>>> they [...] got a minimal update in 2013

>> The update was in 2016 (#911) after NXP took over

> these seem to be the same old 2013 cores in just a different package.

Yes, these are with the original e6500 cores. I've seen no technical details as to what NXP did specifically to lower power and increase memory performance of those SoCs at the same time.
I see now that the 2013 update you referred to may be the e6501 core (#493) to be used in the QorIQ Qonverge B3421 (#515). However, I can find no evidence that this SoC actually hit the market. Thus, I conclude that there's no e6501 out in the wild either.

> PCIe will be limited more by the lack of available lanes
> then version of these lanes.

PCIe throughput quadrupled from v1 to v3, so with v3 you only need a quarter of the lanes for the same total throughput.

> spending 1000€ (being optimistic here) to save 20€
> does not sound like a good plan.

20€ per month is 4 years until redemption ;-)

Quote:

I understand no single cores are produced anymore. Even SAM class CPU has dual core version.
Points are
- All boards except buggy Terrons and SAM440 were not fully supported by OS4
- x1000 was worse case scenario including horrid and buggy CFE and OS 4.2 license
- OS 4.1 FE, Enhancer, hanses Driver were all partial solutions that mended some problems
but forced gfx card change (instead of supporting what users had, which is also decent 3D)
- Each AmiWest kept repeating similar promises, with rare real world solutions
- Only SAM460 and x1000 (and Peg2) got dual boot support to boost exclusivity and sales,
even all boards can have it, with maybe some driver sharing not to force users to have two gfx cards for this
- Killing Acubes Moana port for MacMini is missed opportunity to spread OS4 licensing
and have widely avail hardware
- Only really supported cards on both systems with Warp3D are Radeons 9200-9600 and maybe 3dfx

Its really such approach plus rarely avail and insanely expensive hardware that kills OS4,
not MOS or AROS (that just have better approach)

My court claim is simple users answer to such nonsense, nothing more or less.

More care for users and single board higher run cheaper end price approach, similar to Terron
but non buggy could save the day way better,

If I can say something good of PA Semi, it is power efficiency, nice mem and FPU score
but horrid CPU performance per Mhz, closer to G3. It was obviously ment for laptops, not desktops.

Also, nice FPGA that could fit ECS AGA chipset emulation instead of Xorro Xena could significantly
increase backward compatibility.




[ Edited by vox 09.12.2024 - 12:00 ]

Quote:

Sure if you can afford to route 4 of the limited PCIe lanes to the GPU and if that GPU is PCIe3 or better you will be right back where you started.

As for efficiency, affordability and availability choosing a random non PPC board and hardcoding a PPC EMU to it would a less bad idea.

>>> full Hyperion name, as well as Amiga EON is a bit mythical

>> It's A-EON officially, not "Amiga EON".

> Its obviously Amiga for EONs

No matter what "A-EON" stands for, using "Hype" for correct "Hyperion" is completely different from using correct "A-EON" for "Amiga EON". That's basic logic.

> OK SSolie is still setuping up the newer machine

No, it was said he tried to set up an *older* system, because newer compilers cannot compile old CFE source code.

>>> Timberwolf […] update to FF16 promised

>> No, neither FF16 nor promised (#10).

> Friedens brothers used to have small website with Timberwolf page [...].
> It stated they will update the engine.

Maybe there's a cultural difference regarding what constitutes a promise? The statement was:

"Currently, the code is based on Firefox 4.0.1, but we’re working on bringing it up to the latest build (18.0.1 at the time of writing, although 19.x is more likely)."

That's not a promise to my mind, but merely a description of what they were attempting to do. They failed at this attempt, but this doesn't mean they broke any promise.

> why there isnt some 64 bit G5 optimised Linux?

There are several. I gave you a list of 7 distributions there:
https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=13894&forum=11&start=10 (#11)

> before him best gfx card for OS4 was Radeon 9600.

OS4 never 3D-supported Radeon R300 GPU series. Best in 3D was the 9000 Pro of R200 GPU series as even the 8500/9100 (fastest of R200 series) caused artifacts on the screen.

