AmigaOne DMA Problem

[Floid]

itix wrote:

This talk of the 90MB/s PegII sounds pretty specious now

It only suggest max transfer rate over ATA interface. Like benchmarking ethernet :-)

Actually, my bad (unless there's a grand conspiracy to re-edit everyone's posts when I'm not looking) ... I was still having my coffee, and I misread that as a claimed HDParm result.

The same-sector-read methodology is a sound one for this question, so now I have to wonder what Mikeymike's on about.  Not knowing how the system really plumbs together, I can't really trust numbers from MorphOS (not to say it's cheating, but I don't know how/if it implements a buffer cache) ... but that is what should be expected yoinking the same sector over and over from the disk's cache.

I don't think I see anyone claiming accurate, recently-trialed numbers for the same type of test off a Mai'd board of any sort (if, again, we're trying to figure out whether *something,* bugs, fixes, or user-misconfiguration pushes down the peak theoretical throughput on the A1), so I'll try to hold off speculation while I doublecheck the thread.

[Cymric]

Floid wrote:
Sadly, I'm not finding any HDParm numbers from KT133A/686B users with ATA100 drives (and the KT133/A bugs put a wrench in that thought, too.)  Still, look at these numbers from the 2.4 kernel era, with good, standard hardware... they 'suck!'

I got interested in the numbers for my system, and at first found them despicably low (around 4.5 MB/s for the buffered disk read). I have been using the system for well over a year, and never really noticed this very low value... Duh. Turns out my 2.4.18 kernel did not support the VIA chipset in my PC; I could not even enable DMA manually. So one upgrade to 2.4.26 later, I see 272.34 and 50.00 MB/s on the one hard disk, and 261.22 and 40.00 MB/s on the other. I am genuinely impressed :-). (The system is a KT333/8235-machine, by the way.)

[mikeymike]

The same-sector-read methodology is a sound one for this question, so now I have to wonder what Mikeymike's on about.

In what respect?  Testing whether DMA works 100% properly?

[Hammer]

(if they did Microsoft would have gone out of business years ago), the embedded market is a different ball game.

Refer to www.windowsfordevices.com for examples of embedded Windows devices i.e. Microsoft applies market segmentation for its products. It would be unrealistic for Microsoft and is its value adding partners to offer a full Windows XP install on a limit role computer.  

PS; They made "small" profit on embedded related activities (minus XBOX).

Desktop users don't put their machines through that sort of abuse so machines don't need to be so reliable

What kind of abuse?

[KimmoK]

@KennyR
"How is buying an AmigaONE going to support anything?"

According to AlanRedhouse. AOS4 licence is being paid to AOS4 developers for every sold AmigaOne.

"I really don't see how buying an A1 supports the 'Amiga platform'"

See above.
And I was talking about AmigaOS platform.

" any more than buying a Pegasos "

Buying Pegasos supports the competition of AOS4 and funds the court actions of Genesi trying to get the hold of AOS4. (in the end Genesi would have no reason to let AOS4 live to help the sale of competing HW)

OK that came from my perhaps biased twisted mind ... or pehaps I'm just too DISGUSTED from BBRV actions.

"or a copy of Amiga Forever,"

That is supporting AmigaOS. (kind of, especially if licence money go to right address)

"or downloading AROS."

That is not supporting AOS.

[KimmoK]

@takemehomegrandma

"then the Pegasos is an EXCELLENT choice! All major Linux distros runs fine, and it's only getting better all the time!"

So is the x86 competition, I think.

But yes someone might choose Peg2 for that low power consumption but rich media / powerfull telecom use.

[KimmoK]

@Piru

"Marvell is over 100% faster with HD DMA than ArticiaS. Same southbridge, same HD."

But different motherboard, different HW glues / fixes.

And if one compares peg1 to A1:
Different motherboard, different HW glues / fixes, different drivers, (different OS), different southbridge.

