[Rumor] AMD's Plans For Carrizo, Devastating if True

[patrickjp93]

The original sources are PC Watch and bitsandchips.it.

 

http://wccftech.com/amd-carrizo-apu-targetting-mobile-kaveri-power-desktop-side-2016/

 

The rumor across the pond is AMD is planning to keep Carrizo mobile only until 2016. That could easily mean Zen will not arrive until late 2016 or early 2017. As I said: devastating if true.

 

My response: WTF AMD?! Every time I think they might turn around and make good business decisions moving forward, I see something like this (which usually ends up true) where it seems the businessmen and the engineers just do not see eye to eye... This even suggests Carrizo won't be headed for servers until 2016 either. Please dear God let this rumor be wrong!

[C0LL0SS0S]

Looks like someone wants to get down to ubisoft

[Solkia]

meh.

[Sharkyx1]

I would hope that everyone knows if this is true its because 
AMD has to delay Zen, not because it enjoys being behind, 

 

Looks like someone wants to get down to ubisoft

No me-mes pls

 

keep ubi shitposting to ubi threads.

[Master Disaster]

What's the big deal, so AMDs next apu will be late, it will be two generations behind Intel no matter when its released.

AMD are out of the CPU race and have been for some time.

[patrickjp93]

I would hope that everyone knows if this is true its because 

AMD has to delay Zen, not because it enjoys being behind.

Which will put Zen up against Cannonlake, 14nm against 10nm, AMD's HSA 1.0 with HBM against Intel's completed version (equal-citizen scheduling is inbound). At that point it's going to be hexacore Intel's on the Z-series platform with Gen 9/10 graphics and 96+ EUs. 

[ComradeHX]

Maybe they will do another refresh of FX series in the mean time...

[patrickjp93]

Maybe they will do another refresh of FX series in the mean time...

Keller and Co. have already said that's not happening until Zen. 

[holytoledo]

This is frustrating. So what's the down-low on AMD? Why are they so far behind?

[patrickjp93]

This is frustrating. So what's the down-low on AMD? Why are they so far behind?

You mean apart from 6 years of bad executive management, paying 1.4x what they should have in the ATI buyout, not having controlling interest in GloFo, and not having Intel's warchest and foundry equipment suppliers? I wrote a 20-page essay to the effect of why AMD as a company is suffering so much. These are just the bones and meat of it. The connective tissue where it all meshes together is the real brain fuck.

[Misanthrope]

Hey AMD, here's a plan that could actually work for you: Sell the entire processor division off to someone that has the money to sink into R&D like Samsung and focus on your true strength more: GPUs

[DuckDodgers]

This is frustrating. So what's the down-low on AMD? Why are they so far behind?

Too many causes to pick them all, so here are some brief notes:

They under-invested in a new CPU architecture, without even a B-plan, so they milked the old K8 core for way too long.

Their first quad-core processor was a complete dud and 6+ months late (same story with Bulldozer later).

They paid twice the real price for the merger with ATi in 2006 and later sold the mobile GPU devision for pennies to Qualcomm, oblivious to the mobile market growth potential.

The separation from the chip manufacturing facilities put them in a series of outstanding payments to the new owner, for some old debts and future obligations for production capacities.

[amdart11]

One word, WCCFTech. Take with a grain of salt.

[patrickjp93]

One word, WCCFTech. Take with a grain of salt.

Sourced from Bitsandchips Italy, far from a rumor mill/clickbait site.

[Suika]

I'd rather AMD delay a good chip then release a bad one early.

[patrickjp93]

I'd rather AMD delay a good chip then release a bad one early.

The Carrizo mobile chip engineering samples are already producing some benchmark numbers on SiSoft. It's not looking like a bad chip. This just means greater delays on Zen which will mean AMD remains a process node behind, maybe more if Samsung doesn't buy up GloFo.

[Lazmarr]

The original sources are PC Watch and bitsandchips.it.

 

http://wccftech.com/amd-carrizo-apu-targetting-mobile-kaveri-power-desktop-side-2016/

 

The rumor across the pond is AMD is planning to keep Carrizo mobile only until 2016. That could easily mean Zen will not arrive until late 2016 or early 2017. As I said: devastating if true.

