@koft

Quote:

koft wrote: @COBRA Quote: COBRA wrote: @koft [quote]The rest of your rant is nuts. I have no "chosen" platform. I'm a classic user, an AROS user and recently a MOS user, I'd be an AOS4 user if it was more available for a decent price What would you consider a decent price? Say a price for which if OS4 hardware was available you would buy one.

Quote:

300 dollars including software license.

Are you jocking?.Not even Apple can offer those prices and they used common chinese mobos.Even complete NEW taiwanese PCs are not much less expensive than that
You are MAD, live completely out of reality or are the most big troll ever

Last edited by Fransexy on 25-Jun-2010 at 11:07 AM.

_________________
No PowerPC, No Fun
Make Amiga Great Again

@koft

Quote:

COBRA wrote: @koft Quote: The rest of your rant is nuts. I have no "chosen" platform. I'm a classic user, an AROS user and recently a MOS user, I'd be an AOS4 user if it was more available for a decent price What would you consider a decent price? Say a price for which if OS4 hardware was available you would buy one.

Quote:

300 dollars including software license.

your still baiting koft! ... go on ... go on :-/

Last edited by Framiga on 25-Jun-2010 at 11:11 AM.

_________________

@tiffers

Quote:

tiffers wrote: @koft Quote: koft wrote: Wow, that's a wall of text. TL;DR comes to mind... It did seem a little ranty didn't it. I was trying to make a point, that you are looking at it from 'one' point of view. Life very rarely hold true to a single circumstance. Quote: You're over reacting to the points I have made. A matter of opinion of course Quote: Optimally, A-Eon would be shipping product right now. They originally claimed that we'd have this before summer, guess what, it's summer. A-Eon claims to have to be silent about the CPU because of an NDA agreement, but it's already past the date of their most optimal release schedule. Do you think an NDA cannot be written to make room for slippage in the planned release schedule. Quote: The only way to fit their claim to something logical is if the NDA restricted them from talking about the "secret part" until *after* they were shipping units. That's anal to some absurd degree and it makes no sense whatsoever. I disagree, but I agree it is an anal idea. I don't think they would have such a clause, only release info after it goes on sale. But as I proposed in my previous post, why would they not have a clause which restricted info being released until they were sure the product actually woudl go on sale? ie when the final design goes to manufacture? Or once a number (say 100 or more) boards are sucessfully manufactured and passed quality control testing?

If the source was that anal, they'd never deliver chips to A-Eon.

Quote:

Quote: That's absurd. A-Eon isn't talking about the CPU because they choose not to, and it's part of their bizarre marketing strategy. So, you don't know anyone who imposes absurd restrictions, or has other absurd business practices? It's not impossible, and maybe being related to the fact the SoC is aimed at DoD and other 'sensitive' applications maybe, just maybe, PA Semi went with an overly strict NDA, and just applied it to everyone. I can't wait to hear what the 'interesting story' is behind the CPU partnership (as mentioned on the What is X campaign pages.)

Yes, some companies do absurd things, they tend to go out of business quickly. I see that the pa6t never made any success in the market, perhaps there is a reason for that. That's not saying much, it seems to be in a miles long list of chips that never went anywhere.

Quote:

Quote: In my opinion it's a shame they've chosen to go this route in the PR department. We've seen endless waves of holy wars because of it. I'm glad though, I've gotten sick of all the BS. I picked up an emac earlier today and installed MorphOS and was thoroughly impressed to be honest. Half the power of an A1X1K for a grand total of 256 dollars. And man is that sweet that the cost of the machine + MOS license happened to be a power of 2. LOL Yeah I love numbers like that. And it is a shame that there are such restrictions, I agree. But if the NDA is valid, it's valid. You may opine as much as you like, but it won't change a thing. I think it really is a matter of the CPU supplier saving face if small startups never come out on top after announcing their use of the SoC. Sounds like a fun machine. I'm in the camp of not buying 2nd hand equipment though, and I don't like the look of the eMac's. There is a Mac Mini (1.5GHz) listed on ebay in Australia, but it's $330 (+post I figure) and that's just... not doable I think. Yeah, I expect Mac equipment to de-value a little faster than that. I do buy 2nd hand gear, but not at those rates. But back to the point, I don't think it's BS. I think it's simply lack of experience. I am constantly astounded why ANYONE in this market doesn't use the 'forward look statement' disclaimer. It would really help to quieten some of the raging emotions in this community. tiffers

As for CPU supplier saving face, nah. All a supplier cares about is moving product, they don't care if the customer face plants.

