's connected as a client and runs an HTTP server where you can download files.

 

Boom! Wireless. Server.

 

 

I also almost made a donation and glad now I haven't !

 

So you won't donate to us, volunteers who spend our own time and money to buy the hardware and improve it, because we didn't make a perfect product that can do everything you dream of?

 

Now wouldn't it be nice to have a FAQ upfront on Armbian homepage to clear up all this confusion, updated monthly?

 

But you guys are happy instead spending 80% of your time on repair jobs. Where is the innovation?

 

If people like you didn't complain so much, we might actually have the time and motivation to do "innovation"

 

This is the n'th time you've suggested that we take even more time away from "innovation" to create an FAQ for lazy people like you who can't be bothered to read the forum or mailing list.

 

I will again suggest that you can spend your own time making the monthly FAQ, instead of posting complaints on the forum about how we don't follow your every suggestion.

 

 

@pacifica

 

I actually meant looking at comments on both articles at cnxsoft than the article itself .... It tells you how much respect certain Armbians get trying to answer questions or attempting to abuse others, at more open forums.

 

That is the fun part here. Folks here call everything crap, but keep working on it too !

 

Why?

 

Excuse me? Our answers are 'abusing' you?

 

Please. Get a grip. We're the ones actually improving the features you all are using, and what do we get? You publicly congratulating yourself for not donating money to the cause and complaining that it doesn't work.

 

Why do we work on this at all? Maybe because we bought the hardware, realized that it has limitations, and are working to improve it for our own use. We could certainly improve it and keep the improvements to ourselves. We're not selling a product, we're not required to push our patches upstream.

 

Unlike you, I don't go to internet forums of open source projects and whine that my expectations weren't immediately satisfied. I'm actually trying to improve things to the point where your completely unrealistic expectations ARE realized.

 

 

Armbian wouldn't need many contributors and certainly would have much less arguments/confusion if you guys don't spend 80% of your time on repair jobs and insulting folks as opposed to having a decent FAQ and asking, say OPi, to do eMMC on OPi0 as well, a better WiFi job etc.

 

At least acknowledge the facts: is above any innovation, or even a positive attitude?

 

  1. If you want to see "innovation" go use the Debian image from Xunlong: http://www.orangepi.org/downloadresources/
  2. Do your own research before buying. You're presumably an adult, it's not our job to educate you before you spend money. We're able to spot marketing bullshit before buying, and adjust our expectations accordingly. I'd suggest you file this under a "learning experience" that not everything the salesmen tell you is true. At $10, it's a pretty cheap life lesson.
  3. We'd have a much better attitude if people like you didn't show up to lecture us on how we should be "innovating" more. The best thing you personally can do for innovation right now, is put away your Orange Pi Zero, stop posting stupid missives like the one above, and come back in 6 months.
Link to comment
Share on other sites

Show me where on that page is says "AP"

 

A wireless server can be any computer connected via WiFi which exchanges data. e.g. it's connected as a client and runs an HTTP server where you can download files.

 

Boom! Wireless. Server.

OK, no need to argue about this. Even large companies do make mistakes, false claims in advertising their products and sometimes ignore feedback. Again, Xunlong, by the looks of it, doesn't have a software department - this helps keeps the prices as low as they are, but this means that testing coverage and software situation won't be as good as with more expensive boards from other vendors.

Link to comment
Share on other sites

@Atharmian, no worries. Thanks for suggesting to look at Omega2. Debian/Armbian is a much better fit for my purpose than OpenWRT though.

 

OpenWrt is far more suited to industrial environments. Most OpenWrt hardware uses NOR flash, which does not degrade via reads, NAND does.

 

Wikipedia has more details on this: https://en.wikipedia.org/wiki/Flash_memory#Read_disturb

 

You can even buy OpenWrt devices which use NAND, if you need more storage. These routers have more than enough horse power for almost any M2M application, and you can buy them pretty cheap: https://www.aliexpress.com/item/Original-Xiaomi-WIFI-Router-3-WiFi-Repeater-1167Mbps-2-4G-5GHz-ROM-128MB-Wi-Fi-Roteador/32688259842.html

 

 

If it is true that Armbian-on-OPI0 is a total waste of time from all perspectives, as tkaiser seems to suggest, then that might be true for any of the other boards supported by Armbian too, right? Or would you suggest I buy one of each, try to get Armbian running on it, ask a question here if something fails, possibly get abused verbally for asking and then .... eh, leave? And please, if you can, stay friendly and amicable.