> no RX support making x1000 biggest router ever.

Using a GCN1-based card instead of a GCN4-based one makes it a router? Mind you, MorphOS doesn't even have GCN1 support.

>> Whatever this should prove, I don't get it.

> that MOS is older and longer has OS3 native ports, there was
> momentum of about few years around x1000 release, OS4 development
> gained tracktion, but soon died down to few people.

Fair enough, but I was referring to your Aminet and OS4Depot number mumbo jumbo.

>>> MOS […] supports WarpUp […] and Warp3D.

>> …as does OS4.

> OS3 RTG apps dont work well, WarpOS PowerUP rarely

PowerUP probably won't work, but originally there was WarpOSEmu which only works on G2/G3/G4 and Sam440, and meanwhile there's ReWarp with ReWarp3DPPC which is said to work quite well on all OS4 hardware.

> and Warp3D only up to Radeon 9600.

There was Warp3D_SI for GCN1-based cards and MiniGL4GL4ES for GCN4-based cards, and meanwhile there's NovaBridge for both GCN1 and GCN4. (And again, no Warp3D for Radeon 9600 ever, see above.)

>> MorphOS already has original MUI5.

> Called MUI4

No, as I already told you 2 years ago, it's been called MUI5 since MorphOS 3.8 in 2015:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=11372&start=335 (#336, #337)

> its very unclear how MUI5 came to be.

The MorphOS/MUI developers decided in 2015 to bump the MUI version along bug fixes and/or feature improvements for MorphOS 3.8. Simple.

> Amiga Inc and Hype had to go back to OS 3.1 because OS 3.9
> sources were either lost or required buy out from H&P.

This didn't mean much practically, as all that was H&P's in 3.5/3.9 was some GUIs (Prefs...). Everything else was either already in 3.1, was owned by then-OS4 individual developers anyway or could be re-licensed from the respective developers who often released their components also on Aminet.

> both OS4 and OS 3.1.4 and OS 3.2 started from OS 3.1, not fully
> leveling up to OS 3.9 in some minor aspects up to this day.

You say yourself that it's only minor, but what's really missing functionally in OS4 or OS3.2 from OS3.9? And of course, 3.2 didn't start from 3.1 but from 3.1.4 ;-)

> All boards except buggy Terrons and SAM440 were not fully supported by OS4

What's missing from Sam460 support?

> Only SAM460 and x1000 (and Peg2) got dual boot support

X1000? You mean X5000? And who is to blame for MorphOS Team not supporting the other boards/systems?

> with maybe some driver sharing not to force users
> to have two gfx cards for this

It's not as easy as "driver sharing". MorphOS would have to adopt the whole OS4 2D/3D graphics subsystem for this. See also bigfoot's comment.

> Only really supported cards on both systems with Warp3D
> are Radeons 9200-9600 and maybe 3dfx

With original Warp3D on OS4, Radeon supported was only from 7000 to 9250 (but not 8500/9100), with 9000 Pro being the fastest.

> PA Semi [...] was obviously ment for laptops, not desktops.

Indeed, the PA6T was originally conceived to be offered to Apple for use in a "G5" PowerBook. Apple's switch to Intel was the reason for development direction adjustment (especially regarding the on-chip peripherals) towards embedded applications.

[quote]esc wrote:
Playing a bit of catch-up here, but...from where I sit, it looks like MagicSN likes porting games to OS4, and it looks like Vox is giving grief because Vox does not like Hyperion.

Is that literally the point of this whole thread? Or do I have it wrong?.[/quote]

It started like that, but is oversimplification. MSN basically claims OS4 is more viable market then MOS


>> It's A-EON officially, not "Amiga EON".

> Its obviously Amiga for EONs

No matter what "A-EON" stands for, using "Hype" for correct "Hyperion" is completely different from using correct "A-EON" for "Amiga EON". That's basic logic.

VOX>Both are irony and sarcasm


No, it was said he tried to set up an *older* system, because newer compilers cannot compile old CFE source code.

