AMD's ZEN-Supporting X370 Motherboards to be Shown at "New Horizon" Event

[Raevenlord]

No SR2? Commander Shepard is pissed!

I always think of that when I read through those code names

I signed up online to watch the event, might be interesting.

Indeed. But I fear they will focus a lot on gaming and such, and not on product specifics. Here's hoping I'm proven wrong.

 

[Parn]

Why SATA Express in 2017? Are there any SSDs that even use this interface?

 

[thesmokingman]

Board looks like a pre production havent seen so many jumpers in a long while.

This. I almost thought it was the 90s again there were so many jumpers in one place, lol.

Does the new platform offer more PCIe lanes than intel does? Are motherboards going to offer M2 ports that don't require sacrificing two SATA ports to use? Cause that's what important, not simply regurgitating marketing slides.

AMD never got into monetizing pcie lanes. Let's hope it stays that way!

 

[R-T-B]

WOWOWEO AMD HYPE TRAIN MOAAAR HYPE. Super duper event , mega presentation pictures alll
like have to come supernova cpu smash whole intel to hell..
update: intel cut all cpu s prices 40% DOWN iaam so scary from s7 s7 uuuu

You have hurt my brain. Please... stahp.

 

[X4K4]

jumper switch realy now -.-

 

[TheinsanegamerN]

So that would be two more lanes than Skylake, but older version for all of them? That seems to be an old chipset, is it still relevant?

It is very old, yes. The 990fx was effectively a tweaked 890fx, which was a tweaked 790fx. Each generation tweaked the chipset and added new features, and a new southbridge, but the AMD chipset for FX hasnt seen a major overhaul since the days of the OG phenom.

Relevancy is hard to determine. it still works fine, and since dual GPUs run in x16/x16 mode, you are not loosing any bandwidth compared to non-HDTE skylake. Modern versions have things like M.2 NVMe boot and the sabertooth managed PCIE 3.0 somehow. It has been a very long lived, surprisingly useful chipset design.

However the design is definitely dated, meaning no micro ATX or mini ITX boards, due to cooling requirements and the need of a southbridge.

Zen will be the first major overhaul of the AMD chipset in nearly a decade. Given how well their old one held up, the new one should be fantastic, as long as they dont cheap out on it.

 

[TheGuruStud]

To confuse consumers. AMD's name is tarnished, people see the AMD logo and assume it's trash. But if they see a single letter and the number "7" they assume it's an i7, which means it's good, because marketing. The confusing chipset and CPU naming scheme is being done to try and get users to buy AMD instead of intel, and defeat this notion that AMD=Bad.

It's like that fake design philosophy from apple. "make everything simple to use, because 90% of your consumers are f$%^!(g idiots". Most consumers dont think at all about what they are buying, and AMD is trying to get them to blindly buy their chips instead.

If the customer doesn't know the difference between them, then it doesn't matter. They won't care as long as it works. For them it's not much different than " Bob has an Apple phone and he said they're the best, so I need to get that one." Whether the statement is true or not and regardless of price, the dummy will buy the Apple product. Intel was parroted as superior by idiots and shills even when they had their asses handed to them.

If you think using a couple of the same numbers is deceptive, then you better be ranting about Intel and their decades of illegal business practices that has brought us to the sad state of CPU performance.

 

[TheHunter]

To confuse consumers. AMD's name is tarnished, people see the AMD logo and assume it's trash. ...

Yeah, Bulldozer left a scar by few..

But gimmick names or not, the chip actually seems capable this time.. I think intel will get surprised how good it is.

350€ or 450€ BE 8core vs their 1000€ same if not even better perf., hello

 

[bug]

If you think using a couple of the same numbers is deceptive, then you better be ranting about Intel and their decades of illegal business practices that has brought us to the sad state of CPU performance.

The thing is, comparing 3, 5 and 7 against 3, 5 and 7 is actually the least confusing way to go about it. But that's assuming they are comparable and AMD doesn't pull an RX480 vs GTX1080 again.

 

[Camm]

The thing is, comparing 3, 5 and 7 against 3, 5 and 7 is actually the least confusing way to go about it. But that's assuming they are comparable and AMD doesn't pull an RX480 vs GTX1080 again.

Chicken or egg? Both have used both nomenclatures for most of their cards in varying performance segmants going back a good decade and more.

 

[Melvis]

What was it? Competing against intel Sandy Bridge CPU's?....

The 8350 already did that

 

[Grings]

The thing is, comparing 3, 5 and 7 against 3, 5 and 7 is actually the least confusing way to go about it. But that's assuming they are comparable and AMD doesn't pull an RX480 vs GTX1080 again.

AMD had been using x90 not x80 to indicate their higher end cards since the 280 was discontinued, the 480 is just the replacement for the 380, a card which was made to compete with the gtx960

you really think the £200 rx480 was designed to go head to head with the £600 1080?

 

[AsRock]

Board looks like a pre production havent seen so many jumpers in a long while.

yeah really whats with that ?, all so they seem to think we need a guide to know which way up a jumper goes too.

 

[Frick]

Yeah, Bulldozer left a scar by few..