 

Supporting Linux on these cheap SBCs is not a total waste of time, but it might be for your use case.

 

Yes, you absolutely need to buy multiple boards and test them for your use case. Or you can call a distributor like Digikey and ask them for the right board for your application. You will pay a lot more for this, it won't be as new as the Orange Pi Zero, it will probably use some vendor proprietary toolchain and Linux image, but it will work for 10 years and your customers won't hate your product because you thought it would be good to build it around a $10 SBC from China.

 

@zador.blood.stained, thanks for answering and suggesting to have a look at the NanoPi Air. Are you saying I can expect stability from the OPI0 running Armbian, except for the wireless? Some here (tkaiser for example) seem to suggest that I can't really expect anything with a price tag of <$10 to work at all and that I'm actually rather stupid thinking I could! I am well aware that Armbian is a non-commercial open source project, hence my thoughts to give something back (in this case $$) in appreciation for the work done. I must say though that being the recipient of abusive language (not from you) simply for asking a few questions after reading through the many posts (granted: not all) on this subject is quite effective as a disincentive.

 

No one is saying you can expect stability from an Orange Pi Zero for your use case. You need to test it yourself and figure out whether it's stable or not. We don't have the resources to test the hardware in your use case.

 

Please, email Xunlong and ask your questions. See if you get any response back, or how helpful it is.

 

This forum has a search feature. A contributing factor to this "abusive language" as you put it, is that 20 people ask the same questions over and over again instead of searching the forum for the last time this question was asked, looking at the answer, and then accepting that the situation isn't as they had hoped.

Link to comment
Share on other sites

A contributing factor to this "abusive language" as you put it, is that 20 people ask the same questions over and over again instead of searching the forum for the last time this question was asked, looking at the answer, and then accepting that the situation isn't as they had hoped.

 

 

Well, main (and IMO the only) contributing factor to the abusive language got banned, so it should be nice and quiet here now. For anyone wondering you can read [1] and [2] and decide for yourself if this can be called a constructive discussion.

 

I can only apologize for the use of possibly inappropriate phrasing from one of our devs/moderators.

 

 

Link to comment
Share on other sites

Hello, all!  I got my second Zero in the mail yesterday (512MB w/2MB SPI).  I also ordered the NAS board.  Is there any kind of observations I can make or testing that anyone would like done?

 

Can you perform some benchmarks on the SATA performance? tkaiser has some interesting info about the SATA chip used, it seems like the performance will be above what USB 2.0 to SATA bridges normally offer.

Link to comment
Share on other sites

Can you perform some benchmarks on the SATA performance?

 

Well, even with mainline kernel and switching to performance governor IO performance won't be that great especially with Armbian default settings (designed for IoT use cases --> 408 MHz DRAM clock will lower maximum bandwidth possible on USB2 ports). Obviously this NAS Expansion board is not the best companion for actual OPi Zero so we can only hope for more Zero variants making use of the 13-pin header that feature also Gigabit Ethernet (A NAS with just Fast Ethernet and slow Wi-Fi isn't that great... but that's a no-brainer)

 

The specific USB-to-SATA bridges used here (JMS578) are interesting since they allow the use of the following features:

  • USB Attached SCSI protocol (UAS -- usually an USB3 feature but works with Allwinner SoCs even with USB2 and greatly increases random IO througput if the device behind the USB-to-SATA bridge is not the bottleneck, so typically an SSD, sequential throughput also increases slightly compared to USB's mass storage mode)
  • SCSI / ATA Translation (SAT) which enables S.M.A.R.T. (useful to query health data or even trigger self tests on the device) and the use of hdparm to tweak specific settings of the disk in question
  • TRIM/UNMAP which helps SSDs with performance, efficiency and also less writes (write amplification)

@willmore: please try the Expansion board with mainline kernel, install smartmontools and check manual pages of both smartctl and hdparm for useful stuff. Unlike most 'el cheapo' SATA bridges all the stuff should work out of the box. Though I've not the slightest idea what the prerequisits for TRIM are (kernel support for example) and how to test whether TRIM works or not (but for 'software vs hardware chicken-egg problems' it's great that Xunlong chose JMS578 since now at least some work can start if TRIM currently won't be possible with Linux through USB).

