It reminds me of the Pandora!


I used emacs throughtout my undergraduate career, and struggled with it's arcane keycodes. I picked up vim when I started work, after a few years struggling with the editor in intellisense. The only thing you really need to learn to use it is that it's modal...

I think both Emacs and vi/Vim are great. Emacs is great for its Lisp-like capabilities (you can do incredible things), and I love vi/Vim for its modal paradigm and low requisites.

On other side I consider vi a must: it is a standard command in Unix/Linux, so all of them have that command (some of them have Vim simulating vi mode to accomplish that requisite). In some very small and some embed Linux you only have vi available, so it is a must.

But please, it is better we forget TEXT EDITORS on this thread and we concentrate on Pyra :)

PD: My answer for text editors was a simple help for matzesu as it seems he didn't know about text editors richness in Debian.
 
The problem whit this „just make a CPU Board whiteout the buggy OMAP 5“ is, that there may be a too big loss, if the OMAP 5 CPU Boards will not bought, so there is no money to develop a new CPU Board..
I would wait until good developers found a solution for this Low Power Mode Issue, or whe should first look how much time it’s needs for a boot up
Also : I suspekt that you can already close the Lid and put the Pyra in your Pocket when you have to leave fast, and if you have to leave fast, most of the time you will open the Pyra then after some minutes again in the train or the plane or something, so it’s not sitting more then 10 Hours In low Power Mode..
it would be cool if you don’t have to shut it off, but maybe it’s just now..

My plan was to put my Work Time Libre Office Sheet on the Pyra and just put it out of my pocket and work on this bevor I leave my Workplace, but It’s not that necessary, ..
The Pyra is still the whole Work PC in my pocket, I just have to boot it up rather then just awake it from hibernation mode..


Send from something, i guess ^^
 
Fake news, this was already debunked FFS.

I read it, but if it was fake news it is really bad news, because it implies we could have more Pyra CPU boards with bugged low-power mode CPU. I would prefer there were stock only to serve actual orders and a bit more, so this would force a faster change to new, better¹, and non bugged low power mode CPU/SOC.

Low power bug in OMPA5 is a serious bug for some of us, specially in a mobile device we would use intermittently a lot of times in the day. I know all CPUs have bugs (ask Intel :D ) but this is severe for a mobile device, of course for cars (a lot of clients for OMAP5) it is irrelevant, but our OMPA5 will be running on Pyra.

¹: For better I don't mean necessary faster, but lower power consumption for same work, more open drivers, better doc, and better and more extended support, for example like NXP I think they do.
 
The problem whit this „just make a CPU Board whiteout the buggy OMAP 5“ is, that there may be a too big loss, if the OMAP 5 CPU Boards will not bought, so there is no money to develop a new CPU Board..

No, there is no loss: all OMAP5 on ED stock will be used on all actual orders and xxx more. New CPU will be a later upgrade, for later Pyras after all OMAP5 are sold, in double way: for new Pyra orders after that point and for older owners who can upgrade their old Pyra.

Really a new CPU board with new SOC would not be a loss but a great earning: ED would sell a lot of CPU boards to old Pyra users. I think most Pyra users would want to update CPU.

New CPU board will not be ready tomorrow: it would take at least one or two years in best case.

so it’s not sitting more then 10 Hours In low Power Mode..

Some of my computers are in low power idle/suspend mode more than 10 hours. And if that mode is not running ok, like in OMAP5 (for example it set it to "low" 200 MHz mode), the worst part is that you waste much more battery capacity.

I just have to boot it up rather then just awake it from hibernation mode..

"Hibernation"¹ from RAM (suspend to RAM) is like instantaneous: you don't waste time to boot-up nor to set up all things like you left them. RAM is loaded with all content and only CPU plus slipping devices needs to wake-up

Think on smartphones: always on, instant usage (even when in offline mode).

