Quote:

The NatAmi hardware developers have said that because of the extremely low memory latencies on the NatAmi design (and the slow memory controller on AmigaOnes), they expect performance of their board to be on a par with that of an AmigaOne board

What CPU are we talking about? I've heard the argument of low memory latency before, but that is only interesting as soon as you leave the on-chip level 2 cache. I would be very surprised if any 68k CPU can come close to a G4.

Quote:

So if the NatAmi team is correct, their system will be the fastest Amiga, that can be bought new, for the foreseeable future.

Even if that were true, it's still a 68k CPU.

Quote:

Given that, if hard benchmarks do confirm the performance advantage of NatAmi over PPC Amigas that can be bought new, do you think that Hyperion would reconsider it's "PPC only" stance?

Standard disclaimers applied, my guess would be "not in a million years". If there were any support for a different CPU than PowerPC at this point, it would be for x86. Regardless of what speed you think the NetAmi can achieve, I am quite sure that any cheap x86 board you can buy will run circles around it.

So before we would go through the hassle of porting to a completely abandoned processor architecture, we would go x86 if there were any plans for porting to a different CPU now; alternatively, MIPS or XScale would be a possibility because of their wide-spread use in embedded devices.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Quote:

The A1 will only beat it at pure CPU power on algorithms that fits in the cache.

Don't forget L2. You can construct benchmarks that bypass L2 performance because of data size, but in reality, data access is much more localized. Memory latency is only one piece of the puzzle.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Quote:

they expect performance of their board to be on a par with that of an AmigaOne board

While I don't know what CPU/frequency they have exactly (I think it is 90 MHz), the "low memory latency means it will outperform an 800 MHz CPU" is absolute bullocks. Low memory latency will NOT help you on CPU intensive tasks. And, sorry, any design will have a hard time outperforming the L2 cache on a 750GX, no matter how low the memory latency is.

Quote:

So if the NatAmi team is correct

They are not.

_________________
Thomas, the kernel guy

"I don't have a frigging clue. I'm norwegian" -- Ole-Egil

All opinions expressed are my own and do not necessarily represent those of Hyperion Entertainment

@Rogue @EntilZha

Quote:

What CPU are we talking about?

That CPU does not exists yet. The project is to (re-)build a new 68K like (and compatible) that will reside in a single chip (FPGA based) along with the Natami's SuperAGA. The early draft of the (what it's called for now) N68070 can be seen on the Natami web site.

From the FAQ:
Quote:

The goal of our new advanced 68K CPU is to provide a higher compatibility to 68000 games than the 68040 and 68060 CPUs ever did. Our goal for performance is to exceed the speed of an original 68040 clocked at 200Mhz for normal applications and to archive a performance of a 68040 with 600 Mhz for multimedia applications. [ ... ] The SuperAGA chipset will be 100 times faster than AmigaAGA. The CPU options that we are looking forward to use, will be about 10-20 times faster than the 68040 in the Amiga 4000. Its clear that the system will be able to run AmigaOS very swiftly. Our goal for the system is to have enough power to play games and do divX playback. [ ... ] The "Self-Made" 68K provides a lot of advantages: * Cost reduction * Possible to add new faster and cheaper DDR memory. * Higher clockrate of 150-500 MHz is possible. * Memory latencies will be reduced. * Powerful multimedia instructions could be used. [ ... ] A 100 MHz 68060 has a similar performance as a 100 MHz PPC.

Please, have a look on the Natami project.
IMHO, the Natami team have the ground and skill to build a really interesting machine.

(Edit: Oh, and BTW, I would love to own a Natami box running AmigaOS 4.x ! )

Last edited by AmigaPhil on 06-Aug-2008 at 09:47 PM.
Last edited by AmigaPhil on 06-Aug-2008 at 09:41 PM.

Just make backup first

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@Entilzha

They have talked about PPC cpu and Cell BE too. Dunno if there are any prototype of such machine.

BTW, PPC and Cell could be used only as accelerators in Natami. Is the code 68k oriented, isn't it?

_________________
retired

@ChrisH

Quote:

AmigaOS's Env: & EnvArc: are quite sufficient for most "registry" purposes. But it's use could certainly do with expanding, to handle tracking of what has been installed where,

Have a look at SYS:Prefs/Env-Archive/AppPaths, this is certainly used by AmiUpdate for the purposes you describe, and is probably used by other programs.

_________________
Peter Swallow.
A1XEG3-800 [IBM 750FX PowerPC], running OS4.1FE, using ac97 onboard sound.

"There are 10 types of people in the world: those who understand binary, and those who don't."

@Hans
Quote:

I count 8 graphics-related items in the list.

If it had to do with graphics and/or GUI I counted it.

