Quote:

Since the announcement was authored by A-Eon in Cardiff, I'm guessing that it was Matt who wrote it. I don't know why you dislike Matt, or why you think customer service from AmigaKit for North America customers (or any customers for that matter), is not good, as I have spent many thousands of dollars buying all kinds of Amiga gear from AmigaKit over the years, and have almost never had any problems. The couple of mistakes that were made, were quickly taken care of, in the most professional and courteous manner I have experienced from any tech company.

I would not be surprised if there are a few AmigaKit customers who don't share my feelings, or views on AmigaKit's customer service, but when I think of how many unreasonable and childish Amiga users I have seen on many different Amiga related forum sites, it is no wonder to me that Matt may have had some less than optimal interactions with a few of them (I won't name any names, to avoid stirring up a hornet's nest). If Matt has been less patient with some customers, due to his usual calm demeanor being worn thin from time to time, I think that any of us would probably react the same, or worse, under similar circumstances. Just imagine trying to satisfy everyone in all Amiga related camps, without making anyone unhappy, and not making any enemies? Probably not possible for any of us.

I'm not saying Matt or Trevor are saints, but having met both men several times, they seem like regular good guys, with genuine passion for the Amiga, and desire to help this community as best they can.

Quote:

AeonKit...

Quote:

Yes, that might be apt name for it.

@ David,

I've only had a little interaction with Matt and company.

I do know they sold me a Cocolino PS/2 mouse adapter that draws to much power to be used in the applications I have tried it in, and then misquoted my complaints about the matter on one of the forums they regularly post on.

And returning a relatively low cost item to a company in Great Britain is a exercise in futility as the postage will eat up a large percentage of the original price.

Also, while they do support the community, they have control over the distribution of Aeon hardware and the support and margins offered to the resellers.
And I'm not comfortable with that or the casual way the company does business, its too much like something run by a hobbyist, for hobbyists, rather than a competitive business.
I know our market is small, and we should be grateful for the support we get, but I just expect more professionalism from those I buy from.

That said, I can live with it, and if it helps Trevor's endeavors...I'll just leave it be.

Either way, I can, once again, purchase from Aaron, so I'm cool.

Quote:

I wasn't referring to you, in my earlier post. If your complaints about the Cocolino were made in a forum, instead of directly to customer service, then I would say that was your first mistake. The fact that they misquoted you (IMO) is no big deal, as that kind of thing happens all the time, and miscommunication can be cleared up quickly by contacting the other party directly. I'm not going to get into trying to figure out who was right, or wrong in your dispute with AmigaKit over your Cocolino PS2 mouse, and it is irrelevant to your point that having a USA distributor is better than having to ship something across the Atlantic for warranty repairs or replacement. IMO, it never resolves anything to discuss complaints about a company in a public forum, until all other attempts to resolve the problem via direct communications have failed several times. Maybe you did, or maybe AmigaKit brought up your complaint in the forum, not you (but somehow I doubt it).

As far as I know, AmigaKit does not manufacture the Cocolino PS2 mouse, they only distribute/re-sell them. If it were me, my first attempt to get resolution with the product would be from the manufacturer of the item, not the reseller, specially if the product partially works, but perhaps not in the way you expected, or needed for your particular instance. Many others, including perhaps yourself, may opt for asking the reseller/distributor first, to resolve warranty issues, which is okay too, just not the way I do things usually.

Unless things have changed recently at AmigaKit, I think they respond quickly to customer service questions, and requests, though I have heard of some strange occurrences where one or two customers have waited long periods of time, without a resolution, but I also suspect that a lack of communication is to blame in that instance.

I'm not trying to defend AmigaKit, or Matthew, I'm only relating my many experiences with them, and my personal opinion, which is based on how I have been treated by AmigaKit staff, and well over 90% of all AmigaKit customers I have heard from, in any form. Given that their service to me has been so good, for so long, it is hard for me to imagine that they treat other customers any differently.

Don't sweat it Dave, I got over it.
And I wasn't really directing an inquiry to them on the forum I mentioned, I was warning another user about the issue I'd experienced.
Since they have staff acting as mods they chimed in about it, and misstated the issue I'd had (even though I'd communicated it to them directly).
I usually expect a vendor I buy from to be my first source of assistance in problem solving, but its not a issue related to anything Matt or his staff can work on (its a current supply issue related to the design).

Anyway, no big deal, I'm saving the Cocolino for my A2000 if I ever upgrade that and start using it again.

At least with Aaron, returning hardware will be a practical option.

BTW - Its beginning to sound like this year's AmiWest could shape up to be more interesting than usual.
Possible release of the A1222, other interesting stuff...
I tried to clear the time, but I'm going to be training for a new position during the Fall, so I can't count on that time off.
Otherwise I'd consider buying an X5000/20 motherboard and assembling a system for show exhibition.