ALL POCKET computers I have had, much before smartphones big market, had the same capability or better (there was not hungry Android spyware): I had old Casio and Sharp pocket computers, HP 95LX and HP 200LX (MS-DOS compatible) with always instant-on (the same for HP 48 calculators, that were really pocket computers), Palm PDA, Psion 5mx, or horrible Windows PDA², Zaurus SL-C pocket computer, etc (really a lot of pocket computers). On all of them, some now decades old, you had instant use from suspend mode.

That capability is a must in a pocket computer. It always was a must
, and today it continues being a must. It is a must in smartphone (a cell phone with a personal computer).

Sorry, but I know pocket computes from a lot of time ago, and I love that concept, so I love Pyra. But simply that low power mode bug in OMPA5 is really bad for a pocket computer.

And sorry, but boot up, plus setting all you stuff like it was before turning off, is a waste of time. And letting CPU on 200 MHz "low" mode is a waste of battery, reducing autonomy and generating unnecessary heat (in hot summer it must be not very pleasant).

¹: Other thing was suspend to disk: this sometimes was even slower than boot-up on some old systems. It is decades since I don't use that capability (even I don't know if it is supported on PC today).

²: I hate a lot of PDA because they were bad pocket computers for various reasons, but they were pocket computers.
 
Last edited:
I also use gvim sometimes, which is the graphical version. The one in my terminal doesn't seem to have access to the system clipboard, so if I need to yank and paste anything between apps I need to use gvim.
There are few workarounds.
1) put this in .vimrc, you will need xclip (this will not work on wayland, you will need to replace xclip with wl-copy/wl-paste there). After this you can use Ctrl-C and Ctrl-V and Ctrl-A
Code:
vnoremap <C-c> y: call system("xclip -i", getreg("\""))<CR>
noremap <C-V> :r !xclip -o <CR>
inoremap <C-v> <Esc><C-v>a
noremap <C-A> :w !xclip -i <CR>
2) use autocutsel to sync primary with clipboard, so you can paste in vim using S-Insert.
 
Last edited:
And letting CPU on 200 MHz "low" mode is a waste of battery, reducing autonomy and generating unnecessary heat (in hot summer it must be not very pleasant).

¹: Other thing was suspend to disk: this sometimes was even slower than boot-up on some old systems. It is decades since I don't use that capability (even I don't know if it is supported on PC today).

We have to hope that clever, automatic use of both low power mode and suspending to disc will do the necessaries for the great bulk of Pyra customers until the new board comes out. For example its in the former for an hour after having its lid shut, then if still not opened goes to the latter.

An interesting point you make further up the thread about bringing forward to new board --- from later to sooner --- potentially being better for ED because it allows him to sell to original customers again. My guess is that is a delicate balance in ED's business model; weighed against R&D costs, with the major caveat that nothing can be done until cashflow has stabilised. As a general rule when a new product is out the cashflow goes up and down like a yo-yo for a year, then starts to stabilise, at which point you can afford 'niceties'. We can't assume the money balances in a way that allows new board dev. within even two years, but its plausible.*

*I don't follow Pyra news as much as I could so its likely ED has already discussed all these practicalities of when a new board can be developed.

[...] Other thing was suspend to disk [...] even I don't know if it is supported on PC today [...]

Its supported on Linux Mint, so yes. If you haven't encrypted your hard-disc its nippy enough for me.
 
Powersaving?
I can set the LidClose to shut down the Device and when we maybe get an optimized BootProcess then the booting maybe will go around 5 or 7 Seconds.

I do not worry about that buggy Omap5 Powersaving Modus.

When i want to use the Pyra are 10Seconds Boot Time enough for me.

I am curious about Playstation2 and PSP Emulation and Qemu/Box86 and many other Thing.

Compiling would be more fun with the Pyra and Finaly we got the 4GB for it.

Minecraft should run Fine on the Pyra..yay.

And for compiling Things i will setup a fast SSD external on it with SATA.

And do not forget about watching TV,Netflix,Disney and other Things like dvbt2 Dongles.


And the wonderfull Light Show on everytime the Pyra Boot...many other People will ask or watch what for a wonderfull Device you own.

