Kaby Lake information.

[cytg111]

Way to miss the point.

Dude you got owned.. fess up

 

[Fjodor2001]

Dude you got owned.. fess up

:thumbsup:

 

[frozentundra123456]

Dude you got owned.. fess up

Actually he *did* miss the point.

 

[Phynaz]

Dude you got owned.. fess up

2012
AMD loss
Intel profit

2013
AMD loss
Intel profit

2014
AMD loss
Intel profit

2015
AMD loss
Intel profit

2016
AMD loss
Intel profit


I'd say the adf is getting "owned".

 

[Sweepr]

2012
AMD loss
Intel profit

2013
AMD loss
Intel profit

2014
AMD loss
Intel profit

2015
AMD loss
Intel profit

2016
AMD loss
Intel profit


I'd say the adf is getting "owned".

Ouch.

 

[svenge]

2012
AMD loss
Intel profit

2013
AMD loss
Intel profit

2014
AMD loss
Intel profit

2015
AMD loss
Intel profit

2016
AMD loss
Intel profit

2017
AMD loss, sells more of their physical assets
Intel profit

2018
AMD loss, sells the majority of their patents
Intel profit

2019
AMD loss, barely manages to pay the $600mm of long-term debt owed that year
Intel profit

2020
AMD loss and subsequent Chapter 11 bankruptcy after failing to refinance the remaining 1.5 billion of long-term debit due between 2020-2024
Intel profit

 

[tenks]

2012
AMD loss
Intel profit

2013
AMD loss
Intel profit

2014
AMD loss
Intel profit

2015
AMD loss
Intel profit

2016
AMD loss
Intel profit


I'd say the adf is getting "owned".

Phynaz IS exactly right and didn't get owned. THe numbers guy completely missed the point and made it more complicated then it needed to be. This was what my original comments were referring to. You can't get much more simple than that. Sure things can change, I get it, but you can't ignore the past misteps and failures..especially with a smaller company. Intel has fucked up many times, but they usually buy (R&D) their way back into things if they need to (Conroe). AMD doesn't have that luxury.

 

[RussianSensation]

2020
AMD loss and subsequent Chapter 11 bankruptcy after failing to refinance the remaining 1.5 billion of long-term debit due between 2020-2024
Intel profit

I am going to follow your advice, around December 31, 2019, I am going to short-sell AMD stock, sell all my assets and buy puts with the proceeds. I'll post a bunch of pics with my new Ferrari and a nice house with tennis courts circa 2020. I hope you do the same so we can do a Gumball around the world in our supercars, while donating 24-core Intel CPUs to the poor!

Moving on now to discuss Kaby Lake, not AMD, NV, Via, Matrox, etc.

 

[mysticjbyrd]

Is there any word on how many cores would be offered on this generation?

 

[Arachnotronic]

Is there any word on how many cores would be offered on this generation?

Up to 4C/8T, just like Skylake.

 

[Sweepr]

Is there any word on how many cores would be offered on this generation?

Still quad-core. Cannonlake is rumoured to increase the core count.

 

[mysticjbyrd]

Still quad-core. Cannonlake is rumoured to increase the core count.

Thx, what's your best estimate of the release of Kaby and Cannonlake?

Q3-4 2016
Q2-3 2017

"Then in the second half of 2017, we expect to launch our first 10-nanometer product, code named Cannonlake. "

July 2015
http://www.extremetech.com/computin...-will-introduce-kaby-lake-at-14nm-to-fill-gap

 

[dark zero]

I am going to follow your advice, around December 31, 2019, I am going to short-sell AMD stock, sell all my assets and buy puts with the proceeds. I'll post a bunch of pics with my new Ferrari and a nice house with tennis courts circa 2020. I hope you do the same so we can do a Gumball around the world in our supercars, while donating 24-core Intel CPUs to the poor!

Moving on now to discuss Kaby Lake, not AMD, NV, Via, Matrox, etc.

INBF...
AMD dies at 2018, the same date when VIA needs Intel's response in order to stay in x86 and BAM!, Intel denies it skyrocketing the prices of most of their products.

 

[IntelUser2000]

Thx, what's your best estimate of the release of Kaby and Cannonlake?

Q3-4 2016
Q2-3 2017

"Then in the second half of 2017, we expect to launch our first 10-nanometer product, code named Cannonlake. "

July 2015
http://www.extremetech.com/computin...-will-introduce-kaby-lake-at-14nm-to-fill-gap

Easy.

July 2014 - Haswell Refresh
August 2015 - Skylake
September/October 2016 - Kabylake
October/November 2017 - Cannonlake

It should be 12 months but it never turns out to be.

Also, "second half" unfortunately means "deep into" so July is quite unlikely.

