Any reason to keep going? (for me, not in general)


I love Motorola 68k architecture, it was a dream and far better compared to horrible Intel x86, but 68k is too old today and CISC design lost against RISC design for a lot of reasons.

Well the Apollo-Core is a modern implementation of the 68000 series, same base instruction set, but can do more instructions per cycle, with 64-bit instructions and extra multimedia functions including modern version the Amiga Paula sound chip that can do 16-bit audio and current core development is implementing a 3D accelerator core, this is all a fun hobby project from a former IBM Processor designer that loved the Amiga and 68k.. With a Cyclone V FPGA it roughly has the same CPU power around a Pentium II processor. If a true SoC ASIC could be made it likely could contend with some more modern processors, but this would one be cost prohibitive and legally would be a challenge.
 
  • Like
Reactions: rSl
Well the Apollo-Core is a modern implementation of the 68000 series, same base instruction set, but can do more instructions per cycle, with 64-bit instructions and extra multimedia functions including modern version the Amiga Paula sound chip that can do 16-bit audio and current core development is implementing a 3D accelerator core, this is all a fun hobby project from a former IBM Processor designer that loved the Amiga and 68k.. With a Cyclone V FPGA it roughly has the same CPU power around a Pentium II processor.

Even that modern reincarnation with new instructions, 64 bit, etc, continue being a CISC design. Modern x86-64 continue being CISC CPU and they have to "translate" those CISC instructions internally to execute them in an efficient way more similar to RISC (and they have to support old instructions). Decoding/"traducing"/etc of that "complex" instructions to simpler ones (similar to those on RISC) has a cost in energy, wasted transistors/space, material cost. It is one of the reasons ARM is more efficient.

And RISC-V, being obviously a RISC architecture, goes even further: for example it doesn't have flag register like was usual on most old designs I have know about. It can be shocking at first, but it is done that way because old design (having classical register flag like most architectures) for example implies a more difficult/costly parallelism and instruction reordering.

If a true SoC ASIC could be made it likely could contend with some more modern processors, but this would one be cost prohibitive and legally would be a challenge.

If it could be done in an ASIC it would be faster and much more efficient and even cheaper than a FPGA, if enough number of unites are produced. But there is the problem: probably there is no enough Amiga followers wanting to buy it so no ASIC can be produced at good cost (if you make small number, ASIC is more expensive than standard FPGA).

I remember my C64 Direct-to-TV. It was cheap, fast, little energy consumption, etc, as it used an ASIC instead of FPGA, but the 1st production run was 250,000 units. I bough a few for 10€ each in. And it is supported/emulated in VICE Commodore 8-bit emulator.

PD: A am no Amiga fan today (I have complex reasons to do so), but I continue being Commodore 64/128 and ZX Spectrum fan today :)
 
Last edited:
Can it play full screen video files though? I guess that depends a bit more on what graphics chip your using, but I remember it was the kind of thing that pentium 2 chips struggled at back in the day.
 
Let's go back to threat question:

Any reason to keep going? (for me, not in general)


Waiting for a RISC-V CPU board for Pyra is not a good reason if you want an open CPU/dirvers/support, because RISC-V doesn't imply open design/drivers nor good support.

BUT waiting for a NXP i.MX CPU board could be a very good reason: actually they use ARM cores, but it is the more OPEN CPU/drivers (no closed specs/doc/etc, no blobs) and best support (plus long term availability) on market. In fact NXP i.MX was selected for the more open modern computers I know about: MNT Reform notebook and Librem 5 smartphone. They selected that NXP CPU/SOC for those reasons.

It could be in a future NXP would use RISC-V cores, but that doesn't change the way (it could be RISC-V core designed by other company, for example Si-Five, with its IP (Intelectual property), i.e.: licensed cores, like ones in ARM: having an open ISA doesn't change game: there will be IP on most cores, etc).
 
Last edited:
Waiting a CPU board with a Rockchip RK3399S, like PhinePhone Pro, would be another good reason, because although it is not exactly like NXP i.MX, it would be cheaper, even more powerful and would benefit/share from PinePhone community. It would be a plus for Pyra.
 
Someone could say "but I can buy a PinePhone Pro+keyboard case and I would not have to wait for a Rockchip RK3399S CPU board for Pyra" in relation to my previous message #26).

Someone could say "but I can buy a MNT PocketReform (when it be on sale) and I would not have to wait for a NXP i.MX CPU board for Pyra" in relation to my previous message #25).

Well, the difference is that PinePhone Pro is a smartphone design. You can accommodate a keyboard case to transform it on a big pocket computer but it is much bigger than Pyra (not so pokcetable like Pyra) and it doesn't have gaming controls, only a keyboard. It doesn't compete with Pyra. Even there is a very bad point on PinePhone keyboard: it doesn't have light. I don't understand how they forgot that problem (Even they could have put one only cheap LED on top case like old Thinkpad had).

For MNT PocketReform the same: it is a pocket computer in design (we don't know specifications, but it seems it will be much bigger than Pyra, i.e. not so pocketable), and it doesn't have gaming controls, although it is likely its keyboard will be mechanical.

So none of these other "pocket computers" compete with Pyra. Pyra continue being unique: a REAL POCKET computer with keyboard (and light on keyboard) plus complete gaming controls integrated in its body.
 
Last edited:
[..]