"I am no expert, but to me it seems as if Articia was seriously crippled."

It could be.

IIRC: genesi did not try to get ArtisiaS weirdnesses fixed by SW, because it would have been against their strategy.

I have understood that Pegasos1 does not have drivers that can do reliable DMA without April chip's help.


It will be very interesting to compare (A1, Peg1, Peg2) HDD performance with CPU usage when there exist AOS4 with DMA.

[KimmoK]

@KennyR
"Even if every single A1 owner, potential or present, had bought a Pegasos, there would probably only be about 1500 more sales"

A lot of people chose Pegasos pecause it's cheaper than A1. (I could giv a lot of names)

"That's just not worth the expense."

Still Genesi made heavy push to get Amiga developers and users.

"MorphOS is a toy OS in the world of computing and Genesi know that as well as anyone. Genesi's ambition is far bigger than MOS. As the slogan goes, MOS is a free gift. Smearing a tiny outfit like the A1 for the sake of a tiny OS like MOS is a total waste of effort."

That is something for MorphOS users and AmigaOS fans to think about.

If pegasos prices some miraculously way cover the cost of niche (about 1000 units per board) HW R&D and production, how are the MOS R&D costs covered?

[KennyR]

KimmoK wrote:
"Even if every single A1 owner, potential or present, had bought a Pegasos, there would probably only be about 1500 more sales"

A lot of people chose Pegasos pecause it's cheaper than A1. (I could giv a lot of names)

And? This is how market economics works. The A1 loses out to a better deal, just as the Pegasos usually loses out to an even better one - a PC. The A1 is ridiculously priced and has lost out on many developers and users.

If pegasos prices some miraculously way cover the cost of niche (about 1000 units per board) HW R&D and production, how is the MOS R&D costs covered?

Who says they are?

If we take it in man-hours payment for code, MOS is not being paid for. But then, the funds from selling A1 won't pay Hyperion's wages either. Both systems are developed part time by the coders with real jobs to bring income.

"How is buying an AmigaONE going to support anything?"

According to AlanRedhouse. AOS4 licence is being paid to AOS4 developers for every sold AmigaOne.

"I really don't see how buying an A1 supports the 'Amiga platform'"

I'm sorry, but I don't consider supporting Eyetech or Hyperion supporting the Amiga platform any more than it would Elbox and Titan or Genesi and bPlan. They're just companies. Hyperion got the permissing to use this miraculous name, but the OS is becoming very unlike OS3 in the way its being designed and the A1 is just a Teron. I'm afraid to me supporting it alone is only supporting the marginilisation of the amiga platform - and marginalised in its own niche!

Buying Pegasos supports the competition of AOS4 and funds the court actions of Genesi trying to get the hold of AOS4. (in the end Genesi would have no reason to let AOS4 live to help the sale of competing HW)

Well, even if it was, so? You'd rather have shadow companies owning this OS? KMOS the mysterious don't-know-anyone-except-GarryHaretheCEO could-be-a-IP-laundering tool?

No... I'd rather have a company that can actually deliver hardware, who I know works for it, and I can speak to anyone on its ranks. KMOS are not that. KMOS, for all I can see, have all the flaws of Amiga Inc, and I will never support such a company on faith ever again. They sell stuff with good prices, I buy it. Else the Amiga name isn't going to dazzle me into buying substanard stuff.

"or downloading AROS."

That is not supporting AOS.

It's supporting the Amiga platform. The potential amount of software from AROS could be staggering, and easily ported to any other AmigaOS flavour - yes, even OS4. If AROS users want to call it AOS, who's to stop them? Amiga Inc?

Seeing the difference from supporting the platform proper and just from supporting name-leeching companies is the first step in actually being able to really support it.

[GadgetMaster]

KennyR wrote:
but the OS is becoming very unlike OS3

Err yes. Thats why it is called OS4

Sorry couldn't resist.