VOX> Anyhow no results.

>>> Timberwolf […] update to FF16 promised


"Currently, the code is based on Firefox 4.0.1, but we’re working on bringing it up to the latest build (18.0.1 at the time of writing, although 19.x is more likely)."

That's not a promise to my mind, but merely a description of what they were attempting to do. They failed at this attempt, but this doesn't mean they broke any promise.

VOX> For first and major and only FF implementation it looked like giant leap forward.
Since it remained FF4 and nothing better was source coded, lets call it effective lie and settle.

> why there isnt some 64 bit G5 optimised Linux?

There are several. I gave you a list of 7 distributions there:
https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=13894&forum=11&start=10 (#11)

VOX>Great thanks, all G5 users should use em

OS4 never 3D-supported Radeon R300 GPU series. Best in 3D was the 9000 Pro of R200 GPU series as even the 8500/9100 (fastest of R200 series) caused artifacts on the screen.

VOX>Which just shows how sad was state of gfx before Hans.
VOX>In my case, supplied OS 4.1u5 x1000 could not boot x1000 since it even did not have 2D driver.

> no RX support making x1000 biggest router ever.

Using a GCN1-based card instead of a GCN4-based one makes it a router? Mind you, MorphOS doesn't even have GCN1 support.

VOX>Both are irony and sarcasmCFE outside x1000 is only used to set up routers,


> that MOS is older and longer has OS3 native ports, there was
> momentum of about few years around x1000 release, OS4 development
> gained tracktion, but soon died down to few people.

VOX> Fair enough, but I was referring to your Aminet and OS4Depot number mumbo jumbo.

To be honest, I expected there is way more MOS sw, but have been proven its about same.

>>> MOS […] supports WarpUp […] and Warp3D.

>> …as does OS4.

VOX>> OS3 RTG apps dont work well, WarpOS PowerUP rarely

PowerUP probably won't work, but originally there was WarpOSEmu which only works on G2/G3/G4 and Sam440, and meanwhile there's ReWarp with ReWarp3DPPC which is said to work quite well on all OS4 hardware.

VOX> With software renders, CIA emulations, add on of MUI, AHI and many other tricks situation can a bit VOX>improve, but its still awful. Rarely OS3 clean WB apps work. Biggest step forward was JIT UAE OS4 VOX> that had x1000 optimised binary. But yet even with it SAM460 could do ECS games and x1000 and x5000 VOX>AGA including productivity sw. With RunInUAE combined, it looked for a moment like advanced Amiga.

> and Warp3D only up to Radeon 9600.

(And again, no Warp3D for Radeon 9600 ever, see above.

VOX> EVen sadder then I throught, Acube was right putting 9000,9100,9200 level card onboard.

>> MorphOS already has original MUI5.

> Called MUI4

VOX> Thats my point too. Very moot bussiness having nonexisting MUI5.


> its very unclear how MUI5 came to be.

The MorphOS/MUI developers decided in 2015 to bump the MUI version along bug fixes and/or feature improvements for MorphOS 3.8. Simple.

VOX> I mean OS4 version, and derived OS3 version, even its good. Since MUI 3.8 registration was VOX>impossible, I was happy to finally customize MUI apps, since I do love them, and MagicWB to VOX>Reaction. Only good newer Reaction App I have seen was EMOTION


This didn't mean much practically, as all that was H&P's in 3.5/3.9 was some GUIs (Prefs...). Everything else was either already in 3.1, was owned by then-OS4 individual developers anyway or could be re-licensed from the respective developers who often released their components also on Aminet.

VOX> From user point it includes ClassAct, MUI, WarpOS, AHI etc. These were external extras, but so they were VOX>when supplied as extras with OS 4.0/OS 4.1FE and a necessity for modern AMiga user.

