@cdimauro

Quote:

I'm looking forward to see some real-world application.

Me too.

@cdimauro

Freescale announced successors to the P5020 and P5040 with an AltiVec enhanced version of the E5500 core about two years ago.
These were to be the T5020 and T5040.
As there is no longer any information on these available on Freescale's website, they may be canceled.

I have filed a Service Request asking for clarification.

If Freescale does not enhance the e5500 core in the way it was announced that they planned to do, it would still not prevent me from endorsing the P5020, P5040. or T1024, as AltiVec only provides a small performance increase.

In particular, the T1024 is attractive as it costs a small fraction of the other processors.

But, I think I have already stated that our best move would be an e6500 cored cpu.

And of course as a MorphOS user, at least one system I will have under that system in the future will be AMD64 based.

I like PPCs and will use them as long as a decent product is available, but we all may HAVE to move eventually.

@pavlor

Quote:

pavlor wrote: @cdimauro Quote: I'm looking forward to see some real-world application. Me too.

This will be the crux of the biscuit.
But as I am familiar with some of the people at Varisys, I am expecting a very competent system.

@iggy

Quote:

iggy wrote: @cdimauro Freescale announced successors to the P5020 and P5040 with an AltiVec enhanced version of the E5500 core about two years ago. These were to be the T5020 and T5040.

As I said before, except for the T1 family, the T2 and T4 are based on the e6500 core, so I think that the T5 (if any, but I don't believe anymore; see below) will continue to use a processor of this family, which integrates Altivec.
Quote:

As there is no longer any information on these available on Freescale's website, they may be canceled.

Can be, because all processors vendors are moving to ARM, and some time ago Freescale has announced that she was putting all R&D effort on this family. PowerPC seems to relegated to a legacy plan, like the old 6800, 68000, Coldfire, etc. families.
Quote:

I have filed a Service Request asking for clarification.

Good. Please, let us know.
Quote:

If Freescale does not enhance the e5500 core in the way it was announced that they planned to do,

IMO it's the e6500 that would have an enhanced Altivec (but it requires new development, which is unlikely: see above). The e5500 covers the low-end segment, whereas adding Altivec means that it overlaps the e6500 segment.
Quote:

it would still not prevent me from endorsing the P5020, P5040. or T1024, as AltiVec only provides a small performance increase.

It depends on the application. Some application like mplayer doesn't make sensible benefit from the lack of Altivec. However, other applications do.

Some benchmarks were published here some time ago. I think pavlor can help, if you're interested.
Quote:

In particular, the T1024 is attractive as it costs a small fraction of the other processors.

I took a look at the specs, but it has too low frequencies. A T2080 is a much better option, because it uses the e6500, can run up to 1.8Ghz, has much more L2 cache; but costs a bit more than double. Anyway, we aren't talking of huge amounts of money.
Quote:

But, I think I have already stated that our best move would be an e6500 cored cpu.

Absolutely.
Quote:

And of course as a MorphOS user, at least one system I will have under that system in the future will be AMD64 based. I like PPCs and will use them as long as a decent product is available, but we all may HAVE to move eventually.

At the end you run applications, and you don't "touch" the processor (unless you like to code in assembly ).

Since PowerPCs is a died platform (no new microarchitectures will come), it makes sense to move to something else which can ensure a future for the platform. But that's something which I repeat from long time...

Last edited by cdimauro on 30-Aug-2015 at 02:35 PM.

@iggy

Quote:

iggy wrote: @pavlor Quote: pavlor wrote: @cdimauro [quote]I'm looking forward to see some real-world application. Me too.

This will be the crux of the biscuit.
But as I am familiar with some of the people at Varisys, I am expecting a very competent system.[/quote]
Varisys engineers can only assemble the parts, but they cannot make changes to the processors...

@cdimauro

Quote:

It depends on the application. Like Rob said, maybe an application like mplayer doesn't make sensible benefit from the lack of Altivec, but others do.

Huh. I never said that. I was saying that with the higher performance of the E5500 vs PA6T might be enough to edge it past PA6T in Mplayer. I was not commenting on How well Mplayer utilises the Altivec insstruction set.

There are very few Amiga OS apps that make use of Ativec anyway. Mplayer, Lame and DVplayer are the only ones I can think of right now.

@Rob: sorry, I've corrected the text.

@Rob

Quote:

There are very few Amiga OS apps that make use of Ativec anyway. Mplayer, Lame and DVplayer are the only ones I can think of right now.

And blender both directly via libbffmpeg.so and indirectly via miniGL , which is altivec optimised.

_________________
BroadBlues On Blues BroadBlues On Amiga Walker Broad

@cdimauro

And good friend and I have been going over old documentation and roadmaps from Freescale.

He pointed out that we had discussed the disappearance of the e5500 cored T5s from Freescale's literature in '012.

To quote:
"The Wayback Machine link now shows the site on October 7th already having been without mention of the T5. Thus, the removal of the T5 must have happened somewhen between October 4th and October 7th."

So, any further reference to a T5 MUST refer to a potential e6500 cored product.

Further, all mention of anything related to future e6500 developments has disappeared except for references to a version with virtualized interrupts (and that is not shown on some of the latest maps).

So, at this point, a board that could use either the T1042 or the T2081 (e5500 and e6500 cored respectively) would be what could be our last PPC move.

Then we may have to consider X64 or ARM.

As a MorphOS user, I already know my developers are considering the former.

As an open question to Amiga OS4 users , what are YOU going to do?

@iggy: thanks for the information.

I think that this should make people ponder about it, and accept your suggestion.

@iggy

Quote:

As an open question to Amiga OS4 users , what are YOU going to do?

Pray that those who can decide will recover from this useless PPC fixation, which should've happened years ago already.

