Option to hide useless offtopic threads like this would be cool.

>>> I dont see [...] OS4 library growing.

>> A look at os4depot.net indicates otherwise.

> in my exp, its been down since at least 2016.

So you mean it's not growing as much as you have liked, or growing less than it did at some point in the past. Both are different statements to the one I replied to.

> Above performance will be included.

Unlike BigGun, please don't "forget" to add the note that bustest benchmark on PPC systems runs in emulation.

> I am speaking of UAE OS4 [...] not Petunia.

Oh, I misunderstood then. But that JIT is in no way OS4-specific as it is used on UAE for other PPC OS as well, like MorphOS. And your "PA Semi Altivec" mention doesn't make sense here either, I think.

Quote:

>> A look at os4depot.net indicates otherwise.

> in my exp, its been down since at least 2016.

So you mean it's not growing as much as you have liked, or growing less than it did at some point in the past. Both are different statements to the one I replied to.


I believe it can be easily proven by using its stats (no of file uploads, downloads etc.)
but I have been watching progress for few years, I tell you no lie that development beside enhancer has been down since 2016, both OS and community production like OS4 depot, HunnoPPC, whoever.

> Above performance will be included.

Unlike BigGun, please don't "forget" to add the note that bustest benchmark on PPC systems runs in emulation.

Surely, you will even be credited.

> I am speaking of UAE OS4 [...] not Petunia.

Oh, I misunderstood then. But that JIT is in no way OS4-specific as it is used on UAE for other PPC OS as well, like MorphOS. And your "PA Semi Altivec" mention doesn't make sense here either, I think.

You really missed UAE JIT Warp speed episode?
http://euaejit.blogspot.com/2014/03/ppcjitbeta02-beginning-of-beautiful.html

In 2015 it resulted in 1.0 fast UAE JIT for OS4,MOS and MacOSX, get it!
https://sourceforge.net/projects/euaeppcjit/files/1.0.0-Stable/

OS4 edit since early beta has uae_X1000 file that is PA Semi Altivec optimised.

Its one of few OS4 pieces of code that ever did so, so much of OS4 dream machine.
It gives better results then generic one. Am I clear now?

OK, I am gone to FAQ, and gents and ladies FAQ you!





[ Edited by vox 06.01.2020 - 18:49 ]

>>>>> I dont see [...] OS4 library growing.

>>>> A look at os4depot.net indicates otherwise.

>>> in my exp, its been down since at least 2016.

>> So you mean it's not growing as much as you have liked, or growing less than it
>> did at some point in the past. Both are different statements to the one I replied to.

> I believe it can be easily proven by using its stats

No, you cannot prove by any means that the OS4 software library isn't growing. What you can maybe prove, as I said, is that it's growing less than it did at some point in the past, if anything. Every new upload to os4depot.net grows the OS4 software library.

>>> I am speaking of UAE OS4

>> that JIT is in no way OS4-specific as it is used on UAE
>> for other PPC OS as well, like MorphOS.

> You really missed UAE JIT Warp speed episode?
> http://euaejit.blogspot.com/2014/03/ppcjitbeta02-beginning-of-beautiful.html

Huh? How do you conclude from what I wrote that I don't know this? What I wrote is that this UAE JIT is not OS4-specific, like you just regurgitated.
Have a look who reported your link here:
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=8762&start=63

> In 2015 it resulted in 1.0 fast UAE JIT for OS4,MOS and MacOSX, get it!
> https://sourceforge.net/projects/euaeppcjit/files/1.0.0-Stable/

Better MorphOS versions to be found there:
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=8762&start=66 (Edit4)
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=12&topic_id=12242
https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=12242&forum=12&start=6

>> your "PA Semi Altivec" mention doesn't make sense here either, I think.

> OS4 edit since early beta has uae_X1000 file that is PA Semi Altivec optimised.

No, it's not. You just made this up. There is no AltiVec code in this build, and it wouldn't make sense to withheld AltiVec optimizations from users of OS4 on G4 machines, don't you think? What this PA6T-specific build (and those for G5/PPC970) has is a workaround for a missing instruction (mcrxr) on these CPUs that otherwise would have to be emulated and thus be slower, which has nothing to do with AltiVec. More info:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=8762&start=58
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=8762&start=61

> Am I clear now?

To me you sound as confused as always ;-)

Quote:

As usually Your lack of knowledge makes your opinions wrong.
instruction decoder is only part of cpu.

Quote:

It is history.
It is memories.

68k made by Motorola is original vampire is replica.
AGA made by Commodore is original vampire is replica.

vampire do not share history with Commodore and Motorola, everything was done from scratch.

>> the difference is on the gate level, i.e. on the actual
>> microarchitectural level

> instruction decoder is only part of cpu.

The instruction decoder must be different because the instructions are differently encoded in microcoded microarchitecture vs. hardwired microarchitecture on the gate level.

>> It's just an ISA, nothing more. Anything else is religion.

> It is history.

History is not changed by 68080/Apollo.

> It is memories.

Memories are not changed by 68080/Apollo. You can even revive those memories with 68080/Apollo just by ignoring its ISA extensions.

> vampire do not share history with Commodore and Motorola

Indeed, it continues the history by reimplementing the m68k ISA, like was done before.