_________________

@Fransexy

Quote:

Fransexy wrote: @koft Quote: koft wrote: @COBRA [quote] COBRA wrote: @koft [quote]The rest of your rant is nuts. I have no "chosen" platform. I'm a classic user, an AROS user and recently a MOS user, I'd be an AOS4 user if it was more available for a decent price What would you consider a decent price? Say a price for which if OS4 hardware was available you would buy one.

Quote:

300 dollars including software license.

Are you jocking?.Not even Apple can offer those prices and they used common chinese mobos.Even complete NEW taiwanese PCs are not much less expensive than that
You are MAD, live completely out of reality or are the most big troll ever[/quote]

Efika?

_________________

@koft

Quote:

Efika?

Computer that can´t even be used for common tasks?

However, I think it could be possible to make low cost (eg. less than 400 USD) computer powered by 440EP and something like Silicon motion 502, 1 GB RAM (or even less), small and fast SSD disk and some USB 2.0 connectors. Give it to netbook case with LCD and it will sell really good...

@koft

What the lowest cost PPC motherboard + CPU still in production ?

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

@asymetrix

Quote:

It does not bother me too much as the military buy this processor in the millions.

What ever gave you that idea? If the military was indeed buying "millions", don't you think PA-Semi would have been worth more than $278 million?

Quote:

But looking at Freescales new products, they have similar int/float as PASEMI

How are they any more similar to PA-Semi than to, say, IBM?

@DiscreetFX

[Edit] (I got the context slightly wrong here; That was another thread. However, the sentiment applies here, too --- should I *not* point out that the article referred to is 5 years old, and that in the meantime PA-Semi was sold to someone who has publicly stated they have no interest whatsoever in the products? How is it "comparing Apples and Oranges" to state that the "potential future product" hypothesised in that 5 year old article is about as likely to become available as a dodo-burger from KFC?)


Quote:

This machine is the highest specced Amiga yet and should be a great upgrade for Amiga OS 4.x fans.

Sure. Assuming it actually comes out, *and* is halfway decently supported by the OS. Having two 64 bit cores doesn't do much for the user if the OS only uses one, in 32 bit mode.

However, I fail to see how "the X1000 is going to be a great upgrade" is relevant to "if the XCore chip can be used to speed up 3D rendering, we will do so". It's blindingly obvious that it *can't* be used for that purpose[1], so the statement is completely meaningless. Nothing "On a positive note" about it.

Quote:

Apple and Oranges comparing it to something else unless what you compare it to also runs Amiga OS 4.x.

I call strawman. I did not make any comparison; I merely pointed out the faults in your statement.


[1]: The memory available can hold less than 3000 vertices. There is no FPU. The chip runs at 500MHz, and is not even superscalar.

Last edited by umisef on 25-Jun-2010 at 04:06 PM.

@umisef

Indeed even though I am optimistic, The reality is the same proof of the pudding that Hyperion MP referred to last year or so.

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

@pavlor
>Maybe another example will convince you:

>Sun UltraSPARC IIi 648MHz
>Coremark: 916.09 (2 PThreads!!)

>For comparison same CPU scores 246 SpecInt2000 (comparable to 750FX or >Pentium III)

>I don´t think that 5200B could score 600 SpecInt2000 (comparable to Pentium 4 1.6 >GHz)...

you cant compare specint and coremark.coremark speed up not so much by bigger caches or faster memory.its embedded world small code.

specint is a desktop benchmark with Desktop apps that are in size of 2 megabyte or more and cause on a CPU as 5200 with no second level cache lots code cache misses.

but coremark use only simple routines and as can see, on the 5200b it is only 15%slower in coremark /MHZ as the pa6.

so its sure the caches and faster mem is not need of the pa6t, because the code and data is small(as in embedded apps realistic)

But as can see modern Intel(not ATOM CPU) or AMD are in this benchmark more than 60% faster in coremark /MHZ.

I can tell it once again, because i see no other benchmark of pa6, so i cant look at dekstop relevant benchmarks for pa6.

what we want to do with AOS, the coremark show not very good, but its better than nothing.

