Anyway A-Eon now own DvPlayer and Tunenet. Two programs that would be affected by the FPU emulation speed. So special SPE floating point versions of these should not be a problem.
Most other type of programs should not be affected based on the Linux exprience .

Last edited by Spectre660 on 19-Oct-2015 at 11:28 AM.

_________________
Sam460ex : Radeon Rx550 Single slot Video Card : SIL3112 SATA card

@Spectre660
But games and computers/consoles emulation could be have several problems with this SPE unit.

_________________

@OlafS25

How can something be slow, yet not tested

@Spectre660

Quote:

special SPE floating point versions...

So we can expect some specialized binaries?

@eder

in software you can also test parts of it, f.e. you can define test cases that run against routines to see if there are problems. As I understand it not yet integrated in AmigaOS because AmigaOS is still not running on the hardware.

Last edited by OlafS25 on 19-Oct-2015 at 12:51 PM.

@eder

they have developed the method that they expect to work, yet its expected to be slow. however the developer, or rather managing director, if thats his correct title, hopes it will be possible to accelerate it once the osruns and the patch is actually implemented.

you can read his statement olaf linked to. its quite contrary to "there us no problem os4 already works on the board in question" in my opinion.

Last edited by wawa on 19-Oct-2015 at 12:51 PM.

@wawa

So, if/ when this is sorted out and its on par with Sam 460 at 500 £. This is going to be a bargain in OS4 terms ?

@wawa

The "no problem os4 already woks on the board" did not come from Hyperion Management or AmigaOS developers .

It may have been assumed by those saying this because AmigaOS developers have had boards for several months.

Quote:

wawa wrote: @eder they have developed the method that they expect to work, yet its expected to be slow. however the developer, or rather managing director, if thats his correct title, hopes it will be possible to accelerate it once the osruns and the patch is actually implemented. you can read his statement olaf linked to. its quite contrary to "there us no problem os4 already works on the board in question" in my opinion.

_________________
Sam460ex : Radeon Rx550 Single slot Video Card : SIL3112 SATA card

@iggy

Speculation on my part .

Quote:

iggy wrote: @Spectre660 Quote: special SPE floating point versions... So we can expect some specialized binaries?

_________________
Sam460ex : Radeon Rx550 Single slot Video Card : SIL3112 SATA card

@Spectre660

Quote:

Spectre660 wrote: @iggy Speculation on my part . Quote: iggy wrote: @Spectre660 [quote]special SPE floating point versions... So we can expect some specialized binaries?

According to HyperionMP your speculation is wrong

Quote:

HyperionMP: So no need to either recompile the OS (the kernel is obviously always specific for a certain CPU) or any applications.

Last edited by Zylesea on 19-Oct-2015 at 02:15 PM.

_________________
My programs: via.bckrs.de
MorphOS user since V0.4 (2001)

@Zylesea

'No need to' is not the same as 'cannot benefit from'.

There is no need to compile altivec binaries for example.

_________________
BroadBlues On Blues BroadBlues On Amiga Walker Broad

@Spectre660

Quote:

Speculation on my part .

Oh, I appreciate your candor.
In some cases it would be preferable.
And using the non-standard fpu incorporated into the e500v2 ought to be a better solution for software that needs an fpu than emulating the standard unit.

At this point, I'm reserving further judgement.
A-eon invested the funds and the time creating this board, its going to happen.
So, if its price is right and its performance is a least on par with the Samantha, then it was a good investment.

And nothing precludes A-eon from further introductions.
After all, the e5500 core I would appear to be so fond of is a dead end.
So while I would have preferred a T10XX solution, neither that core or the e500 will be developed further.
Only the e6500 core is shown in current roadmaps of future introductions.

We have the X5000, and hopefully one day we may even see an e6500 cored cpu in its successor.

In any case, PPCs ought to be viable for at least 5-10 years, and I'm extremely fond of this ISA...so, OK.

@Spectre660

Quote:

Spectre660 wrote: @wawa The "no problem os4 already woks on the board" did not come from Hyperion Management or AmigaOS developers . It may have been assumed by those saying this because AmigaOS developers have had boards for several months.

right (post #466 here):
Quote:

Emulation of FPU instructions will be handled transparently without significant impact on performance. So no need to either recompile the OS (the kernel is obviously always specific for a certain CPU) or any applications.

you dont even need to speak it out yourself. people will read from your lips, whatever they wish for, and fill in the blanks as facts. and then all is fine again (post #468):

Quote:

Thread over :D

;)

Last edited by wawa on 19-Oct-2015 at 04:17 PM.

@OnlyMe

Welcome!

@Spectre660

Quote:

Spectre660 wrote: Anyway A-Eon now own DvPlayer and Tunenet. Two programs that would be affected by the FPU emulation speed. So special SPE floating point versions of these should not be a problem.

In this case you need that the kernel supports the 64-bit wide registers of the e500v2, which requires some work.

This also slows down the task switching and the interrupts handling (at least), in general (so not only for the "SPE application").

Quote:

Most other type of programs should not be affected based on the Linux exprience .

What's your Linux experience? Have you used Debian PowerPC (so, NOT the version with SPE unit support)?


@Seiya

Quote:

Seiya wrote: @Spectre660 But games and computers/consoles emulation could be have several problems with this SPE unit.

Only slowdowns, but they can run without any other problem, since the FPU (and "string" instructions) emulation is totally transparent to the existing software.

I have asked ACube on their Facebook page about the Sam460CR and they told me they have no intention of stopping its production.

Are we sure ACube are involved with A1222?

_________________

@Everblue

they might not have an "intention" but if they ocassionally do a 50 units run and the other board will be 10 times more in one run as apparently announced, what will effectively be the outcome? any thoughts? ;)

@cdimauro

Quote:

What's your Linux experience? Have you used Debian PowerPC (so, NOT the version with SPE unit support)?

Belive me Spectre660 have more experience than you with linux PPC and PowerPc ...

_________________
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

@Everblue

Why acube have to stop a production to a motherboard for sure less expensive compared the Tabor?
440 will exit from the scene totally and 460cr will be the entry machine

_________________
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

@tlosm

Quote:

tlosm wrote: @cdimauro Quote: What's your Linux experience? Have you used Debian PowerPC (so, NOT the version with SPE unit support)? Belive me Spectre660 have more experience than you with linux PPC and PowerPc ...

Of course: I've ZERO experience with both.

BTW, there's no need to have experience to discuss about some technical stuff. The FPU (and string instructions) emulation drops down the performance, and it's quite obvious if you know how it works.

The only way to avoid it is to have ad-hoc binaries that natively use the SPE unit (but this require some changes to the o.s., which slows down a bit the execution for all applications).

For 68K FPU emulation it isn't a big problem, because the JIT can be reworked to directly generate SPE instructions, but this requires a consistent change of Pentunia, since it's also written in assembly.

Anyway, I've asked for some information.