Even if MorphOS 3.10 isn't released by then, it could be possible to arrange for a beta copy (after all, Pampers had a copy for evaluation), and Bigfoot might even have a beta Radeon HD driver available by then. Mark's on a business trip right now, but the last time I exchanged messages with him it sounded like he was progressing quite nicely.

Then again, like everything, 3.10 and the accelerated video drivers...done when they're done, eh?

Anyway, don't take my grumbling for a sour attitude, I always have high expectations for everyone (including myself), and it makes me appear more gruff than I actually am.

> Otherwise I'd consider buying an X5000/20 motherboard

Tired of waiting for the X5000/40? :-)

Quote:

Well, its hard to justify buying an X5000/20 when our current hardware seems just as capable (outside of the issue of video card support).
Were we to have PCIe G5 support, I'm not sure I could even justify the wait.

Quote:

I'd say if we ever get support for the quad G5 w/PCIe HD video card drivers fully supporting 3D, the cost of the 5040 X5000 would be very hard to justify. What do you get? It's new, it has newer SATA hardware (AmigaOS4.1FE SATA driver has at least 1 person complaining or questioning how good it is), it might have faster memory (haven't looked at the exact numbers), and the X5000 will be capable of running both AmigaOS4.1FE, and MorphOS3.10.

If you don't mind paying the price to buy an X5000 for those few advantages, after the MorphOS Dev. Team supports the PCIe G5, (and I think you will, given your fondness for PPC hardware), then go for it.

I doubt that the MorphOS Dev. Team will use their limited time to support the quad G5, UNLESS they want to offer MorphOS users the ability to run a cheap and fast system, which also supports PCIe video cards. The SAM460 is hard to find for sale, and not really cheap, considering the performance you get for your money. If there is no way to shut off, or put to sleep, the unused 3 extra G5 cores, running MorphOS3.10 on the quad G5 will be very energy inefficient.

I wonder if Bigfoot is using an X5000 to write the new RadeonHD drivers, or only using the X5000 to test the driver once it is done?

[ Edited by amigadave 11.07.2017 - 10:38 ]

> What do you get?

PCIe v2 (vs. PCIe v1).

> it might have faster memory (haven't looked at the exact numbers)

See comment #666.

> If there is no way to shut off, or put to sleep, the unused 3 extra G5 cores,
> running MorphOS3.10 on the quad G5 will be very energy inefficient.

Yes, that's why the lower clocked dual-core models of the PCIe-based PowerMac G5 would surely be worth considering.

Quote:

I never thought to ask. He has a SAM 460 as well as the X5000.
The only thing he did mention was that some of the longer cards were a challenge to fit in the X5000's case.

[ Edited by Jim 12.07.2017 - 13:30 ]

Quote:

When I spoke to Aaron last week he said that this was directly between him and Trevor. Aaron doesn't have much nice to say about Matthew at all from our conversations. Apparently the final straw was Matthew's demand that distributors not selling in another's area then turning around and selling the X5000 to the US for less than the US distributor.

Quote:

8x PCIe v2 lanes = 16x PCIe v1 lanes. So it's the same bandwidth, right?

Speaking of bandwidth, looking at this document (page 11)...

https://amigaone.files.wordpress.com/2017/06/trm_cyrus_1-1-1_aeon.pdf

...does this mean that the PCIe 1 slot (I assume that slot 1 is the 16x slot) only has four dedicated lanes, and four switched lanes, shared with the other PCIe slots (i.e. the 16x, 4x, 1x, 1x, and "Xorro" slots)?

> 8x PCIe v2 lanes = 16x PCIe v1 lanes. So it's the same bandwidth, right?

Right, but while in theory there shouldn't be compatibility problems, some PCIe v3 cards have been reported to not work in PCIe v1 slots.

> the PCIe 1 slot [...] only has four dedicated lanes

Yes, the x16 slot is connected by only four lanes, as was announced last year.

> and four switched lanes, shared with the other PCIe slots
> (i.e. the 16x, 4x, 1x, 1x, and "Xorro" slots)?

According to the block diagram on page 8, the four switched lanes have nothing to do with the x16 slot, but are shared via PCIe-PCIe switch between x4, x1, x1 and Xorro slots as well as PCIe-PCI switch (for two PCI slots).

Quote:

So the X5000 PCIe x16 slot only has 4 PCIe v2 lanes connected, effectively meaning 1/2 of the theoretical bandwidth of the x16 PCIe v1 slot of the PowerMac...?


Quote:

Ah, I managed to scroll past that block diagram!