Quote:

not all the graphics related stuff is just eye-candy

Better graphics for games is not eye-candy? Then what is it? If that were true the industry would never have progressed beyond Pong!

And as Rob inquired...
Quote:

Why do people keep asking if there is an upgrade price for OS4 owners. How many people have an A1 without OS4?

Because they are asking us to fully re-purchase the OS for our A1's. My questions stands - what makes it worth the price of a new OS and not a lower cost upgrade price? It's not like we are able to buy it with new hardware to run it on.

Is printing fixed? Is there a SANE library so I can hook my scanner up to it? TWAIN? Etc.... Hyperion talks about "industry standard" apps. Hans mentions two (Firefox, Blender) and only alludes to **possible** others. They want us to pay for a whole new engine for our computer. I'm trying to find out if it is worth the asking price. Am I asking too much?

@Rogue @EntilZha

Thanks for a plain and to the point answer

I'm a software guy so I'm not qualified to discuss the details, but as I understand it the "secret sauce" is that NatAmi's CPU will access memory with zero wait states, as opposed to a PowerPC which will always have some latency. So register to register ops on an 800MHz 750GX will spank the N68070 (which will be running at ~200MHz inside the FPGA), but the minute you hit memory the '070 will complete the read or write in one cycle whereas the 750 will need to wait a dozen or so cycles for the memory access to complete. The idea is that in "real" programs, the mix of loads+stores to register ops is such that the '070 will have the edge.

That's the theory, anyway. I trust that Thomas and Gunnar know what they're doing, but benchmarks will reveal how well that works in practice.

@Rogue
@EntilZha

Thanks for the updates guys. This is great news.

---
redfox
µA1-C

@DL
Quote:

If it had to do with graphics and/or GUI I counted it.

I still count eight, including the GUI stuff.

Quote:

Quote: not all the graphics related stuff is just eye-candy Better graphics for games is not eye-candy? Then what is it? If that were true the industry would never have progressed beyond Pong!

Without the added functionality, new application and game ports would not be possible (note that I said applications[b] too, not just games). A Firefox 3 port will [b]never happen without the Cairo library. Likewise there are a whole series of other apps out there that require Cairo, or better OpenGL support. Does that sound like just eye candy to you? I see that as essential functionality to get more software.

Quote:

Is printing fixed? Is there a SANE library so I can hook my scanner up to it? TWAIN? Etc.... Hyperion talks about "industry standard" apps. Hans mentions two (Firefox, Blender) and only alludes to **possible** others. They want us to pay for a whole new engine for our computer. I'm trying to find out if it is worth the asking price. Am I asking too much?

Note: I did not say that Firefox would be included. It is not. I gave it as an example of a project/app that needs some of the extra functionality that comes with Amiga OS 4.1.

What do you mean with "is printing fixed?" Printing works on OS4.0, provided that you have a printer with an Amiga driver. If SANE were included, it would be listed in the set of new features as that would be a big one.

Your "industry standard apps" quote is meaningless out of context. This is what Hyperion said: [b]"...wealth of substantial new functionality which is a prerequisite for the deployment of industry standard applications..."[b] So, these industry standard apps will not necessarily be on the CD. However, if these "industry standard applications" are ever to be ported, they will need the new functionality in OS 4.1.

Quote:

Because they are asking us to fully re-purchase the OS for our A1's. My questions stands - what makes it worth the price of a new OS and not a lower cost upgrade price? It's not like we are able to buy it with new hardware to run it on.

Every single person that has the hardware necessary to buy this, has a copy of Amiga OS 4.0. There is no point in having a "new" price and an "upgrade" price. If they were selling a MacMini version, then they might charge them a higher price because that would not be an upgrade.

Maybe you should wait and see what others say once they buy their copies before deciding to buy. There's no need to be one of the first. Just realize that if you don't buy OS4.1, over time there will be more and more software that you can't run because they require OS 4.1 functionality.

As a developer, I really want to get my hands on OS 4.1 for the extra functionality. I see great new potential.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

Quote:

The SuperAGA chipset will be 100 times faster than AmigaAGA.

Wow. That just screams compatibility issues to me. I hope you can clock it back to the old speed for games etc.

The thing I don't get about their claims for a super-fast CPU is how do they expect an FPGA to outperform PowerPC? Do they really think a couple of hobby engineers with an FPGA can beat IBM's hardware guys making a dedicated processor? Sure it might just about beat a 601 or a low speed 603/604, but a G3 or a G4? Come on. Only in very contrived benchmarks.

You know, you can get FPGAs with embedded PowerPC cores for the exact reason that a CPU actually done on an FPGA isn't likely to match an actual PowerPC.