And when coremark show for pa6 so slow values in compare to AMD or Intel CPU, i cant believe that a PA6t give the speed of a 1,6 GHZ Intel or AMD CPU or G4 in specint, or other

The fastest PPC i think is the IBM Xeon in the xbox360.
it reach with his 3,2 GHZ AMD Athlon 1,6 GHZ speed in specint or blender benches you can see.Iook at the playstation 3 values, i think the PPC is on the Cell and Xeon same.Xeon only have 3 real PPC cores

Last edited by bernd_afa on 25-Jun-2010 at 05:36 PM.
Last edited by bernd_afa on 25-Jun-2010 at 05:35 PM.
Last edited by bernd_afa on 25-Jun-2010 at 05:34 PM.
Last edited by bernd_afa on 25-Jun-2010 at 05:33 PM.

@umisef

We dont know how many Pa-Semis are out there. For all we know they are keeping secret to its use. Closing down a company dont mean anything if the military payed them off or other companies have licenced the IP for years.

My thoughts on similarity is they - Freescale/IBM are using the PA-Semi designs as a blueprint for their own processors.

looking at integer and float performance is strikingly similar.

All A-Eon need to do is contact the IP owner or manufacturer and guarantee discounted producution for 3 years or even request multiple cores.

_________________
Download 499.26 Mbps, 659.94 Mbps Upload :)

@stedy

Quote:

I can't find the CPU in the PCI listings, that should show up. The device A000, has the device code of 8, indicating it is the 8th device on the PCIe bus.

So you *can* find it....

Quote:

If it is the root complex, it should be the first device.

....it just isn't where you expect it.

Well, considering there are two of them (for the two cores), they can't very well both be the first device, can they? However, there is only one host bridge, and *that* is at device address 0.

And if you look at the device tree available here (search for "1959" in that page), you will find that the device addresses on a real-world, definitely-not-faked PA6T are exactly as those shown by ShowConfig on the X1000, except for differences related to the SERDES setup (i.e. the machine the device tree was taken on is using a couple of the PA6T's ethernet MACs, and seems to have a different collection of PCIe ports).

So if it were a fake, it would be a quite accurate one. And considering how hard it was to find a source for the actual real-life PA6T device tree, the question of "why would anyone bother" would need answering.

@bernd_afa

I think you overvalue Coremark benchmark. The examples I gave clearly show that one can´t blindly compare one Coremark result with another.

Quote:

The fastest PPC i think is the IBM Xeon in the xbox360.

970MP is faster...

"IBM Xeon"

It's called Xenon.

and some kind of block diagram...
http://www.xbitlabs.com/misc/picture/?src=/images/news/2004-04/xbox2_scheme_bg.gif&1=1

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

By the way in the latest Amitopia (at VFC) Trevor confirmed it's a 2Ghz processor running underclocked.
Well everybody knows, but receiving confirmation was nice

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

@DAX

however, these cpu are embedded cpu, don't forget you.

Last edited by Seiya on 25-Jun-2010 at 10:03 PM.
Last edited by Seiya on 25-Jun-2010 at 10:03 PM.

_________________

@Seiya
(what did you mean?)

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

@Seiya

Quote:

however, these cpu are embedded cpu, don't forget you.

The story I heard about PA6T was that PA Semi intended it to be the G5 CPU in Apple iBook/PowerBook G5 laptops, since IBM's were so horribly power hungry and Freescale wasn't doing a G5 class chip. Apple pulled the rug out from under them by changing to Intel instead, and PA Semi refocused on markets they could still sell to. It wasn't intended by design to be an "embedded cpu", but there's no reason not to sell to embedded customers if they want it.

Why does "embedded cpu" have to have a stigma attached to it anyway? Not all "embedded cpus" are limited to Efika class performance/features you know.

_________________
All glory to the Hypnotoad!

@KimmoK

Quote:

"IBM Xeon" It's called Xenon.

Actually it's called Waternoose.
The motherboard is called Xenon but the CPU seemed to pick up that name for some reason.


--

As for the PA-Semi performance, I expect it to completely destroy the G4. In fact I expect it to come in somewhat faster than an equivalently clocked G5.

It's got a very fast memory interface and it was designed by some of (if not the) best people in the industry.

_________________
Whyzzat?

benchmark pa semi

http://www.coremark.org/benchmark/index.php

Last edited by Seiya on 25-Jun-2010 at 11:04 PM.
Last edited by Seiya on 25-Jun-2010 at 10:26 PM.

_________________