Ouch!

Still cheaper (and available) than PAT6 ;)

pampers,
Quote:

Small comfort when I can buy two complete Cortex A15 based systems for the price of the most costly of those.

Update:

> http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=P4080&tab=Buy_Parametric_Tab
>
> 530 USD for the 1.5 GHz version, 442 USD for the 1.3 GHz version and 368 USD for the
> 1.2 GHz version. All versions in quantity of 100 pcs.

Some weeks ago, Freescale silently released the QorIQ P4081:

http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=P4080
http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=P4080&tab=Buy_Parametric_Tab
http://www.freescale.com/files/graphic/block_diagram/P4080_40_81_BD_IMG.jpg
http://www.freescale.com/files/32bit/doc/fact_sheet/QP4080FS.pdf

Differences compared to the P4080:
- 1.2 GHz instead of 1.5 GHz max frequency
- 1 instead of 2 GbE controllers
- price of 157 USD instead of 297 USD for the 1.2 GHz version (and 118 USD for the 1.0 GHz version)

More PCIe controllers, good price.

Thanks.

> More PCIe controllers

No, the product page has it wrong. It's correct on page 3 of the fact sheet PDF though, which says 3 PCIe controllers for P4040, P4080 and P4081.


Edit: It's been corrected on the product page since.

[ Edited by Andreas_Wolf 10.03.2013 - 20:36 ]

Irritating.
When they say PCIe controllers, are they talking controllers or lanes?

> Irritating.

True. But taking into account that the P4040 has 3 PCIe controllers not 4, I'd say it's a safe bet that the fact sheet PDF has it right and the P4081 has also 3 not 5.

> When they say PCIe controllers, are they talking controllers or lanes?

The technical data and nomenclature provided by Freescale may be irritating at times, but "controllers" indeed means controllers here :-)

That's cool.
3 PCIe lanes would be of limited utility.
Any idea how many lanes each controller has access to (or the total #)?

I never thought I'd find the e500mc that interesting, but the pricing makes it pretty attractive.

> Any idea how many lanes each controller has access to (or the total #)?

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=37074&forum=14#693507

Update:

> Prices for QorIQ P5010 (Avnet), P5020 (Avnet) [...] are now public [...].
> They are as follows in quantity of 100 pcs.:
>
> P5010:
> 1.6 GHz: 340 USD / 210 GBP / 260 EUR
> 1.8 GHz: 360 USD / 225 GBP / 275 EUR
> 2.0 GHz: 400 USD / 250 GBP / 310 EUR
>
> P5020:
> 1.6 GHz: 400 USD / 250 GBP / 310 EUR
> 1.8 GHz: 420 USD / 265 GBP / 325 EUR
> 2.0 GHz: 470 USD / 290 GBP / 360 EUR

30% lower prices now to be found directly at Freescale:

P5010:
1.6 GHz: 240 USD / 160 GBP / 185 EUR
1.8 GHz: 255 USD / 170 GBP / 195 EUR
2.0 GHz: 285 USD / 190 GBP / 220 EUR

P5020:
1.6 GHz: 285 USD / 190 GBP / 220 EUR
1.8 GHz: 300 USD / 200 GBP / 230 EUR
2.0 GHz: 335 USD / 220 GBP / 255 EUR

Andreas_Wolf,
Quote:


Getting processors has never been the problem. It's getting a board developed and built is the expensive built.
Add in this cost and you suddenly understand why the OS4 guys send so much for what are basically very low end boards (and I'm not even talking about the X1000).

At the sorts of volumes you can sell to MorphOS users building your own makes no financial sense. OTOH going with the existing systems (i.e. Macs) makes a lot of sense.

>At the sorts of volumes you can sell to MorphOS users building your own makes no financial sense.

It didn't make financial sense to build the X100o either, but it got done.

The only problem I have with the P5010 and P5020 is the lack of AltiVec.

Great reference, btw, Andreas.
"not enough PCIe"
2x, 2x, 2x - Man, not even adequate for a video card

[ Edited by Jim 03.03.2013 - 03:46 ]

> Getting processors has never been the problem. It's getting a board developed and built
> is the expensive built.

I know that. My price listing is a continuation of a sub-thread that started in this thread with someone else's statement on the QorIQ P5 series:

"there are pretty many other things within the package for real high end networking (several 10 Gb controllers, packet inspector, buffers, security stuff). Not that MorphOS gets hindred by these compounds, but I guess they are rather bad for the wallet."

Looking at the prices helps quantify how bad for the wallet the non-desktop features of the P5 chips actually are.

> Add in this cost and you suddenly understand why the OS4 guys send so much for
> what are basically very low end boards (and I'm not even talking about the X1000).

I've been understanding this fact for so long that I can't even remember the time when I didn't.

> At the sorts of volumes you can sell to MorphOS users building your own
> makes no financial sense.

True, that's why a combined platform for OS4, MorphOS, AROS/PPC, Linux/PPC, Haiku/PPC and whatnot would make more sense. ACube seeks to market the Sam460ex not only to OS4 users but also to AROS/PPC and MorphOS users. And it's no secret that Trevor would have liked MorphOS on the X1000 as well (besides OS4 and Linux/PPC), and that he hopes to see MorphOS running on the X1000's successor(s).

> going with the existing systems (i.e. Macs) makes a lot of sense.

Yes, there's no doubt about that.

> 2x, 2x, 2x - Man, not even adequate for a video card

Hey, in that other thread you say you're seriously considering a Sam460ex with its 1st-gen x4. That's the same speed as a 2nd-gen x2 of the QorIQ P4 :-)

I know, I know Andreas.
And who is to say that it might not be possible to combine the output of two controllers for X4?

And the SAM is only something I am considering because I'd like to play around with OS4 and still have the option of using a better OS. ;)

Really, I think when G5 support is introduced, I'll wind up using just my Powerbook and a G5.

I already have an R300 video card for the G5.
And since I will be dual booting that system with OSX, I will finally get to use the Sonnet Tempo SATA X4P card I have lying around.

> who is to say that it might not be possible to combine the output of two controllers for X4?

I don't think that's technically feasible. According to Neko, splitting the lanes coming from a controller is not possible, so I guess combining the lanes of two (or more) controllers isn't either. But I'm no hardware expert, so I wouldn't know for sure. Do you?

Andreas_Wolf,
Quote:

I have no clue.
What did he mean by not being able to split lanes?
Multi-plexing lanes or seperating the two from each controller to equal 2 X1 connections?

I would hope we would be able to run seperate X1 connections.
An X2 connection doesn't seem that useful.

> What did he mean by not being able to split lanes? Multi-plexing lanes
> or seperating the two from each controller to equal 2 X1 connections?

Judging from his posting I linked to, the latter.

> I would hope we would be able to run seperate X1 connections.
> An X2 connection doesn't seem that useful.

According to what Neko wrote, you can run an x1 connection from an x2 controller, but only one, not two. So you'd be effectively throwing away one lane.
While an x2 connection doesn't seem that useful at first sight, you can use a PEG slot (x16) with an x2 connection, similar to how it is done on the Sam460ex.

With the higher sppeds of this connection, I guess X2 would be adequae (if you can find a PCIe video card that doesn't force the bus to operate at an earlier/slower speed).