AMD Confirms Key "Summit Ridge" Specs

[patrickjp93]

5 hours ago, Bouzoo said:

Exactly, INTEL RECALL. Tell me, when was the last AMD recall that you can remember? I can't remember if they had any, so statistically Intel had more. What would happen if it would happen. What would happen if AMD has no faults on their upcoming CPUs? Now look at this, nothing. Jesus Christ, you're just fantasising about things against AMD for the sole reason of doing it. 

Barcelona.

[Bouzoo]

12 minutes ago, patrickjp93 said:

Barcelona.

Correct me if I'm wrong but that was not a recall, it was "only" a bug. They stopped the shipments and fixed it with the next batch. 

[patrickjp93]

4 minutes ago, Bouzoo said:

Correct me if I'm wrong but that was not a recall, it was "only" a bug. They stopped the shipments and fixed it with the next batch. 

The only difference is time of discovery.

[shdowhunt60]

2 hours ago, SuperCookie78 said:

This is just like AM1. This means we don't have to change our mobos because of chipset reasons. 

This actually sounds pretty damn nice. It means that Carrizo can be a reasonable investment while one waits for Zen and same goes for whatever Zen+ is.

[Energycore]

6 hours ago, zMeul said:

plus, it adds multiple failure points in the CPU - un fucking believable 

The END is coming!

 

*points a silence effect at doomsayer*

[Energycore]

Jokes aside, explains the need for additional (>1300) pins on the socket.

[Doobeedoo]

Good Zen is nearing. Hopefully to see great frequency cause there sure is ton of potential.

[Blake]

 

8 hours ago, zMeul said:

yes, I know about that - dumb move

why do you think they reverted back their design .. eh!?

Why don't you tell me, you are clearly more knowledgeable then the Intel engineers. /s

 

Seriously, all your doing is arguing/having a hissy fit over speculation and assumptions.

[Blake]

3 hours ago, patrickjp93 said:

The only difference is time of discovery.

No. Stop making crap up.

By that logic your implying that intel also recalls a shit tonne of products, you know the engineering samples that are not for public release?

[zMeul]

7 hours ago, patrickjp93 said:

Barcelona.

wasn't there a recall for the instruction set bug discovered with BSD, in 2012 (ish) ? or they fixed with a micro-code?

[patrickjp93]

4 hours ago, Blake said:

No. Stop making crap up.

By that logic your implying that intel also recalls a shit tonne of products, you know the engineering samples that are not for public release?

No, because those are free samples, and they fully function anyway. Withdrawing products from sales channels due to defects is a recall. A rose by any other name is just as sweet. AMD got lucky with timing and that's it.

[shdowhunt60]

7 hours ago, Doobeedoo said:

Good Zen is nearing. Hopefully to see great frequency cause there sure is ton of potential.

It sounds like the architecture is up to the task, but I question if Samsung's LPP will be up to it. If it goes to 4.2ghz, I'd be happy.

[Paragon_X]

I wonder if we will see Thunderbolt on zen and new APUs.

[RexinOridle]

Well, a lot of people are upset because you won't have the choice of your chipset but the same thing happened when memory controller, Northbridge etc was moved onto the CPU. What this really means that you won't have to worry about motherboard traces, thus latency, and PCI-E, SATA, USB will have direct and open connection (more bandwidth) to the CPU, which should get rid of bottleneck. I'm looking forward to 8 PCI-E 3.0 X16 running all at the same time.

I think SATA is dying and USB-C is taking off. I think the future of connection will be types of plug n play PCI-E.

[Eroda]

since its on die i dont think AMD will at least to start with have multiple "chipsets" unless they made that entirely modular as well allowing them to laser cut off features if not then different features will require to much work right?? they might just have 1 for CPU and 1 for APU  but who knows how this will all shake out

[Helly]

Wow feels like all these AMD topics are just becoming all about reacting to whatever zMeul has to bitch about. Can we stop doing that please? Makes these topics really annoying to read.

 

It all sounds like a good idea to me. It just means you need a better CPU to get more pci-e lanes. Intel has that already with their top I7's having 40 lanes and the lower ones have less. I believe in AMD mostly because of what was said earlier, they mess this up and they're history.

 

These days i'm also mostly hoping AMD kicks ass just to see what zMeul will bitch about when it's actually Intel/Nvidia lagging behind..... seriously dude, just stop reacting to these topics. We know you hate AMD you don't have to tell us every time a news post pops up...