Not outside the enthusiasts and gamers. What did them in there were the slow low end APU laptops and the terrible TDP of the bigger models. You saw a cheap HP laptop with an AMD sticker on it? Be prepared for it being completely unusable, partly because the 1.3GHz CPU and partly because of bloatware that suckled up precious CPU time.

 

[R-T-B]

The 8350 already did that

Amd has yet to beat even nehalem in raw IPC... Let alone Sandy Bridge.

Zen will hopefully bring better parity there.

 

[Melvis]

Amd has yet to beat even nehalem in raw IPC... Let alone Sandy Bridge.

Zen will hopefully bring better parity there.

IPC doesnt mean everything when it comes to overall performance http://www.guru3d.com/articles-pages/amd-fx-8350-processor-review,1.html
http://www.guru3d.com/articles-pages/amd-fx-8370-and-8370e-processor-review,1.html

and Zen will be double the 8350 and seems to be on par with the 6900K

 

[P4-630]

The 8350 already did that

Ok, Ivy Bridge then...

 

[bug]

IPC doesnt mean everything when it comes to overall performance http://www.guru3d.com/articles-pages/amd-fx-8350-processor-review,1.html
http://www.guru3d.com/articles-pages/amd-fx-8370-and-8370e-processor-review,1.html

and Zen will be double the 8350 and seems to be on par with the 6900K

It kind of does. Overall performance is IPCxCoreCount. And since few applications scale well beyond 2-4 cores, it usually comes down to IPC. That's literally the only reason AMD have had their asses handed to them for so many years: their IPC has been stuck in A64 era.

 

[Melvis]

It kind of does. Overall performance is IPCxCoreCount. And since few applications scale well beyond 2-4 cores, it usually comes down to IPC. That's literally the only reason AMD have had their asses handed to them for so many years: their IPC has been stuck in A64 era.

But thats not really AMD's fault, its developers been lazy and not address and or unlocking the performance that can be there if used correctly.

Once you find programs that use cores correctly you be surprised how fast a 8 core can really be. I have been playing DOOM this past few weeks on Vulkan and it uses all 8 cores really well so much I got a 30FPS boost.

Dont get me wrong I understand where your coming from and its true that IPC is king in the past but when a bit of software address core count and scales it correctly the IPC gains become alot smaller.

 

[bug]

But thats not really AMD's fault, its developers been lazy and not address and or unlocking the performance that can be there if used correctly.

That's the exact argument intel used to defend Pentium4. If you wrote code specifically for it, Pentium4 could be faster than AthlonXP.
Also, a program that prints numbers from 1 to 1000 simply cannot be multithreaded.

 

[Brusfantomet]

It kind of does. Overall performance is IPCxCoreCount. And since few applications scale well beyond 2-4 cores, it usually comes down to IPC. That's literally the only reason AMD have had their asses handed to them for so many years: their IPC has been stuck in A64 era.

Only thing is, its not IPC x CoreCount, its IPC x CoreCount x Clockspeed, otherwise a stock i7 5820 and i7 5930 would be exactly the same performance.

 

[Melvis]

That's the exact argument intel used to defend Pentium4. If you wrote code specifically for it, Pentium4 could be faster than AthlonXP.
Also, a program that prints numbers from 1 to 1000 simply cannot be multithreaded.

Huh? im guessing your talking about HT? because back in those days pretty much everything was coded for intel CPU's, its only been in very resent yrs that AMD has been more recognized and coded for. Just shows how slow the P4 really was back then and HT well it only gives at most 25% extra performance over a normal core.

Athlon XP had to have a windows patched for it to work correctly back then (Windwos XP) and the FX (bulldozer/piledriver) had to get 2 patches in Windows 7 for them to work correctly also, and without the Patch the FX's run super slow. (Must be Patched Manually)

 

[Enterprise24]

IPC doesnt mean everything when it comes to overall performance http://www.guru3d.com/articles-pages/amd-fx-8350-processor-review,1.html
http://www.guru3d.com/articles-pages/amd-fx-8370-and-8370e-processor-review,1.html

and Zen will be double the 8350 and seems to be on par with the 6900K

8350 need 4.7ghz to compete with i7-950 @ 4ghz in gaming. So they need to beat Nehalem first. Let's alone Sandy Bridge.

 

[bug]

Huh? im guessing your talking about HT? because back in those days pretty much everything was coded for intel CPU's, its only been in very resent yrs that AMD has been more recognized and coded for. Just shows how slow the P4 really was back then and HT well it only gives at most 25% extra performance over a normal core.

Athlon XP had to have a windows patched for it to work correctly back then (Windwos XP) and the FX (bulldozer/piledriver) had to get 2 patches in Windows 7 for them to work correctly also, and without the Patch the FX's run super slow. (Must be Patched Manually)

That's not what I meant. Similar to AMD's GCN needing async compute to be able to use all the hardware simultaneously, P4 required special compilation flags to avoid emptying their long pipeline too often.
Of course, the industry look at those requirements and gave intel the finger.

In both cases, chip manufacturers think developers must adapt to their silicon and ignore the existing software.

 

[Melvis]

8350 need 4.7ghz to compete with i7-950 @ 4ghz in gaming. So they need to beat Nehalem first. Let's alone Sandy Bridge.

Claps very slowly, guess you havent been reading the posts have you? and there is more to computer performance then just gaming.

Also you should look in my specs and you might learn something