all this talk about freescale is a dead end with power ??. but is power8+ and 9 not comming or ??

@retro

I think people are most interested in this topic, is people who has the most agent PowerPC.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@retro

Quote:

but is power8+ and 9 not comming or

You'd have to ask Trevor, everyone else is bailing to X64.
(Actually, that's not 100% accurate, there is also the T2080 laptop projects that Acube is helping to design).

And NXP will still have PPCs ten years from now, so supply isn't an issue.

Who know if Power will ever make it down to the desktop level again.

_________________
Amiga! "Our appeal has become more selective"

@Srtest

Quote:

What you are writing sounds to me at its basic form as computer logic. What happens when you need to compare to elements that are based on a different kind of logic or at least doesn't correspond to each other?

I don't like the idea of different "logic"'s, unless you're talking about ternary logic, but I don't think you are, because you'd be way off base.

I think you mean this: What happens when you need to solve a different problem that requires a different approach?

My answer: there are standard solutions for all these problems. For example, (I may get some details wrong here so I'd appreciate correction) there's a standard for Bluetooth communication. There's a standard for the internet (TCP/IP). There's a standard for web pages (original HTML, then HTML+"Java"script, then HTML+"Java"script+CSS). There's a standard for how to write files to a device. (Lots of standards for that, actually.) And so forth.

Most operating systems have standard APIs that take care of standard problems in a standard way. The Java virtual machine programmer need "merely" figure out how to translate Java's approach to the underlying OS.

Suppose, for instance, you need to send data from a client to a server. It's up to the programmer to encode the data in a way the server can understand it, but he'd have to do that on any platform, and he'd be doing it more or less the same way on each platform; with Java he does it that one way, once. The programmer then invokes the appropriate Java libraries, which in their turn implement networking standards and communicate with the host OS just as any ordinary program would.

If he were writing it in a different language, for instance C++, he wouldn't be able to do that in the base language and its libraries. The standard C++ library does not offer a networking interface; a programmer either figures out how to compile code for 10 different clients (say) or he chooses one of several cross-platform libraries that are not part of the C++ standard (such as QT).

Quote:

Is Java like the web in itself or is like the web if it can characterize it first?

I don't really understand what you're saying. I think what I just wrote above addresses this, especially when I say that there is a standard way for communicating across the web, and Java's libraries offer an interface to do this, which the programmer can use, and the JVM then translates to the underlying architecture (OS+hardware).

Quote:

I understand the enviroment argument I just don't get the merit of that as amiga is also a kind of corresponding signals of a language.

Amiga isn't a language, so I don't understand what you're saying. Amiga is an OS and/or an idea about computer organization.

AmigaOS offers certain API's that allow one to do certain things; these include intuition.library for user interface, dos.library for disk devices, and so forth. But, for instance, there wasn't a standard networking library in Amiga until OS 4.

Amiga hardware offers certain features that weren't found anywhere else in 1985. You could access those through the APIs, but you could also "bang the metal" if you wanted to. Banging the metal used to be necessary but it's generally frowned upon these days, because you want code that's readable, maintainable, and so forth.

Let's suppose then that someone wanted to implement Java on AmigaOS. This would require two major steps. First, she'd first have to port a compiler, which might actually be the easier part of the job, because the bootstrapping compiler is probably written in C, which "should" port easily enough. But then comes the hard part of writing a virtual machine for AmigaOS. She has to write code that will recognize when a Java program is requesting to open a window, and she has to turn around and tell intuition.library to do that. That's not so bad, but what does she do when the Java program requests to open a network socket and communicate? She has to sort through AmigaOS' networking library, figure out how AmigaOS sets up and conducts the communication, then translate Java API functions into AmigaOS functions.

And so forth.

Quote:

If AmigaDE was made then perhaps it would be able to be something that can benefit us more than a something like java that simply has its own basic elements which are easy to move from platform to platform but maybe hard to shape to benefit you specifically.

The point of a general-purpose programming language like C, C++, Java, and so forth is to move from platform to platform. What distinguishes Java from C or C++ is that (a) Java doesn't compile to bare machine code, so that the programs can be sent via internet and run most anywhere they end up, and (b) Java has a much, much larger and (in many cases) better-designed standard library.