:lol:  :-D

[KimmoK]

@KennyR

"Hyperion got the permissing to use this miraculous name,"

Affect of the blue pill I assume.
Hyperion made a deal to continue AmigaOS development.
AmigaOS4 is based on AmigaOS code.

"but the OS is becoming very unlike OS3 in the way its being designed"

It's more like OS3 than how MOS is, especially the design.

(I agree there are things that make wonder how the end result will work, time will tell if the Amiga-like feeling is lost, so far it seems to function OK (according to those who have tried it), same for MOS. )

"and the A1 is just a Teron."

(nothing to do with the AmigaOS platform, but)
And peg is just a CHRP / POP ?

"I'm afraid to me supporting it alone is only supporting the marginilisation of the amiga platform - and marginalised in its own niche!"

Well, people are ofcourse free to support what ever they want. (I was planning to, but I'm not sure I will get peg+mos any more, not untill some changes are made @ Genesi)

I've heard that some have bought M$ products even.




>Buying Pegasos supports the competition of AOS4 and funds the court actions of Genesi trying to get the hold of AOS4. (in the end Genesi would have no reason to let AOS4 live to help the sale of competing HW)

"Well, even if it was, so? You'd rather have shadow companies owning this OS? KMOS the mysterious don't-know-anyone-except-GarryHaretheCEO could-be-a-IP-laundering tool?"

KMOS does not have any reason to kill AmigaOS (or to kill the multi-HW strategy).    ( again fully personal opinnion after applying my common sense )

"No... I'd rather have a company that can actually deliver hardware, who I know works for it, and I can speak to anyone on its ranks."

I rather have companies that can deliver hardware and a company that deliver AmigaOS.

"KMOS are not that. KMOS, for all I can see, have all the flaws of Amiga Inc, and I will never support such a company on faith ever again."

To me KMOS does not have the flaws of Amiga Inc. But it neither has any merits either. It's pretty neutral. Perhaps the CEO seems too sane for Amiga market. (no, wait, I forgot the business card grazyness)

"They sell stuff with good prices, I buy it."

They also say cow meat from UK is cheap. I do not buy it.


"If AROS users want to call it AOS, who's to stop them? Amiga Inc?"

No. KMOS has that option. They already acted when the threat against AOS became apparent in the thendic-Amiga court issue.

"Seeing the difference from supporting the platform proper and just from supporting name-leeching companies is the first step in actually being able to really support it."

I think helping AOS4 in any way is the best way to help to keep AOS alive.
Other than that, supporting MOS (without supporting Genesi, if possible), AROS, Amiga SW/HW developers is far better thing to do than supporting Amiga.Inc.


yet one update...
btw. "supporting name-leeching companies" made me build the following reminder of who might get money when one buys A1/AOS4:
- KMOS: AOS defence in court & something else that is not visible yet
- Hyperion: AOS development
- Eyetech: the first company to arrange new AOS4 compatible desktop HW

IMHO: it is better option for Amiga OS fan than the main alternative ( why support AOS imitator OS that is planned to be "something else" if the "original" is also there, why support "other than AOS HW platform" rather than the existing AOS HW platform etc. )  

Anyway... I have my opinnion, I let people have theirs. (correction:  I try to let people have theirs, sometimes I fail. Sorry for that.) :-(

[KennyR]

@gadgetmaster

No, it's called OS4 because the people behind it hated bPlan and weren't content to see MorphOS become the only Amigalike system being marketed, and so put up with AInc's horrible licencing to make an OS which is now almost totally superfluous.

[KennyR]

Hyperion made a deal to continue AmigaOS development.
AmigaOS4 is based on AmigaOS code.

A deal with Amiga Inc, a proven con outfit. I see no connection with the real AmigaOS here.

And OS4 is not based on AmigaOS code, that's just a stupid claim from years back. It's engineered from the ground up using autodocs to be OS3 compatible, just like MOS and AROS. That has become very clear from its greater compatibility to autodocs than to actual OS3, which breaks a lot of software. Eventually Hyperion will fix this I don't doubt, but it shows straight away that OS4 was not based on original, hardware dependent OS3 68k C/ASM code.