 

[AtenRa]

Easy.

July 2014 - Haswell Refresh
August 2015 - Skylake
September/October 2016 - Kabylake
October/November 2017 - Cannonlake

It should be 12 months but it never turns out to be.

Also, "second half" unfortunately means "deep into" so July is quite unlikely.

Kabylake Desktop is Q4 2016, that is October to December(most probably November 2016). And since Skylake will be a 5Q product, im assuming Kabylake will too and that will put Desktop Cannonlake to Q1 2018.

 

[mysticjbyrd]

Kabylake Desktop is Q4 2016, that is October to December(most probably November 2016). And since Skylake will be a 5Q product, im assuming Kabylake will too and that will put Desktop Cannonlake to Q1 2018.

Idk, I think intel will do anything it takes to make sure they release Kabylake before AMD can launch Zen. AMD is seemingly always late to party!

Heck, with the added cores Zen could even be a threat to Kabylake's performance, particularly in multithreaded tasks. I am kind of doubtful that will include gaming though seeing as console are limited to 8 threads.

 

[AtenRa]

Idk, I think intel will do anything it takes to make sure they release Kabylake before AMD can launch Zen. AMD is seemingly always late to party!

Heck, with the added cores Zen could even be a threat to Kabylake's performance, particularly in multithreaded tasks. I am kind of doubtful that will include gaming though seeing as console are limited to 8 threads.

Kabylake seems it will be release the same time as ZEN, Q4 2016. But, they dont directly compete.

 

[mrmt]

Projected 2015 Mobile/PC group shortfall c.f 2014 PC profits:
Q1 = 1.4, Q2=2.1, assume Q3=1.7, assume Q4=1.4
Total shortfall = 6.6 billion USD.
If 4.2 billion is attributable to Mobile,as per 2014, then PC shortfall for 2015 is 2.4 billion.
If 3.4 billion is attributable to Mobile, then PC shortfall for 2015 is 3.2 billion USD.

Your numbers are flawed. On Intel investor relation site you can find the quarterly comparisons of the PC groups with the mobile numbers included, they do not come even close to what you are posting here.

 

[Eddward]

10 bit HEVC/VP9 in hardware is a big update, much more important than another 5% CPU speedup. This is what I'm looking for. So it definitely isn't just a renamed Skylake. Good to know that Kabylake is compatible with 100 series chipset.

Based on this and the fact that KabyLake has never been in the roadmap before, I really think that KabyLake is just Cannonlake uarch. on 14nm.
Intel is pushing Cannonlake back because of 10nm process..uarch itself should be ready as originally planned for 2016.
It's just impossible to make this core changes in such a short time. So yeah, most likely not the same situation as Haswell - Haswell refresh. In this case it was basically the same silicon with Mhz bump.

 

[frozentundra123456]

Cannonlake even if on time would not have been a new uarch. Skylake is the "tock", cannonlake was scheduled to be the "tick", i.e. shrink of the same uarch. So I expect cpu wise, not much change except maybe a clockspeed bump like haswell refresh. I believe there are supposed to be more significant igpu improvements (if anybody cares).

 

[crashtech]

Could be that since shrinks are becoming so difficult and taking longer, we will see more µarch changes within the same node moving forward. Surely Tick-Tock as originally envisioned can't continue forever.

 

[Arachnotronic]

Could be that since shrinks are becoming so difficult and taking longer, we will see more µarch changes within the same node moving forward. Surely Tick-Tock as originally envisioned can't continue forever.

Tick-Tock is dead. Say hello to "Tick-Tock-Tock"

 

[Sweepr]

Tick-Tock is dead. Say hello to "Tick-Tock-Tock"

Yes, the way I see it, Cannonlake is Kaby Lake's 'tick'. Except this tick will probably bring more cores at the same TDP for both mobile/desktop (and improved iGPU).

 

[mikk]

Based on this and the fact that KabyLake has never been in the roadmap before, I really think that KabyLake is just Cannonlake uarch. on 14nm.

I don't think so.

Kabylake= Gen 9.5
Cannonlake= Gen10

 

[Eddward]

I don't think so.

Kabylake= Gen 9.5
Cannonlake= Gen10

Of course, they have to change plans accordingly.
Cannonlake core in 2017 will not be the same Cannonlake as Intel planned years ago. Maybe I'm wrong but there are several reasons for this option.

@frozentundra123456: by new uarch I mean updates in core and iGPU... refresh like Haswell refresh has nothing new...Mhz bump, maybe little bugfixes and unlocked instruction set, it requires max new revision/stepping on the same silicon... in Kabylake there will be updates, it's basically another 14nm tock with tick improvements...