You say yourself that it's only minor, but what's really missing functionally in OS4 or OS3.2 from >OS3.9? And of course, 3.2 didn't start from 3.1 but from 3.1.4[/quote]

VOX>Starting with video player (no matter how bad, functional), browser, improved CD player
VOX>As user I do count everything AsyncWB, good IP stack, more datatypes etc. When you include 2 VOX>officialand some unofficial Boing Bags its quite a lot
VOX>https://amigan.1emu.net/releases/BoingBags3&4.readme

VOX>Surely large drive support, file systems and critical under hood improvements and some OS 4 VOX>VOX>backporting is significant milestone, I dont underestimate effort put to it. I just say as end user I understand why some users stil stick to OS 3.9 patched,

VOX>Plus OS 3.2 existed according to WB Nostalgia :DDDD
https://www.gregdonner.org/workbench/wb_b32_40.html

> All boards except buggy Terrons and SAM440 were not fully supported by OS4

What's missing from Sam460 support?

VOX> Small FPGA support and there is some gfx card limit linked to Uboot

> Only SAM460 and x5000 (and Peg2) got dual boot support

X1000? You mean X5000? And who is to blame for MorphOS Team not supporting the other boards/systems?

VOX> My bad x5000, I wish it was x1000. I mean this as Trevor presenting as person that supports all systems, I dont blade MOS team.

> with maybe some driver sharing not to force users
> to have two gfx cards for this

VOX>It's not as easy as "driver sharing". MorphOS would have to adopt the whole OS4 2D/3D graphics VOX>subsystem for this.

Yes, I ment also needed libs etc. sharing if we are to really end "war of all wars" some unity would be neeed, and sharing. Even VIA, AMD and Intel and OS2 and Windows did it.

> Only really supported cards on both systems with Warp3D
> are Radeons 9200-9600 and maybe 3dfx

With original Warp3D on OS4, Radeon supported was only from 7000 to 9250 (but not 8500/9100), with 9000 Pro being the fastest.

VOX> Yup its even worse then I troughts, where 9100 is best model, but rarest.

> PA Semi [...] was obviously ment for laptops, not desktops.

Indeed, the PA6T was originally conceived to be offered to Apple for use in a "G5" PowerBook. Apple's switch to Intel was the reason for development direction adjustment (especially regarding the on-chip peripherals) towards embedded applications.

VOX> I believe Apple betrayed PPC alliance because judget investment is too costly, its easier to use VOX> offshelf and profit more plus aquire PA Semi team for iPhone use. Good move.
VOX> But at least they had decency to put some good m68k emu layers in early MacOSX PPC, universal
VOX> binaries and PPC emulation in early Intel MacOSX and now x86 emu on ARM.

[ Edited by vox 09.12.2024 - 18:10 ]

> CFE outside x1000 is only used to set up routers

...as well as Smart TVs and Blu-ray players, and -- which is the main reason for its use on X1000/Nemo -- it was used in the PA6T evaluation boards Electra and Chitra. Btw, you and me already had this discussion 7½ years ago:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=11137&start=685 (#686, #689)

>>> OS3 RTG apps dont work well, WarpOS PowerUP rarely

>> PowerUP probably won't work, but originally there was WarpOSEmu
>> which only works on G2/G3/G4 and Sam440, and meanwhile there's ReWarp
>> with ReWarp3DPPC which is said to work quite well on all OS4 hardware.

> [...] CIA emulations [...] JIT UAE [...] ECS games [...] AGA [...] RunInUAE

I was clearly discussing your assessment of OS4's PowerUP/WarpOS software compatibility here. Not sure why you suddenly change this to Amiga chipset emulation topic.

> Acube was right putting 9000,9100,9200 level card onboard.

Radeon M9 on Sam440ep equals Radeon 9000. 9100 wouldn't have worked well with OS4.

>>>> MorphOS already has original MUI5.

>>> Called MUI4

>> No, as I already told you 2 years ago, it's been called MUI5
>> since MorphOS 3.8 in 2015:
>> https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=11372&start=335 (#336, #337)

> Thats my point too. Very moot bussiness having nonexisting MUI5.

For the umpteenth time: Official MUI5 has been existing since MorphOS 3.8 in 2015.