_________________
Playstation Network ID: xeron6

who says that these features are eye-candy didn't understand how a modern operating system like OSX works. Well done to all involved

_________________
ikir

@xeron

ECS was faster than OCS and AGA was faster than both of those. Any software that's broken by the speed of SAGA was broken already.

Quote:

Without the added functionality, new application and game ports would not be possible (note that I said applications[b] too, not just games). A Firefox 3 port will [b]never happen without the Cairo library.

But it's not like that library needs to come with or be part of the OS, by force (it's not unlikely that they put it in "contrib" anyway). Anyone can port it for any version of any OS. So even if OS 4.1 did not came with Cairo it would be possibly to port apps which need it.

@Georg
Quote:

Quote: Without the added functionality, new application and game ports would not be possible (note that I said applications[b] too, not just games). A Firefox 3 port will [b]never happen without the Cairo library. But it's not like that library needs to come with or be part of the OS, by force (it's not unlikely that they put it in "contrib" anyway). Anyone can port it for any version of any OS. So even if OS 4.1 did not came with Cairo it would be possibly to port apps which need it.

Sigh. Do I need to add boring qualifiers to everything that I say? Yes, it's technically possible to port Cairo to any system that outputs graphics. However, having it in there with the system cuts down one large requirement. Would you bother porting a game if it required you to port X, Y and Z, etc., first? I doubt it.

BTW, there's actually a Cairo port on os4depot already, but I bet that the one in OS4.1 is properly integrated with the system, and a lot faster. I can't remember what the issues with it were, but it wasn't perfect (lack of hw-acceleration perhaps). No-one seems to have used it. Having it as part of the system really does make a difference.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

@all of you

Well, we have to think about how hard it was to make Amiga computer family as survivor.

I think that for the OS4.1 isn't the price so expensive. First because of hard work of the developpers team which spend time time and time to keep regular updates for OS4.0 and release of 4.1.
2nd, because of the "war for licence", the team hasn't given up. Yes, we should wait for new updates, but they were made;
3rd, in comparaison with Windows, for example, it cost here (Switzerland) depending the version about 125-375 Euro and with million of users world wide!! For Amiga there are (an hypothetic number) 50'000 regular users world wide. Then proportionally it should cost a lot of more!
Finally 4th, it is an hope for developpers with new implemented funcionnalities!

I've already made a preorder to my dealer (Relec) and i've never think it's to expensive!

What I regret is that I don't have to much money for making regular donations to developpers.

So, don't think "will it has a low price for updating" or "it's to expensive" but think "This will help our great family"... The Amiga Family! Ok?

CU

_________________
RedOneX (SAM 460ex) and an A1 which now rests quite.
Amiga OS 4.1.6 and CinemaWare games.
Classic and NG softwares.

@Rogue

Quote:

A single-file database is, in general, too vulnerable to damage. ENV: and ENVARC:, however, are seriously misused for too many purposes that it was never meant to be.

The problem whit ENV: and ENVARC: is that programs use it store configuration information, so when you reinstall your OS, the program stops working, the same thing is true for the registry system, developers shuld consider using the progdir: for configuration files, how ever I can se problems whit multi user profiles if all configurations are in progdir:

I suggest:

PROGDIR:CONFIG/ for default configuration files.
ENV:USERNAME/ for user profiles.

ENV: and ENVARC: for hardware profiles only?

Older programs using ENVARC: now can be redirected to PROGDIR:CONFIG/
Older programs using ENV: now can be redirected to ENV:USERNAME/

hardware configuratons can be redirected to ENV:HARDWARE/ OR ENVARC:HARDWARE/

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@Swoop Quote:

Have a look at SYS:Prefs/Env-Archive/AppPaths, this is certainly used by AmiUpdate for the purposes you describe, and is probably used by other programs.

Don't have OS4 :( so will have to take your word on this. The only problem with using EnvArc for this kind of thing is that each EnvArc entry should really be a (small) single line of text, not an entire 'database', so that you can have a script go IF "$variableName" EQ "something" for EnvArc:variableName .

Adding "installation tracking" in an Amiga-like way is not trivial, but I'm sure it could be done with some thought. The main point is that such a tracking system should be accessible from both Shell & ARexx. If you use EnvArc like it was supposed to be used (see previous paragraph), then using it from Shell/etc is easy - otherwise you need some kind of Shell command to parse/modify the database (and such a database should be stored in human-readable (text) format to allow easy manual editing when/if it gets corrupted).

Last edited by ChrisH on 08-Aug-2008 at 03:52 PM.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@Rogue

I have a question, when are OS4 patches for the Kaminsky DNS vulnerability due to be released? All unpatched DNS clients are vulnerable (ie it's not only a server thing!), regarless of OS, and victims are vulnerable to untraceable (and unblockable) phishing attempts. The exploit is out in the wild already, so anyone using OS4 for any web browsing is vulnerable.