So i do not worry about a little Powersaving :)
 
Last edited:
No, there is no loss: all OMAP5 on ED stock will be used on all actual orders and xxx more. New CPU will be a later upgrade, for later Pyras after all OMAP5 are sold, in double way: for new Pyra orders after that point and for older owners who can upgrade their old Pyra.

Ok, I know you mean it for good, and you're right in part in some things, but this comment is absolutely false.
The cost of the chips is very small in comparison to the costs involved. It's possible the most expensive chip on board, but the real cost is the development cost of a new CPU board.
We're speaking of around 1100 orders now (I'm not sure could be 1000 could be 1500) and maybe a couple thousand or 10000 in the future.
This is very few. The cost of developing the board divided by the boards produced must still be much higher than the price of components.
It's not a matter of throwing away stock, it's a matter of throwing away hardware development time, cost of prototypes, time and equipment for testing, simulations, etc.

So we must do with the current CPU board and hope for mitigations in software development. I don't know, maybe with time we get the device to hibernate, and maybe later someone comes up with some software stack for the Cortex-Ms that can
keep enough functionality without idle mode while cortex-As and much of the system is hibernated, so that they can keep connected to networks and wake the system up in time. I'm not even trying to be realistic, just daydreaming, but it's maybe
more realistic than a new CPU board soon. That'd be very expensive.
 
Ok, I know you mean it for good, and you're right in part in some things, but this comment is absolutely false.
The cost of the chips is very small in comparison to....

...more realistic than a new CPU board soon. That'd be very expensive.

I think you are simply wrong and/or you didn't read nor understand my comment:

1/ Board upgrade have been always a promotion by ED for Pyra. It is not I may want other CPU board, it is this is in Pyra DNA: Pyra was designed and promoted with CPU board upgrade as a future upgrade.

2/ Soon? I am not saying CPU upgrade would be soon. In fact I assume and I said in my previous comment it would take at least 1 or 2 years at best once development initiated. "At best": knowing Pyra delays it would be not the best in-time.

3/ I think the worst part in Pyra development was not CPU board, but case and other small parts. I think the more time consuming and delaying was plastic parts (case, etc): It was a real nightmare. In Pyra there are a lot of development apart from CPU board (keyboard-base board, display board, keyboard,...).

4/ Upgraded CPU boards will be sold with enough profit or not sold. I think they would not be bellow 100€, but well above (it may be around 200€?). Obviously chip cost+PCB construction will be much lower than boards selling price plus development, so there will be a good profit.

5/ Sales for CPU board upgrade will be higher than Pyra sales, because you sell that board on new Pryas and you sell that spare CPU board to thousands of existing Pyra users (to upgrade OMAP5 board). So development cost would be shared by more sales than Pyra (I think most OMPA5 Pyra users would want to upgrade).

6/ Obviously a CPU board is much cheaper than a complete Pyra. And all other components on Pyra had a development cost. Do you think case an other plastic parts were free on development cost?
 
Last edited:
Powersaving?
... the booting maybe will go around 5 or 7 Seconds.

When i want to use the Pyra are 10Seconds Boot Time enough for me.

It may be good for you, but 5-10 seconds is a lot of waiting time.

Sorry, but I have had lots of pocket computers (as I listed on previous message) and ALL of them were on from suspend instantly: nor 1 second lost.

A pocket device like this will be used a lot of times, even in same day, and waiting 5-10 seconds each time is not a good enough, worse if you consider all pocket computers from decades ago were on instantly from suspend.

If you see Pyra as a Nintendo DS, it may be enought for you, but obviously for a lot of us it is much more than a NDS: it is a pocket computer, and while we will be running emulators on it, that won't be its only use.

Obviously we have to suffer that problem and live with it for a few years at least (until CPU board upgrade be on sale), but that doesn't imply we have to say there is no problem.
 
Last edited:
Then now the IMPORTANT question is: how much TI OMPA5 do you have on stock (or you can get from other sources than TI, as TI doesn't sell it anymore)? I.e.: How may Pyras with OMPA5 can you supply until you have to switch to other SOC actually on market (like NXP ones with extended support and guaranteed long time on market)?