It's more like OS3 than how MOS is, especially the design.

Nope. MOS is more like OS3 in every way. OS4 is rather incompatible, uses the MMU like crazy (Commodore NEVER used the MMU), and is furthermore incompatible in other conceptual ways (like renaming all the language catalog dirs in ENGLISH, breaking all old installers). OS4 has departed from OS3 in MANY ways. It's probably the most alien to the OS3 structure of all the current AOS clones.

I rather have companies that can deliver hardware and a company that deliver AmigaOS.

That's a pointless argument, because if tomorrow Hyperion lost the right to call OS4 AmigaOS and Genesi got it, I'm damn sure most of the now-KMOS4 users wouldn't suddenly buy Pegasos just to run AmigaOS.

As for Eyetech...they don't deliver hardware. They simply rebadge it and resell it.

What is the actual name worth, except hype and marketing? What is it worth to the user? Consider that.

To me KMOS does not have the flaws of Amiga Inc. But it neither has any merits either. It's pretty neutral. Perhaps the CEO seems too sane for Amiga market. (no, wait, I forgot the business card grazyness)

That's the whole point - what do you actually know about KMOS? Even less than you did about Amiga Inc, I'm pretty sure. How can you possibly trust them?

They also say cow meat from UK is cheap. I do not buy it.

Irrelevant. It would be relevant of the Pegasos hardware were horribly buggy. It's not.

No. KMOS has that option. They already acted when the threat against AOS became apparent in the thendic-Amiga court issue.

I'm willing to bet KMOS has just as little court power as Amiga Inc. I could be wrong, but they don't seem to be clamping down on small entrepreneurs using the Amiga OS4 name without royalties yet, do they?

I think helping AOS4 in any way is the best way to help to keep AOS alive.
Other than that, supporting MOS (without supporting Genesi, if possible), AROS, Amiga SW/HW developers is far better thing to do than supporting Amiga.Inc.

Well, we differ in the first point. I have the genuine feeling OS4 was introduced just to spike bPlan's plans early on. I agree with the second point - supporting actual developers and trustworthy companies supports the Amiga. Sadly, the list of trustworthies is rather short.

btw. "supporting name-leeching companies" made me build the following reminder of who might get money when one buys A1/AOS4:

Let me amend that a bit. Who gets money when you buy A1/OS4: Amiga Inc (licencing fees); Mai logic; Eyetech; Hyperion. None of which are particularly deserving.

IMHO: it is better option for Amiga OS fan than the main alternative ( why support AOS imitator OS that is planned to be "something else" if the "original" is also there, why support "other than AOS HW platform" rather than the existing AOS HW platform etc. )

OS4 is an imitator OS too. Try to think of it as if no-one had the Amiga name. You have HyperionOS, bPlanOS, and OpenROS. Which is most 'Amiga'?

[minator]

@Hammer

I was being sarcastic!

In any case I was referring to the difference in reliability between older MS desktop OSs and embedded stuff.  If a computer crashes you reset it, if your CD player did the same you probably take it back.  We do not expect a high level of reliability from desktop computers, they are getting better but that's why I said "years ago".

What kind of abuse?

Being on 24/7 for months if not years.

The Win NT/2K/XP line is much better at this of course but even then I've seen crashed ATM machines running NT embedded, I've never seen one with OS/2 crashed...

[mikeymike]

Ok guys we're in danger of going well off-topic to the point of no return.  I'm going to lock the thread for 24 hours, please think about whether your planned/intended replies to current posts are on-topic or not (ie. is it about the A1 apparent DMA issue... not Pegasos, not x86, not Windows embedded, not mindless advocacy...).

If the thread continues to go off topic after the 24 hour lock, it'll be locked permanently.