> I wonder how the 604e at 366MHz compares to the Efika's @400MHz? I think I
> remember someone referring to the Efika's CPU as a G2, or G3 class PPC CPU,
> so I imagine that the Efika would be faster, even if the clock rates were identical.

The core of the MPC5200B is the G2_LE (also called e300c0), which is a slightly enhanced 603e core (also called G2).

http://www.nxp.com/docs/en/reference-manual/G2CORERM.pdf

So in essence, the MPC5200B should be slightly faster per clock than the 603e. The fact that the 604e is also faster per clock than the 603e should be able to at least compensate this. I would like to see proper CPU benchmark results between 604e and MPC5200B myself.

> What I don't understand is why tlosmx would sell his fantastic A4000 [...] after he
> already had purchased an X5000-5020 [...]. Maybe he sold the ultimate A4000
> to purchase both the 5020 and 5040 X5000's?

"was need money for buy the x5000"
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=42094&forum=33&start=200#803082

> Didn't the RAM speed problems show up on the 5020, or is it only the 5040 that he is
> complaining about which has the RAM speed problems?

Isn't that the same twice ("not on the 5020" vs. "only the 5040")? Either way, he clearly complained about both systems, with the X5000/20 being even slower than the X5000/40.

> I hope he can recover enough of his investment by selling both X5000 systems

Does he still own the X5000/20? Or did he sell the X5000/20 to buy the X5000/40?

i sold my x5000/20 (i had it from september 2015 ) a month ago i have the 5040 from september 2016
the ram performances are poor in both machines.
soon i will sell x5000/40 too,no reason for me to have it... all what i sow and was gave to me by this experience make me really disgusting... i dont wanna see nothing more that come from a-eon or amigakit.

for os4 i can will use qemu sam460 or winuae/fsuae, for mos i hope will have one day it running on my g5 quad.

my yesterday video of quad with ubuntu mate 16 video

[ Edited by tlosmx 18.09.2017 - 23:40 ]

> The 604 core was a deadend at Moto/Freescale/name_of_the_week,
> all newer CPUs/SOCs can be traced back to the 603.

The 603 lineage stopped with the e300 (MobileGT / MPC512x and PowerQUICC II Pro / MPC837x) and the e600 (MPC7448 / MPC86xx) short of a decade ago, so has turned out as a dead end as well. The MPC8xx / MPC5xx / e200 and e500 lineages (including e5500 / e6500) which the newer Qorivva and QorIQ P/T (as well as older PowerQUICC III / MPC85xx) are based on are not derived from the 603 / e300 / e600 lineage.

Quote:

There opens a great opportunity for you, Jim!

i have already a buyer jim. he is collecting money for it.
i will sell my complete system. X5000/40 is not priceless

Can we get back on topic?

Has any news at all on the status of MorphOS been posted on the machine? For now I have my G5 and X5000 on a KVM but would love to just have the one system up.

Quote:

Why don't you talk to one of the developers? Maybe they could use another beta tester.

Quote:

I'd love to get in on that, but I don't think the 250X card in my machine will be supported. That and I've never got a reply back from any when in IRC or here in the past. I never thought to ask as a result.

Quote:
i'd love to get the opportunity to test MOS for X5000. i've been dying to start running it on my X5K since my pegasos2 is really on its last legs. i'd even be happy to help demo it at amiwest if amigadave can help make sure i point all the cool bits. there are quite a few potential converts at amiwest from both the 68K enthusiasts and (increasingly) disenchanted AOS4 users. and i expect MOS to smoke AOS4 on the X5K in terms of performance.

-- eliyahu

Quote:

Well, I did offer to loan you a Radeon HD 6750, but I don't know what state the drivers for the 6000 series are in.
The R600 and R700 drivers should be close enough to release for a functioning beta, but that would mean a switch to something like a 4650 (or 70), or a 4850/70/90 card.

I've shipped the R700 cards I had on hand to Mark, but I'd be willing to source something for a demo.

And a 4870 or 4890 would probably still do well under OS4.1.

Further, we might be lucky, and find that Mark has a serviceable 5000 or higher driver.

Quote: works but no 3d. all 4xxx 5xxx 6xxx (NI) series dont have 3d on os . Only some cards of SI series is 3d accelerated with warp3d si and warp3d nova.

I'd be willing to put in a new drive with it and another video card if it booted right into MorphOS. Switching back the original drive and video card would be simple enough to get OF4.1FE back going when it was done.

Hopefully I can get to Amiwest and find a room to crash in. Heck all I need is a couch.

Quote:

You ought to talk to David "AmigaDave" Morris. He probably has a table arranged and may need someone to split a room with.
And it sounds like the hotel is overbooked, so it couldn't hurt to try to work something out.
Besides, I'd be willing to bet that David would love to have a demo X5000 set up.
Heck, makes me wish I could attend.