The OMAP5 is still in production. So it's not only available from stock but even still actually produced.

For some of us a low-power idle mode and suspend mode, like we have on all smartphones and actual computers, are a very valuable point, because Pyra is a "mobile" device like a smartphone but in pocket computer format. Even on my desktop PC (actually a convertible/notebook computer) I use suspend mode a lot of, because a lot of times I want it to be where I left it instantly when I back to it. In a pocket device I would value that capability even much more, because I would use it in the move, and I will have a lot of on-suspend changes on it.

I know. I love the function on the Pandora as well.
But there's not much you can do if the chip has a bug without a known workaround.

Even my good old calculators (HP 48/50 TI89/92) have suspend mode: I turn them on and all is exactly where I left them (obviously on HP48 suspend mode, where RAM is powered on and real time clock continues beating, is rated in years, something no modern smartphone can handle). And my old Sharp Zaurus pocket computers have proper idle/suspend mode.

As said: It's not a missing feature, it's a bug.
It's not that TI hasn't implemented it - but due to a bug, it can crash while being in that low-power mode.
 
The cost of the chips is very small in comparison to the costs involved. It's possible the most expensive chip on board, but the real cost is the development cost of a new CPU board.
We're speaking of around 1100 orders now (I'm not sure could be 1000 could be 1500) and maybe a couple thousand or 10000 in the future.
This is very few. The cost of developing the board divided by the boards produced must still be much higher than the price of components.
It's not a matter of throwing away stock, it's a matter of throwing away hardware development time, cost of prototypes, time and equipment for testing, simulations, etc.

One OMAP5 is around 50 EUR.
That's a lot compared to other chips (Rockchips are around 8 EUR).
The costs of the full CPU board (12 layer PCB, 4GB RAM, 32GB eMMC, PALMAS, etc.) is around 180 EUR, so 130 EUR without the OMAP5.

Development costs for the CPU board were around 150,000 EUR.

However, what you shouldn't forget is:
* CPU board was a lot more complex than planned due to the OMAP5430 not being available. This added more layers and made the design 10x more complex.
* It was incredible hard to determine whether known bugs are hardware or software due to the fact that the OMAP5 software support stopped - old drivers from TI don't work with modern kernels and as the OMAP5 is not used in any other public Linux device, there's not a lot support either.

So if we use a common SoC which has good Linux support by the manufacturer and the community and which is available as consumer version (not industrial one), development of the CPU board will take a LOT less time.
Instead of 6 months pure layouting, it could probably be done in 1 month.

Nikolaus was even thinking about creating a quick iMX Board that fits our mainboard for a test ;)
 
The OMAP5 is still in production. So it's not only available from stock but even still actually produced.

I am going to "hate" TI :D

Of course I understand TI is more focused on industrial clients and offers long term supply, like NXP does (HP used Samsung SOC on previous calculators and have suffered from short-production life, so finally they have swichted to NXP because this guarantee extended availability).

I am not going to ask when do you expect CPU board upgrade would begin development, because I suspect you don't know it in this moment, and if you would, it could be you don't want to make it public until future appropriate moment.

As said: It's not a missing feature, it's a bug.

We know it. That was clear from beginning.
 
And I shall add that instead of using 'cl' which the l i guess stands for letter, you can use 'r' for replace so one keystroke less and that would be 0(zero)rx<Esc>
FWIW, r only takes a single character of input, so you can't use it to replace a single character with a longer word, but on the positive side, since it doesn't enter you into edit mode, you don't actually need to press Esc to get out again. Just 0rx will replace the first character with an x.

'l' (lower case L) is just one of vi's ordinal directions. 'h' is left, 'k' is up, 'j' is down and 'l' is right. You can ch to change the character before your cursor. You can also cj or ck to replace multiple lines, both the line you're on and the specified direction also (if you just want to change the line you're on, cc will do that), but in vi mode at least that doesn't drop you in insert mode after it does its work. In vim mode, it works more like you'd expect for a c command.
 