Quote:

The net doesn't need to benefit you as it is a benefit in itself. I would think a programming language should have merit for different uses and implications in a way that is not generic.

I'm guessing you haven't heard of Turing completeness. If a programming language is "Turing complete" then it can solve any problem that any programming language can solve. It might be harder in some languages to solve a certain problem than in others, but it can still be done. Java is a Turing-complete language.

So, for instance, a Java program could work with the Xena chip. It would take a lot of work, and in this particular case might be much harder than using C or C++, but if someone were to write a Java library for Xena it then it would be quite easy to work with it, and would probably be easier than working in C or C++. Of course, it would only work with a particular hardware combination, and would use a non-standard virtual machine, but it could be done in principle.

This gets us back to Android, for instance. Android doesn't actually implement all the standard Java libraries, and its virtual machine works differently than the standard Java virtual machine. This is due to various design choices made by Android, based on their particular needs/desires. But they all run every Android app (with some limitations) and Android Studio will even give you a rough estimate of how many devices can run an app you write, based on which API you target.

Quote:

I have a feeling a lot of smart devices don't use Java if all they need is talk to each other and either be about enabling a few features or be based on a streamlined version of linux that both can communicate directly and fit a wide area.

Either I don't understand you or you don't know what you're talking about. But Android is the OS for the vast majority of mobile and "smart" devices (not talking about embedded devices; that's another story altogether) and the standard programming language for Android is Java.

You talk about having a core that you can meld to suit your needs better. Every programming language seeks to do that, including Java: it has a small core (the language itself and its virtual machine) with an enormous library, which you can view as a toolbox to help you meld the language to suit you better. It isn't a one-trick pony like some languages which are designed only for a particular task.

Quote:

This sounds like what Linux gives you with its abstraction of api and layers and backends and then you also have Mozilla and their contribution to that. Maybe i'm wrong I don't know if we do it similiarly here.

Sure, but that's what the Amiga APIs do, too (Intuition, Exec, DOS, etc.). It's not just the datatypes libraries; the APIs themselves abstract away the hardware and give one a nice(r) framework to communicate with the computer. That's what the IFF file format was for, too.

One of the nice things about Amiga is that they were one of the first home computers to offer a well-though-out API that abstracted the hardware and allowed you to write code that was insulated against future upgrades. People who wrote software that banged the metal were stuck on a particular architecture, so for instance a bunch of games that worked on some Amigas didn't work on others. But people who wrote to the APIs were generally insulated against that sort of thing: a program that worked on AmigaOS 1.0 would still work on AmigaOS 3.9 (usually) without even recompiling. They'll probably work without emulation on AmigaOS 4.x if you just recompile the code.

Java takes this idea to a different level.

_________________
I've decided to follow an awful lot of people I respect and leave AmigaWorld. If for some reason you want to talk to me, it shouldn't take much effort to find me.

@retro

Quote:

retro wrote: all this talk about freescale is a dead end with power ??. but is power8+ and 9 not comming or ??

As far as I understand this

https://en.wikipedia.org/wiki/POWER8

and this

https://en.wikipedia.org/wiki/POWER9

Is only used and produced by IBM.

Freescale is not involved in Power8 or Power9 it seems they stopped using newer PPC ISAs in their embedded processors.

I'm not sure if Freescale still exists or if is nothing more as a brand name used by NXP.

Last edited by OneTimer1 on 23-Oct-2017 at 08:34 PM.

@OneTimer1

Power8 and Power9 are open platforms, with companies other than IBM participating (even Chinese companies).

If NXP isn't interested, that simply their choice.

If they wanted to update their 64 bit PPCs, all they'd need would be to incorporate IBM's VMX instructions (since the latest PPCs from Freescale are bi-endian).

Then their products would still be supported by Linux.

BUT, this is NOT happening.
Again, that's NXP's choice. No one is excluding them from the party.