[Doobeedoo]

3 hours ago, shdowhunt60 said:

It sounds like the architecture is up to the task, but I question if Samsung's LPP will be up to it. If it goes to 4.2ghz, I'd be happy.

Yeah I see it at 4Ghz+ too. At least for quad core. 

[Notional]

7 minutes ago, Helly said:

Wow feels like all these AMD topics are just becoming all about reacting to whatever zMeul has to bitch about. Can we stop doing that please? Makes these topics really annoying to read.

 

It all sounds like a good idea to me. It just means you need a better CPU to get more pci-e lanes. Intel has that already with their top I7's having 40 lanes and the lower ones have less. I believe in AMD mostly because of what was said earlier, they mess this up and they're history.

 

These days i'm also mostly hoping AMD kicks ass just to see what zMeul will bitch about when it's actually Intel/Nvidia lagging behind..... seriously dude, just stop reacting to these topics. We know you hate AMD you don't have to tell us every time a news post pops up...

Completely agree. The problem is that uninformed people might believe @zMeul's FUD propaganda. The entire point of a forum like this, is to get informed, not misinformed, so people have to shoot down his nonsense to prevent uninformation to spread.

 

As for Intel/NVidia, they both seem to be stuck in a rut, and not really innovating properly anymore, which is a shame.

[shdowhunt60]

8 hours ago, Doobeedoo said:

Yeah I see it at 4Ghz+ too. At least for quad core. 

For the quad? I was expecting for the octa.

[Energycore]

As long as AMD doesn't pull an asshole move like including only 16 PCI-E lanes on budget CPUs or even going down to 12 / 8 I think it will be fine.

 

Moving the chipset to the CPU die does decrease the motherboard product stack. It's good that consumers don't have to look at 45 models anymore, but it's also going to be an interesting change for board manufacturers. Perhaps they also win out because they don't have to design that many SKUs.

[Notional]

1 minute ago, Energycore said:

As long as AMD doesn't pull an asshole move like including only 16 PCI-E lanes on budget CPUs or even going down to 12 / 8 I think it will be fine.

 

Moving the chipset to the CPU die does decrease the motherboard product stack. It's good that consumers don't have to look at 45 models anymore, but it's also going to be an interesting change for board manufacturers. Perhaps they also win out because they don't have to design that many SKUs.

Well a lot of high end motherboards already has specific controllers on them to add features, like extra SATA ports, NIC's, etc. There is no problem for motherboard vendors to do the same on these AM4 boards. This approach should mean we will get faster systems and better boards.

[Energycore]

1 minute ago, Notional said:

Well a lot of high end motherboards already has specific controllers on them to add features, like extra SATA ports, NIC's, etc. There is no problem for motherboard vendors to do the same on these AM4 boards. This approach should mean we will get faster systems and better boards.

Will that also mean it's going to be harder to identify whether a motherboard has a specific feature set? For instance whether it supports RAID, how many SATA ports it has, etc.

Cuz before we were at least guaranteed the features that came with the mobo's chipset.

[Notional]

3 minutes ago, Energycore said:

Will that also mean it's going to be harder to identify whether a motherboard has a specific feature set? For instance whether it supports RAID, how many SATA ports it has, etc.

Cuz before we were at least guaranteed the features that came with the mobo's chipset.

I guess it depends if the feature set is defined properly on the CPU itself.

 

However I'm sure we will see a proper spec list on the individual boards. If the board has 8 sata ports, I'm sure it's guaranteed they all work at the same time. It's just a question whether they use the internal southbridge or a dedicated controller on the board. Shouldn't matter that much in the end though. I guess we will find out soon enough.

[DevilishBooster]

4 minutes ago, Energycore said:

But how else are we going to ensure a non-biased recommendation to the people who need one?

By doing what lots of people do on the forum already and just ignore him. These days I don't even really read anything he types. If I see his avatar or name somewhere I just automatically skip over whatever it is. And if people are ever asking for input just give your opinion and back it up with fact if at all possible. Just trust me when I say that no good will come of trying to debate with zMeul on the pros/cons of AMD products. 

 

If you want to talk about this further, I suggest we take it off thread to a PM.

[Doobeedoo]

1 hour ago, shdowhunt60 said:

For the quad? I was expecting for the octa.

That would be amazing for stock, though we can't know how clocked it will it be. But OC for sure.