Post your AmiWest 2016 questions here

[ssolie]

I'm still at the AmiWest 2016 show today (Sunday) and some of you might have some questions.

Please post your questions here and I'll try to get back to you.

Steven Solie
AmigaOS Development Team Lead

[ssolie]

...is the X5000 is officially released?

Yes

And, when will the production of P5040 based motherboards commence?

Better ask A-EON that question.

[ssolie]

What were the books in a picture posted by Robert? A forthcoming print copy of updated RKM set?

Eldee printed those from the wiki as anyone can do right now. See MediaWiki documentation for further details.

[ssolie]

Hi Steve, I know you can't give timelines but please tell me that muti-core support for the X1000 is on the radar or at least the  'must do' list

Yes, it is on the radar. We need to ship the X5000 and A1222 first.

[ssolie]

I've the feeling that porting AmigaOS to new platform is slowing down the release of AmigaOS 4.2. If true will we see a speed up on other part of the OS (when X5000 and Tabor port is done)?

That is not possible since it is all one code base. We simple name it "4.2" when the time has come.

[ssolie]

In the past you talked about a new roadmap, is this roadmap done now?

There has always been an internal road map but the more "external" one is certainly not out yet or you would have seen it.

[ssolie]

Any news about an updated browser?

Well, that is outside the scope of AmigaOS and I don't have any news myself. Maybe A-EON is working on one?

[ssolie]

So an official statement would be nice.

You are wise to ask for direct clarification on things. Plus, news can change from day to day. Assuming you heard something or read something correctly is never a good idea.

[ssolie]

Great to know that there is available product heading into the holiday season.

Yes, it has been a long haul for everyone involved but it is close. Release candidates are being tested as I type.

@all
Any more questions on what I said or exhaled?

I plan to keep checking a.org and I wont be posting on any other public forums due to time constraints.

[ssolie]

Is this something you take into account while developing AmigaOS for it to be faster? It also state that Clang is better than VS, are you considering using Clang over GCC?

We don't really need those types of performance gains until we take care of the low hanging fruit first.

Much larger optimization gains can be had by utilizing A-EON's hardware to the fullest. There are several DMA channels sitting around doing nothing on the X1000 for example.

I am much more interested in order of magnitude optimizations.

[ssolie]

Will Alice breaks the 128 MB limit ? or will have a functionnal uaenative.library (allowing Wazp3D kind of ppc/x86 driver) ?

That project has nothing to do with Hyperion so I'll leave it for others to answer.

[ssolie]

Is there any update on the Hans-Jörg's new 3D API?  His last blog on the subject was all the way back in 2011.

Sorry but I have no update on the status of the AmigaOS Gallium port at this time.

Warp3DNova was born out of the frustration in waiting for the Gallium hardware drivers. It also means A-EON is not dependent on the efforts of those working on Gallium and they can support older cards as required. A-EON is focusing on content and sometimes that means giving the OS a boost to get that content on there.

I was playing AmiCraftNova at AmiWest on my Sam460ex. Whether that works via Nova or Gallium or some other engine I don't really care and I doubt end users will either.

[ssolie]

Hans-Jörg mentioned in his blog that the API (not Gallium) will also be used for hardware acceleration of the OS as well as applications.  If Nova turns out to be faster than Gallium it would be nice to able to drop in a more efficient driver for that purpose.

Using the GPU hardware is the point. It doesn't really matter which is "faster" because neither Gallium nor Nova are doing all the work. They are just a means to enable the GPU.

I think you said last year that the X5000 is using the X Kenel from day one, would I be right to assume the same of the A1222.

The so-called X-Kernel is just a label given to a source code branch. We just needed a way to tell the main trunk apart from the experimental multicore kernel. It will all become just the kernel again and users need not be concerned.

[ssolie]

I known it's on stand by status right now but have you guys found a good solution to multicore support ? I remember sometime ago (a year?) you were studying different solutions.

I actually don't know all the details myself but Thomas did figure something out and is implementing things. The implementation certainly didn't hit any road blocks yet.