_________________
Amiga! "Our appeal has become more selective"

Only time will tell if there will be Power Architecture CPUs/SoCs suitable for our niche after 10 years from now.
Now we just know that NXP (ex Freescale) is "not planning" to do it.

IBM Power is an overkill currently.
Perhaps it's a safe bet that someone (from China) will continue some PowerPC-like production (via OpenPower), but it's unsure if it would be sane option for us.

+
So far it would seem these will be the last ones from NXP: T2081(4/8 cores), T4161 (8/16cores) and T4241 (12/24cores). Plenty of power to use for next 10+ years, but that's it.

NXP future high performance chips seem to look like this:
https://www.nxp.com/products/microcontrollers-and-processors/arm-based-processors-and-mcus/qoriq-layerscape-arm-processors/the-lx2160a-multicore-communications-processor:LX2160A

Last edited by KimmoK on 24-Oct-2017 at 05:04 AM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Arm will gave all needed power for the nxp present and future network Device.
The ppc cpus are big expensive and to hot compared the arm cpu.
Plus are just for network and not for desktop computing ... Dont are in line with today cpu calcs power, have reall poor derser Lines and poor RAM speed. T2080 is better compared the p5040 in RAM speed but have Less mips for thread awerall performances arte not equiparabile with an old G5 970fx or 970mp.

_________________
I love Amiga and new hope by AmigaNG
A 500 + ; CDTV; CD32;
PowerMac G5 Quad 8GB,SSD,SSHD,7800gtx,Radeon R5 230 2GB;
MacBook Pro Retina I7 2.3ghz;
#nomorea-eoninmyhome

@KimmoK

NXP offers ARM based lowprice servers so when they bought Freescale and talked about reducing costs by common operations like development it was obvious that this would mean that PPC development ends. They will produce and sell existing PPC processors as long that is justified by profits and as long they do not have sectors that are more profitable replacing it. Perhaps there are still contracts they have to fulfill, but in my view it is a matter of time until they stop producing PPC. That is business, NXP is not interested in technology but making profits and Freescale is no longer independent. Aeon will certainly not decide it because it is a tiny niche for them. The IBM processors certainly will be produced for some time and of course developed but they are too expensive however you look at it. So finally dawn is there for current NG systems, at leasr as long they are based on PPC.

@tlosm

Expensive or not in production, NXP is more or less based on ARM, Freescale offers ARM too, if the buy makes sense for NXP (and certainly was reason why they bought Freescale) then the concentration on ARM and common development and production to safe money. The other areas where Freescale was involved like PPC are not interesting in a economic sense, and heritage is not somoething business man normally are interested in. They will produce as long it is enough profitable to justify it, when this is no longer the case they will drop it. And they will not invest more than they need to. It is about milking cow until cow is not giving any more milk...

@thread

it looks like the Linux people WANT an ARM motherboard, but not using a mobile processor.

https://www.pcworld.com/article/3184230/computers/the-anatomy-of-a-powerful-desktop-with-an-arm-chip.html

It must have:

Processor on a card - ARM Rockchip RK3399 processor (or better x2)
DDR4 memory slot(s)
USB 3 slots(s)
NVMe slot(s)
SATA 3 slots(s) 6gbs, x8
gigabit slot(s)
PCI-E x7
GPIO header
USB-C

Linux and Android (for a games machine set top box)

"minecraft 8k resolution", UNDER 900 USD

With high specs, A-EON could enter the server space arena
Linux machines build for users who need Video Edit/ high performance systems.

In time these systems could run AmigaOS (obviously)
Going to need a few Electrical Engineers to create motherboard & drivers, small dedicated team just for linux only.

_________________
Download 499.26 Mbps, 659.94 Mbps Upload :)

Qualcomm CEO Steve Mollenkopf said Qualcomm will let its $47 billion NXP offer expire.

@KimmoK

Removed Spam

Last edited by _Steve_ on 28-Sep-2023 at 11:26 PM.

@jack1999

Another thread from the dead brought to you by SPAM.

_________________
I sold my SAM460ex lite... waiting for money to buy a Raspberry Pi... or a Classic A1000 with Buffee... or an A1222... and OS4.3 FE update 11