Quote:

See my reply to Acill in the other thread regarding his sharing a room and attending AmiWest. I'm certain we can work something out.

Correction:

>> the x16 slot only has [...] connected [...] 4 lanes. But the data sheet [...] speaks
>> about many controllers but *not available simultaneously* (depends on HW
>> configuration perhaps?), so with the known other features of the Tabor, does
>> x4 lanes for graphics sound reasonable?

> Yes, it sounds reasonable to me that one of the three PCIe controllers uses
> four of the six SerDes lanes.

I don't know what I was thinking two years ago when I answered that question, but of course it doesn't sound reasonable at all to use 4 out of 6 available SerDes lanes for PCIe, and also use 2x SATA and GbE from the P1022 which requires 3 more SerDes lanes, adding up to 7 required SerDes lanes in total.

More details:
http://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=12151&forum=11&start=39

Edit: mystery solved

[ Edited by Andreas_Wolf 15.06.2018 - 13:16 ]

Quote:

Hey, don't beat yourself up over that, I pretty much ignored Tabor until the discussions about the X5000's limitations started.
Then I thought 'wait a minute, how many lanes can the P1022 provide?'.
It doesn't add up, does it?
They could be using something that multiplexes or switches PCI-E lanes, but not having a complete set of specs on Tabor, I don't know.
Didn't we have a screenshot of that board at one time?

> They could be using something that multiplexes or switches PCI-E lanes

Of course, they could be using a PCIe bridge to multiplex 2 input lanes to 4 output lanes, but I don't think anything could be won by multiplexing or switching lanes that all go to the very same slot anyway. They could as well multiplex 2 input lanes to 16 output lanes and claim a fully connected x16 slot ;-)

> Didn't we have a screenshot of that board at one time?

A screenshot of a board? :-) There's a shot in the A-Eon press release linked to in my comment you quoted the link to above, but it doesn't help me to identify the solution to this mystery. A block diagram would come in handy.

> "In order to bundle forces, A-EON Technology considers acquiring AmigaKit."
> http://translate.google.com/translate?sl=de&tl=en&u=http://www.amiga-news.de/de/news/AN-2017-01-00060-DE.html

Has it happened yet?

"I see a filing was made in September in Australia and now reads: Accepted
Correspondent address:

Leaman Computing Ltd
3.07 Waterloo-on-Quay Apartments
28 Waterloo Quay
Pipitea
Pipitea 6011 Wellington
NEW ZEALAND

Seems to be a matched address for Trevor"
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=41775&forum=2&start=320#805757

> Same goes for MorphOS port, I have no doubts it will support more out of box
> (proper SATA and Ethernet driver [...])

Does MorphOS 3.10 support the GbE controller of the P5020? I couldn't find anything in the release notes.

So Aaron Smith over at Amiga on the Lake is trying very hard to get in touch with somone frm the MorphOS team that will give him a proven hardware configuration for the X5000 that will be compatible. He wants to sell pre configured machines with MorphOS on them and include a full licensed working system. They have yet to reply back to him. It seems like a huge miss here for them to generate some sales. if anyone can help out and get them in touch it would be great to see!!!

Well, you know how to get in touch with me, and he's free to email me. But in general a good combinatio of hardware would be a Radeon HD 6450 or Radeon R5 230 on the graphics side, and a Realtek 8168/8169/8111/whichever on the network side.

The two Radeons are basically the same card, just rebranded. Both passively cooled (yay) and both available brand new from stores for 40-50 euro.

Unfortunately for the Realtek card, there's a hundred different variants of this card, and I can't guarantee that they'll all work. However, if he does manage to get hold of a card that doesn't work, then we can debug it remotely.

>> Does MorphOS 3.10 support the GbE controller of the P5020?

> [...] a Realtek 8168/8169/8111/whichever on the network side

I take this as a no ;-)

I really hope the Radeon 7750 is supported at some point. I'd hate to have to downgrade to a worse GPU (paying for even more hardware) to support an OS that was released to be compatible with the X5000. The fleet of current X5000 owners have the 7750. Very frustrating, though I do appreciate all the developers' efforts at building a compatible OS.

I know how difficult it is to support the many wide range of Radeon chipsets. Hans de Ruiter spends many months writing driver code full time for A-EON. AMD's release schedule of new chipsets is continuous and this has forced us to move to Polaris for the next forthcoming driver release.

Credit where credit is due: Bigfoot has done remarkably well here to support the list of cards published in the 3.10 release notes whilst juggling the other development work associated with this large release. We are very grateful for the help and support both he and Frank have extended to us with the X5000 product.

Quote:

IMHO, supporting the integrated network controller would be a huge waste of time (if it couldn’t be done very quick and easy thanks to some open code/documentation or whatever). P5020 is a dead-end, in very limited numbers, and AeonKit puts 8139 cards in the systems anyway (though 100 megabit only). However, they did it on the Sam460, so...