Link to comment
Share on other sites

Can you perform some benchmarks on the SATA performance? tkaiser has some interesting info about the SATA chip used, it seems like the performance will be above what USB 2.0 to SATA bridges normally offer.

 

Can do.   I have a few USB<>SATA bridges I can use to compare it to.  I can also do tests with those bridges on a USB3 host to show the headroom available.

 

 

Well, even with mainline kernel and switching to performance governor IO performance won't be that great especially with Armbian default settings (designed for IoT use cases --> 408 MHz DRAM clock will lower maximum bandwidth possible on USB2 ports). Obviously this NAS Expansion board is not the best companion for actual OPi Zero so we can only hope for more Zero variants making use of the 13-pin header that feature also Gigabit Ethernet (A NAS with just Fast Ethernet and slow Wi-Fi isn't that great... but that's a no-brainer)

 

@willmore: please try the Expansion board with mainline kernel, install smartmontools and check manual pages of both smartctl and hdparm for useful stuff. Unlike most 'el cheapo' SATA bridges all the stuff should work out of the box. Though I've not the slightest idea what the prerequisits for TRIM are (kernel support for example) and how to test whether TRIM works or not (but for 'software vs hardware chicken-egg problems' it's great that Xunlong chose JMS578 since now at least some work can start if TRIM currently won't be possible with Linux through USB).

 

The potential reuse of the 13 pin header was what I was thining as well.  The OpiZ+NAS is a silly combination. A Z2 with an H5 and GigE would make a lot more sense.

 

I'll do as you requested and report back here when I have something.

Link to comment
Share on other sites

Can do.   I have a few USB<>SATA bridges I can use to compare it to.  I can also do tests with those bridges on a USB3 host to show the headroom available.

 

Well, you need really fast SSDs or a RAID when combined with HDDs to be not bottlenecked by the storage behind the bridge. With JMS567 (USB 3.0) I was able to get sequential transfer speeds up to 400 MB/s with a decent USB3 host controller (Intel) and I would assume this will be the same with JMS578 now (it's USB 3.1 but only 'USB 3.1 Gen1' which is 5Gbps too. You would need USB 3.1 Gen2 to be able to benefit from 10Gbps data rates but will then be bottlenecked by SATA implementation/spec since here data rates max out at 6 Gbps).

 

These are results with different SSDs and USB-to-SATA bridges tested on a rather low-end 2 year old MacBook Pro (the slowest 13" variant with just a dual core i5): http://kaiser-edv.de/tmp/3zK9lI/ (please note the differences depending on the test profile, those SSDs that implement Samsung's 'TurboWrite' look good when tested with data sizes that fit into TurboWrite buffer and slow down a lot if the size exceeds the buffer and also block size of the tests -- 128KB vs. 1 MB -- makes a huge difference in numbers)

 

I would assume you get identical numbers with Xunlong's Expansion board but I doubt a bit an USB3 capable Zero (we most probably will never see :) ) could use 5Gbps 'SuperSpeed' via such a GPIO header (same when trying stuff like that with jumper wires). So still what JMS578 makes special/interesting is not USB3 capabilities but support of those other standards/specs I listed above.

 

Even more off-topic: When you look at the last two screenshots from the above link you see performance of the PCIe SSD in this low-end MacBook Pro. Part of my job is preparing whole server installations inside VMWare Fusion (on a faster MacBook Pro using also a faster SSD). It's always fun if you test a virtualized server on a laptop and then deploy the installation on a customer's vCenter hardware and everything storage related and especially databases slow down as hell... since storage in data centers is way slower compared to what you get with 'the average laptop' today :)

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