 

My response: WTF AMD?! Every time I think they might turn around and make good business decisions moving forward, I see something like this (which usually ends up true) where it seems the businessmen and the engineers just do not see eye to eye... This even suggests Carrizo won't be headed for servers until 2016 either. Please dear God let this rumor be wrong!

Isn't AMD's Zen meant to be a completely different x86 architecture designed by Jim Keller and a bunch of others; if so then it'll likely still be coming in late 2015/early 2016.

[patrickjp93]

Isn't AMD's Zen meant to be a completely different x86 architecture designed by Jim Keller and a bunch of others; if so then it'll likely still be coming in late 2015/early 2016.

It's designed for 16/14nm FinFET, which Samsung isn't even done delivering to Global Foundries. Late 2016/2017 is the reasonable timeframe given Samsung themselves haven't made anything more complex than NAND flash and a controller on the stuff.

[ManWithBeard1990]

Hey AMD, here's a plan that could actually work for you: Sell the entire processor division off to someone that has the money to sink into R&D like Samsung and focus on your true strength more: GPUs

Yeah sure and rename their company to ATI while they're at it...

You DO realise AMD has a vision here, right? Intel can't keep up its manufacturing lead forever. HSA is about more than just scheduling tasks between CPU and GPU. I agree that's a nice in-between goal but it's meant to be a layer in between to execute code on truly heterogeneous devices. Not just CPUs and GPUs but FPGAs, DSP's, the lot. A lot of AMD's resources are going into this and I agree they've dropped the ball on IPC performance a bit but hey, they've much less resources to work with.

 

And about their true strength: it is having both CPUs and GPUs. Intel has both too but even with the manufacturing advantage they're trailing with GPU architecture.

[holytoledo]

You mean apart from 6 years of bad executive management, paying 1.4x what they should have in the ATI buyout, not having controlling interest in GloFo, and not having Intel's warchest and foundry equipment suppliers? I wrote a 20-page essay to the effect of why AMD as a company is suffering so much. These are just the bones and meat of it. The connective tissue where it all meshes together is the real brain fuck.

 

Whoa buddy, that sounds quite in depth. I would like to take a glance through that if you don't mind. May I see a copy?

[patrickjp93]

Whoa buddy, that sounds quite in depth. I would like to take a glance through that if you don't mind. May I see a copy?

When I have wifi again I'll happily send it to you. (Spending Thanksgiving break with housemate's family in Chicago, with no wifi in the house ;(   Boo me...)

[Misanthrope]

Yeah sure and rename their company to ATI while they're at it...

You DO realise AMD has a vision here, right? Intel can't keep up its manufacturing lead forever. HSA is about more than just scheduling tasks between CPU and GPU. I agree that's a nice in-between goal but it's meant to be a layer in between to execute code on truly heterogeneous devices. Not just CPUs and GPUs but FPGAs, DSP's, the lot. A lot of AMD's resources are going into this and I agree they've dropped the ball on IPC performance a bit but hey, they've much less resources to work with.

 

And about their true strength: it is having both CPUs and GPUs. Intel has both too but even with the manufacturing advantage they're trailing with GPU architecture.

 

Look if there was something to show me that vision has even remote chances of becoming a reality, I'd be all for it. Sad thing is that dreams take money, the bigger the dreams the more money it takes. We're talking billions here for their size of dreams. I mean it would be good for the industry for them to keep trying, but it would be even better if someone who could really threat intel like Samsung would get a jump start and really push the fuck out of the technology way faster than AMD could dream to right now.

[RazeG3SG1]

Before any WCCF tech hate coming, I doubt AMD is that stupid. But, with how they were trying so hard for shoving cores onto FX, this maybe true as well. Rumor swings both ways as usual.

[bigboy678]

Hey AMD, here's a plan that could actually work for you: Sell the entire processor division off to someone that has the money to sink into R&D like Samsung and focus on your true strength more: GPUs

 

Id love to see AMD sell itself off to Qualcomm. They have big pockets and they want to break into the server market. This would be a match made in heaven

[RexinOridle]

WTF WTF WTF WTF WTF WTF WTF AMD? WTF?