So none of these other "pocket computers" compete with Pyra. Pyra continue being unique: a REAL POCKET computer with keyboard (and light on keyboard) plus complete gaming controls integrated in its body.
I haven't used either of those 2, but I guess, they have working sound (over cable) and so far all devices (inkl. Pyra) can't be used with a big screen as Desktop replacement. So if it is about a pocket computer ypu want to use as desktop in a dockingstation there is literally nothing atm. So for a pocketcomputer for someone who isn't into gaming, the pine is the cheaper device and has sound.
And I AM very sad about that, but the Pyra just isn't working for me...(yet, at least)
 
I haven't used either of those 2, but I guess, they have working sound (over cable) and so far all devices (inkl. Pyra) can't be used with a big screen as Desktop replacement. So if it is about a pocket computer ypu want to use as desktop in a dockingstation there is literally nothing atm. So for a pocketcomputer for someone who isn't into gaming, the pine is the cheaper device and has sound.
And I AM very sad about that, but the Pyra just isn't working for me...(yet, at least)
I've run into more issues with being 32bit arm than anything else.

Audio has some minor annoyances, but easy enough to deal with.

package support for 32bit is rare/garbage these days.

That said, things might be better with a Manjaro base.

Wish I knew where to begin, I'd be putting in time getting some of these things fixed.

I could probably pull alsa settings, mine work fine. It doesn't switch over on plugging in headphones but pavucontrol can fix that easy enough...

I wonder if jack would be better than pulse...
 
I've run into more issues with being 32bit arm than anything else.

package support for 32bit is rare/garbage these days.
Not sure I've had similar issues, but that said I build a lot from source when I can't find a package.
That said, things might be better with a Manjaro base.
Can't imagine how, trading a much more supported OS with a fringe neckbeard OS.

I could probably pull alsa settings, mine work fine. It doesn't switch over on plugging in headphones but pavucontrol can fix that easy enough...

I wonder if jack would be better than pulse...
Have you updated the OS in a while?, there is now a radio button in the mixer that you can select audio output to Headset, Speaker or USB or Bluetooth if connection...

Jack doesn't make for a good replacement for Pulseaudio, the Audio issues are a driver issue with the AESS system that HNS is working on.
 
Can't imagine how, trading a much more supported OS with a fringe neckbeard OS.
First of all, the neckbeards are diminishing at an alarming rate...

Second, Manjaro has been killing it with the pine64 stuff.

Third, ed was talking about going Rockchip so... Manjaro is an obvious choice and not exactly "fringe" in the specialty hardware realm.

IMO, Debian has been pretty meh since the loss of Ian, and Manjaro has surprised me with how well it works these days.
Have you updated the OS in a while?, there is now a radio button in the mixer that you can select audio output to Headset, Speaker or USB or Bluetooth if connection...
I stay up to date. But I also play a bit deep sometimes, probably time to start fresh. But I'm also a long time Linux guy so I tend to go straight to Pavu... then alsamixer if that doesn't work.
Jack doesn't make for a good replacement for Pulseaudio, the Audio issues are a driver issue with the AESS system that HNS is working on.
I am aware, but a well working Jack system is magical at audio routing. That said, a well working Jack system has been rare in my experience.
 
Don't forget pipewire. It's designed to do everything that both jack and pulse does with a single installation. Perhaps not surprising that debian isn't quite at this cutting edge yet though.
Not exactly. Jack is a special project of its own. For the most part, pulse already does what we would need Jack for as regular users. As much as I dislike pulseaudio... the days before it were pretty rough. I doubt pipewire will do as much as Jack. Probably about the same as pulse already does.

Pipewire was intended to bring video into the mix iirc, but later somewhat evolved to replace pulse. It was a weird evolution and I haven't kept up to date on pipewire much since it was announced.
 
Pipewire was intended to bring video into the mix iirc, but later somewhat evolved to replace pulse. It was a weird evolution and I haven't kept up to date on pipewire much since it was announced.
I'd advise you to check it out again. I adopted it over a year ago now on archlinux (just as an audio server, I don't use video much myself), and it's proved perfectly stable and functional since.
 
I'd advise you to check it out again. I adopted it over a year ago now on archlinux (just as an audio server, I don't use video much myself), and it's proved perfectly stable and functional since.
Oh, yea, I think I was using it on Suse, it was fantastic. I recently went back to Slackware, I haven't done much there other than some gaming.

It's my desktop which I generally don't like to mess around with too much. I've gotten spoiled by all these wonderful fanless devices as of late, and use them for all my other computing needs. So I haven't dug too deep into Slack.

Fast distro though.
 
..... So for a pocketcomputer for someone who isn't into gaming, the pine is the cheaper device and has sound.
And I AM very sad about that, but the Pyra just isn't working for me...(yet, at least)

We need more time to see if Pyra problems are solved. Some problems, for example no sleep mode on OMAP5, only will be solved with next CPU board, because OMPA5 on Pyra, being designed for cars, doesn't support it (they don't need it). For me it is a big disappointment as I own a lot of old pocket computers and all of them have a very big autonomy on sleep mode from which they can resume instantly.

On other side, PinePhone keyboard is a nightmare too: look at this comment #126 on other thread
And Pinephone with that keyboard is really big in size. Even it isn't backlighted (at least they could put a small LED like on old Thinkpad).
 
Back
Top