AMD: Southern Islands (7*** series) Speculation/ Rumour Thread

But, he needs to DOUBLE CONFIRM it. *SCNR*

Could very well be the case, but why didn't Intel then have PCIe 3.0 officially supported? Or can you just not validate something with something else that still isn't final/validated itself?

Intel will advertise Gen3 support when Gen3 products are available, don't sweat it. Why do you think all MB makers promote it?
 
That bears no logic. Why would, when MB makers are promoting it without available products, Intel refrain from doing just the same?
 
"If you ask our competitor, PCI-Express 3.0 is a big deal," says John Fruehe, director of product marketing for servers and workstations at AMD. "If you ask anyone else, it doesn't make a stinking difference. The important thing is that PCI-Express 3.0 forces a platform change that only benefits a few select applications. We'll be there when it is relevant. For us, it is more important to time it right than to be first to market."

http://www.theregister.co.uk/2011/11/14/amd_opteron_4200_6200_launch/

So, don't forget, PCI-E 3.0 for AMD next gen doesn't make a stinking difference.
 
I have given up competely on AMD after BD. The best hope for CPU salvation lies with ARM64.

I don't care about PCIe3 making a stink or an aroma anymore for AMD CPU.
 
Fruehe's statements don't really apply to graphics cards, not that current boards exercise the current PCIe bus enough to make it a significant concern going forward for most users.
 
Because MB makers don't make CPU's? They only make Gen3 ready MB's. Straight up Logic for me.

Sorry, that doesn't make sense. You have a CPU that supposedly is PCIe 3.0 ready, but is not advertised to it because there's a lack of validation hardware (that was the first argument). Now, you have mainboards, that in fact are advertised as PCIe 3.0 ready, but with what hardware did the vendors qualify those, if not even Intel could get their hands on something to validate?

Or do you imply that for MB makers it is enough to adhere to the specs, no further validation needed?
 
Or do you imply that for MB makers it is enough to adhere to the specs, no further validation needed?

I am not saying that this is the case. But it's quite likely that it is. Motherboard makers just have to draw traces. The controllers and PHYs are on the CPU and not on the Southbridge, for which they would have to assume responsibility.

Besides, I find the argument that there are no PCIe3 devices to validate a tad fishy.
 
Sorry, that doesn't make sense. You have a CPU that supposedly is PCIe 3.0 ready, but is not advertised to it because there's a lack of validation hardware (that was the first argument). Now, you have mainboards, that in fact are advertised as PCIe 3.0 ready, but with what hardware did the vendors qualify those,
There is plenty of validation hardware available from the usual suspects.

if not even Intel could get their hands on something to validate?
Did they officially state that somewhere?

Or do you imply that for MB makers it is enough to adhere to the specs, no further validation needed?
Nope, this needs to be lab-tested and verified for a long period.
 
Fruehe's statements don't really apply to graphics cards, not that current boards exercise the current PCIe bus enough to make it a significant concern going forward for most users.

is PCI-Express 3.0 a pure bandwidth upgrade increase in specification or are the new features supported as well? Are these new features important to only PCI 3.0 cards or is 2.1 going to benefit?
 
The primary improvements are related to bandwidth, with an increase in the transfer rate and a decrease in the overhead associated with a transfer, to arrive at doubled bandwidth.

PCIe 2.1 is described as including many of the other enhancements for 3.0, but without the bandwidth increase.
 
Bad PR more like.

Why PR?
That's simply an example of how rumours based on misinformation are being born and start to circulate. Give me a single memory chip manufacturer that is ready with mass quanitities of xdr2. If there is no one out there, then this whole thing is a meh, bs. :devilish:

Charlie @SA says that Intel confirmed PCI-E 3 on SB-E was working by using prototype AMD boards.

Sorry, but may I ask you for a link to prove it? :mrgreen:
 
Why PR?
That's simply an example of how rumours based on misinformation are being born and start to circulate. Give me a single memory chip manufacturer that is ready with mass quanitities of xdr2. If there is no one out there, then this whole thing is a meh, bs. :devilish:

It's bad PR because they just fired the guys who knew the details of 7000 series and were running the account. Newbie apparently responds/retweets everything that mentions the brand they're engaged to promote.

Sorry, but may I ask you for a link to prove it? :mrgreen:

I gave you a link, it just wasn't in URL format. ;)
 
The primary improvements are related to bandwidth, with an increase in the transfer rate and a decrease in the overhead associated with a transfer, to arrive at doubled bandwidth.

PCIe 2.1 is described as including many of the other enhancements for 3.0, but without the bandwidth increase.

So, perhaps having PCIe 3 mainboards will benefit not just the next gen of PCIe add-in boards, but what about previously released AIB's that under 2.0 boards couldn't use all their features? PCIe 3 is a superset of 2.0, so 2.1 cards will have features available under 3.0 boards that weren't accessible under 2.0 boards.

If you sold millions of PCIe 2.1 devices on the basis that PCIe 2.1 was better than 2.0, wouldn't you want to be consistent in being supportive of the next superset that unlocks those features that all the time and money went in to providing?

Conversely, doesn't this also show that PCIe 2.1 is nothing to anyone, never more than most whimsical fleeting marketing fancy talk and there only to decorate the box with bigger numbers? Which is it, is PCIe 3 needed or not? If it's not needed why not spend the resources on something else like expanding the dev/rel program so skyrim, bf3, rage etc. all get working drivers from day 1 instead of 3 months late?
 
(edit: disregard this first sentence, misread last sentence above )Devrel to get PCIe 3.0 working for game or to use low level management features that mean nothing to a game?

Fruehe's position is on the server side, he makes no statements concerning the client x86 side, and his involvement in servers gives him no authority to speak for the graphics group.

PCIe 2.0 versus 1.0 is barely relevant for most of the market for AIB graphics products.
Unless a x16 slot is being split for multi-card setups and this may require trifire to even be noticed, it is not significant.
With the doubled rate in 2.0, the saturation point is even further away. If AMD cards support 2.1, it may be so that they match the broadest range of platforms possible on their marketing checkboxes.
I have not seen anyone note a difference switching from 2.0 to 2.1, and anecdotal evidence that so far it means nothing.
 
Last edited by a moderator:
Back
Top