>>> its very unclear how MUI5 came to be.

>> The MorphOS/MUI developers decided in 2015 to bump the MUI version
>> along bug fixes and/or feature improvements for MorphOS 3.8. Simple.

> I mean OS4 version, and derived OS3 version

I also explained this 2 years ago (see link above).

>> all that was H&P's in 3.5/3.9 was some GUIs (Prefs...). Everything
>> else was either already in 3.1, was owned by then-OS4 individual
>> developers anyway or could be re-licensed from the respective
>> developers who often released their components also on Aminet.

> it includes ClassAct, MUI, WarpOS, AHI etc. These were external extras,
> but so they were when supplied as extras with OS 4.0/OS 4.1FE

I doubt that MUI was in 3.5/3.9 and that ReAction (ClassAct) and AHI were extras with OS4. And your list of components that could also be obtained from Aminet etc. during 3.5/3.9 times proves my point that non-access to 3.5/3.9 was a non-issue.

> OS 3.2 existed according to WB Nostalgia :DDDD
> https://www.gregdonner.org/workbench/wb_b32_40.html

Yes, I already pointed to this some years ago.

>>> All boards except buggy Terrons and SAM440 were not fully supported by OS4

>> What's missing from Sam460 support?

> Small FPGA support

Sam440ep and Sam440ep-flex also have an FPGA. Why do you call them fully supported then?

> and there is some gfx card limit linked to Uboot

This has nothing to do with (non-)support by OS4.

>> who is to blame for MorphOS Team not supporting the other boards/systems?

> I mean this as Trevor presenting as person that supports
> all systems, I dont blade MOS team.

So you blame Trevor for MorphOS not supporting X1000/Nemo? The MorphOS Team once stated Ben Hermans being part of the Belgian company A-Eon CVBA as reason for non-support. Somehow, Hermans' departure and the closure of the company didn't make them reconsider.

>> MorphOS would have to adopt the whole
>> OS4 2D/3D graphics subsystem for this.

> Yes, I ment also needed libs etc. sharing [...].
> Even VIA, AMD and Intel and OS2 and Windows did it.

That's because these companies have provided the OS2 and Windows drivers for their hardware components themselves. You can't compare this with MorphOS and OS4.

> 9100 is best model, but rarest.

Best and rarest of the R200 series was actually the full 8500. 9100 was equal to trimmed 8500 LE.

> I believe Apple betrayed PPC alliance

AIM alliance had been defunct for ages in 2005.

> plus aquire PA Semi team for iPhone use. Good move.

PA Semi acquisition was 3 years after Apple's switch announcement. I think that's more like two distinct moves. Some people tend to see a connection here, but I think there wasn't.

> they [...] put some good m68k emu layers in early MacOSX PPC

It lasted almost to the end. The Classic Environment was still there in 10.4. The last PPC version (10.5) was the first PPC version without it.

@Vox I mean, neither are really all that viable. If you're porting games here with the intent to make a lot of money (vs just doing it for the hobby you enjoy and hopefully earning a few bucks) you're doing it wrong. I guess what I'm really trying to say about the whole MOS vs AOS4 thing is...who cares? I'm glad people take it upon themselves to write software for any of these platforms and support all of them when I can. I'm not here to split hairs.

One in all, yes, we discussed it over the years, to keep things
summerized many choices with OS4 has gone from bad to worse, making it harder for users
to make updated and usable OS and requiring more and more 3rd party purchases.

Unlike that, MOS offers good experience out of box and quite extensive OS support.

Add to that unavail and overpriced hardware, its simply destined to be doomed.

There were chances to make common ground on hardware supported by both camps,
but that never came standardized (i.e. offering MOS licensed and bundled)

Also, healthy trade of would be sources to port Iris and Wayfarer for gfx subys and RadeonHD drivers,

Then selling e.g. SAM460x and x5000 with user choice of one of both OSs would be fair and benefitial.
That is unless we can really unite all efforts towards one single usable OS, which is still a pipe dream.

[ Edited by vox 10.12.2024 - 18:44 ]