One OMAP5 is around 50 EUR.
That's a lot compared to other chips (Rockchips are around 8 EUR).
The costs of the full CPU board (12 layer PCB, 4GB RAM, 32GB eMMC, PALMAS, etc.) is around 180 EUR, so 130 EUR without the OMAP5.
Wow, the OMAP 5 is actualy more expensive than an entire Raspberry Pi 4! :confused: Which also has a faster CPU afaik (ARM Cortex-A72). And I'm not an expert but I would say, your CPU board is also to expensive for what it is.
 
Wow, the OMAP 5 is actualy more expensive than an entire Raspberry Pi 4! :confused: Which also has a faster CPU afaik (ARM Cortex-A72). And I'm not an expert but I would say, your CPU board is also to expensive for what it is.
A72 isn't particularly faster than the A15, just more power efficient and the Pi4 has 4 of them. It was also not out at the time of Pyra's initial development so I'm not sure the point of the outcry here...

Any new CPU board development would take sometime to get started and there is likely no money to start development on it until he can make it back by shipping the Pyras
 

Edit: Also Eben Upton is a higher up at Broadcom, so I'd be very suprised if he didn't get a discount on the Broadcom manufactured components.

From memory he did, because the Pi was originally just for school use so was seen as a charitable thing. And now it's grown beyond just being for that purpose he probably gets a bigger discount since they sell millions upon millions of the things!
 
Development costs for the CPU board were around 150,000 EUR.
I'd say that's cheap. But I'm no expert.
Thank you for the info. But then there were free simulations, I suspect free or cheap (beyond what was paid) labour by experts (Hi, Nikolaus, and that maverick, who got rid of the tiler chip ?), etc.
I don't mean to correct you, of course. You're better informed and can't honestly put a transparent price on help you got for free,
but we can't assume we can always develop a board that cheap. You manage in a way that makes that easier, but it's still not guaranteed.

However, what you shouldn't forget is:
* CPU board was a lot more complex than planned due to the OMAP5430 not being available. This added more layers and made the design 10x more complex.
* It was incredible hard to determine whether known bugs are hardware or software due to the fact that the OMAP5 software support stopped - old drivers from TI don't work with modern kernels and as the OMAP5 is not used in any other public Linux device, there's not a lot support either.

ACK. And that will happen again for any new CPU board.
If not that, then something else. Unknowns are to be expected when one develops.
Crappy support from vendors is to be expected. It's unfortunately a consequence of the sales model.
Even when there's support today, vendors change management, product lines get dropped, you never really know.
But not all vendors are the same, and it is wise to go to vendors with good support history, you are right.

So if we use a common SoC which has good Linux support by the manufacturer and the community and which is available as consumer version (not industrial one), development of the CPU board will take a LOT less time.
Instead of 6 months pure layouting, it could probably be done in 1 month.

Nikolaus was even thinking about creating a quick iMX Board that fits our mainboard for a test ;)

So 150000€ more for a test ? Largesse.
I don't mean to discourage you to develop a NXP or RK board. Go ahead, express yourself, it's not my money. It's not my creativity.
Etnaviv or lima/Panfrost will always be better than PowerX for my taste. I might be ripe to buy it. I'd love it to be available.
And the amount of kernel support is an important factor. You are right.
I just want to keep customer expectations somewhat low. We love positive surprises. But we won't get any if you go under.
So I'd like to have the current CPU board out first, as many sales as possible, and as many people as possible hacking on it
and seeing what it is good for and what not for. There's a lot of software development pending, that won't really happen until
the first hundreds are out there.

Wow, the OMAP 5 is actualy more expensive than an entire Raspberry Pi 4! :confused: Which also has a faster CPU afaik (ARM Cortex-A72). And I'm not an expert but I would say, your CPU board is also to expensive for what it is.
The rapsberry is build in hundreds of thousands or millions. That is what lowers the price.
Whoever wants a rapsberry can buy it easily. We don't need Evil Dragon for that. But I'm not interested anyway in a SOC that needs blobs to boot.
 
Back
Top