Quote:

Unless you really want the Boing Ball engraved case, I would suggest that you only buy the motherboard for the X5000, not a complete system, to save money, and to get yourself a case that can accept all lengths of video cards. Getting a complete system that has the OS already loaded and tested on the hard drive or SSD is appealing, but you are more than capable of assembling and setting up an X5000 system for yourself, and can get exactly what you want, instead of getting what the retail seller wants to put together for you.

>Right, but while in theory there shouldn't be compatibility problems, some PCIe v3 cards have been reported to not work in PCIe v1 slots.

Its more complicated than that.
I haven't finished discussing this with AMD, and using later cards in a v1 slot will probably need some experimentation.
As I've said before, the Radeon HD 6570 will work in a G5.

BUT, there actually may be problems using v3 cards in v2 slots as well.

This is a partially quote from a message I received from AMD tech support today, "I believe that the HD7000 series GCN1 GPUs should work with the older PCI-E 2.0 revision Motherboards, however you may encounter issues if you attempt to use anything newer as we have had reports from customers that had compatibility issues when using the new RX400 and 500 series with PCI-E 2.0 motherboards."

SO, NXP lists the PCIe controllers in the P50xx as PCIe 2.0/3.0, but what has Aeon/Varisys implented?
If those slots ARE only v2, they may not work with future AMD based video cards.

AND, does AMD's statement about v2 apply to v1?
If so, the G5 could potentially support the 7000 series cards, if not the 6000 is pretty good (and probably the best we are likely to get support for in the next driver upgrade anyway).

Its a story that is continuing to evolve, but I personally am moving forward on putting together an X5000 (although I do have a Quad 2.5 GHz G5 sitting in the wings).

> So the X5000 PCIe x16 slot only has 4 PCIe v2 lanes connected,
> effectively meaning 1/2 of the theoretical bandwidth of the
> x16 PCIe v1 slot of the PowerMac...?

Yes, exactly. And also half the bandwidth of the X1000 PCIe v1 x16 slot.

'theoretical bandwidth'...hmm, in the real world, not terribly significant.

Again, what is worrying me more is will these boards work with later video cards?
So far, no one has even shown one working with a GCN gen2 card.

> NXP lists the PCIe controllers in the P50xx as PCIe 2.0/3.0

They list it as "2.0/3.0" in one place and as "1.1/2.0" in others (like fact sheet and product brief). I firmly believe that "3.0" is not correct.

> what has Aeon/Varisys implented?

Their press release linked to above clearly states "Gen2".

> does AMD's statement about v2 apply to v1?

I think v3 cards not working in v2 slots also won't work in v1 slots, and that v3 cards working in v2 slots not necessarily work in v1 slots.

>I think v3 cards not working in v2 slots also won't work in v1 slots, and that v3 cards working in v2 slots not necessarily work in v1 slots.

Seems probable.

So the G5 likely tops out at the 6000 series, whereas the X5000 supports the 7000 series, most of the R5, 7, and 9 cards, and possibly some higher cards.

Sam460 is PCIe v1.1 and the X1000 is PCIe V1.x (unsure which rev) and both work with 7000 and RX series cards. A Sam440 with a PCIe-PCI bridge card tops out with the 6000 series but it's suggested that the limitation is in Uboot rather than the bridge.

Quote:

Thanks Rob,
We've been tossing this around, and it seems likely that AMD's tech support staff is just throwing around qualifiers out of over caution.

I haven't been able to find any logical reason that even v3 (or v4) PCI-E cards shouldn't work in a v1 slot.
The SAM440 isn't supported by MorphOS (only OS4), and PCI-E to PCI bridge components (at least for video cards) aren't supported yet (unless one considers to PCI-E to AGP bridge chip used in some of the cards we support, and that's not the same thing AFAIK).

> both work with [...] RX series cards.

Really? ;-)

http://www.amd.com/en/rx-series

Quote:

Yeah...that's not the info I'm getting out of AMD.
What they are saying is that the RX cards may very well NOT work with older slots, and that they have customers reporting issues using this combination.

And a quick search of the internet does seem to confirm that some users are having problems.
Whether its a hardware or firmware issue isn't easy to figure out, but it does look like we're going to have to work this out on our own.

In any case, everything looks good up to GCN gen1, which we don't even have support of yet.
So our concern is probably premature.

AND AMD's conditional statements could just be cover for an inability to address all possible combinations of older and newer hardware.
After all, they threw in qualifiers for pre v2.0 slots when used with GCN cards, a combination that seems to be working.

Well here's to the hope of MorphOS being released soon for the X5000. I just had an order placed to build me one! Fingers crossed its as nice as my G5 is when I get it next week some time.