_________________
Proud Amigan since 1987
Owner of various Commodore and a SAM440ep\OS4.1FE
See them on http://retro.furinkan.org/

@Nibunnoichi

Quote:

Pray that those who can decide will recover from this useless PPC fixation, which should've happened years ago already.

I don't disagree with this entirely. I think the X5000 will be PPC's last hurrah as a desktop system, and probably a pretty fine one too, if one can afford it. But there are probably a few more years in the architecture via emulation on other systems.

I read on the FS-UAE blog that 2.6.0beta1 is now able to run AmigaOS 4.x via emulation. Has anyone tried this, and if so, how fast is it?

http://fs-uae.net/blog

Last edited by bison on 02-Sep-2015 at 06:37 PM.

_________________
"Unix is supposed to fix that." -- Jay Miner

@bison

I'm pretty sure there will be PPC based hardware introduced after the X5000.
And Freescale will keep e6500 cored products available until at least 2020.
But, yes, eventually we will probably have to move on.

However, I still intend to buy an new PPC based system.
If we move to X64, we just become an alternative to Windows, OSX, and Linux.

Just as an investment in where we are now, the X5000 is very attractive.

@bison

Quote:

I read on the FS-UAE blog that 2.6.0beta1 is now able to run AmigaOS 4.x via emulation. Has anyone tried this, and if so, how fast is it?

I use WinUAE, but both emulators share relevant code. 500-1000 MHz G3 in integer operations (depending on application), 100 MHz G3 in FPU operations (on Core i5-2500 3.3 GHz). Other limitations: 128 MB RAM, 4 MB VRAM, 10 Mbit LAN, no Compositing, no 3D, no USB etc.

@iggy

Quote:

If we move to X64, we just become an alternative to Windows, OSX, and Linux.

Isnt that a good thing though?
Sticking with ppc makes amiga-oids *not* an alternative for most people.
Going x86/x86-64 doesnt stop amiga-oid fans using the system (unless they never were fans in the 1st place).
Also, ppc makes the ppc an alternative to linux, bsd, older osx. Lots of people use Linux already in addition to Amiga OS/MOS on their ppc hardware.
Going ARM makes it an alternative to android, linux, windows 10, etc.

That arguement always seemed a cop out to me.
Using x86 never stopped me loving amithlon. Have had dedicated boxes for years. Ditto AROS.

Unless a person has no faith in their OS of choice that arguement holds no weight.
Some people like ppc and thats fine, but worrying that you'll be an alternative to another system is, quite frankly, ludicrous.

@fishy_fis

Quote:

Some people like ppc and thats fine, but worrying that you'll be an alternative to another system is, quite frankly, ludicrous.

Well said.

The biggest reason why the Amiga is marginalized is because it is tied to special-purpose, overpriced hardware.

It never ceases to amaze me when people claim that software development is prohibitively expensive. Yet instead of investing resources in making the AmigaOS platform agnostic or reinventing it entirely, money is always devoted to perpetuating the status quo.

The PPC isn't a viable option in the near term and it won't even be an option in time. We should have started preparing for this long ago.

@iggy

Quote:

If we move to X64, we just become an alternative to Windows, OSX, and Linux.

And how is that a bad thing? Or let me ask it another way; How is that worse than what we have now?

_________________
All the way, with 68k

@iggy

I did most of the work of making schematic and layout symbols for T1 (with T2081 support built in, as it is basically a reduced feature set except for some new power pins) yesterday. Had already split out pin listings from datasheet and run them through my awk scripts to make eagle scripts, now I have all the symbol pins (allthough most symbols are still just a row of pins, as sorting them in sensible functions groups is a tedious but simple job and not necessary until the symbols actually get put in a schematic) and the complete device with connections from pin to physical pad. DDR controller is 100% done, as I usually like to start with that.

So at least one percent done towards my own T1/T2 design

I agree with most of the assessments here, T2 makes the most sense from a core perspective, but T1 for entry level really is very good.

T1024 would be a good idea if it wasn't a dead end, design wise. With T1020 and not using QE and DIU, you have an altivec in the upgrade path.

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@bison

Quote:

bison wrote: @Nibunnoichi Quote: Pray that those who can decide will recover from this useless PPC fixation, which should've happened years ago already. I don't disagree with this entirely. I think the X5000 will be PPC's last hurrah as a desktop system, and probably a pretty fine one too, if one can afford it. But there are probably a few more years in the architecture via emulation on other systems.

I have a SAM and i really enjoy it, but the problem is that when it will eventually break i'm not so sure i could/want to afford a >1000€ computer.
On the other hand i like AmigaOS4 and i don't want to switch to an alternative (AROS is my plan-b) or emulate it, so it's a bit dark out there...

_________________
Proud Amigan since 1987
Owner of various Commodore and a SAM440ep\OS4.1FE
See them on http://retro.furinkan.org/

@iggy

Quote:

iggy wrote: @bison I'm pretty sure there will be PPC based hardware introduced after the X5000. And Freescale will keep e6500 cored products available until at least 2020. But, yes, eventually we will probably have to move on. If we move to X64, we just become an alternative to Windows, OSX, and Linux.

2020 is less than 5 years away, if you really want/need to move on you need to start now, not one minute before 2020.
The "windows/osx/linux alternative" is a story told us since years and i've always found it to be a pretty ridiculous one, 'cause if i want to run one of those i already can do it regardless what Amiga i have while, on the other hand, if i want the Amiga i'll run it regardless of the ability to have another OS on the same hardware. It's a totally moot point.

_________________
Proud Amigan since 1987
Owner of various Commodore and a SAM440ep\OS4.1FE
See them on http://retro.furinkan.org/