Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Skylake and Broadwell both for Q2 this year
- Thread starter Liffey
- Start date
- Sort by reaction score
Even on this graph it isn't Q2.. The abnormally early ones are more so at the beginning of Q3 than in Q2.
Where is the article is this attached to?
Last Spring Intel's CEO claimed Broadwell by Xmas ....
http://vr-zone.com/articles/can-guarantee-broadwell-holiday/77580.html
Kind of hard to believe that Intel is going to nuke Broadwell U 3-5 months after launching it. If it had launched on its original schedule that might have made sense.
Skylake graphics may be up 50% in some cases but the x86 subset isn't likley to so that in generic x86 code.
If you mean Skylake will be generally 50% faster than Broadwell, what is the basis for that? If you mean integrated GPU, iMacs use discrete GPUs (except for bottom end of 21.5" lineup).
I have never seen any credible rumor that Skylake will be have 50% higher integer or floating point performance than Broadlake at the same clock rate or core count, or even at any shipping clock rate.
Intel says that Skylake will be more of a CPU upgrade than previous generations but that may not be saying much when you look at the recent CPU speed increases which have been about 5% to 10%.
The mobile Macs or select mobile Macs will get Skylake first which could be late fall, to excite holiday sales.
The desktop variants will probably come about the middle of 2016. Skylake for the Xeons will probably show up late 2016 or early 2017.
This can all change depending on Intel's strategy and success with Skylake chips.
The mobile Macs or select mobile Macs will get Skylake first which could be late fall, to excite holiday sales.
The desktop variants will probably come about the middle of 2016. Skylake for the Xeons will probably show up late 2016 or early 2017.
This can all change depending on Intel's strategy and success with Skylake chips.
Intel is most likely going to roll these out Core M (lowest power) , U (ultra book) , H (laptop) , S ( desktop ) order. The longer Apple doesn't drop a Core M model the less likely they are going to drop another one with late Fall Skylake version.
If Intel drops Broadwell Desktops in Q2 '15 then it probably will not take till middle of 2016 for Skylake to roll out. This graph's depiction of roughly 2 months later doesn't seem creditable (less than a quarter is a ridiculously short product lifespan as lead offering). [ If it was Broadwell BGA (soldered to motherboard) desktop CPUs and then Skylake LGA (socketed) CPUs then I could see the short period as they would be somewhat non overlapping products. ] But over 12 months isn't ( past mid Q2 '16) is a bit too much time.
Technically, Skylake Xeons will likely show up around the same time as the desktop Skylake as a subset are based upon the same fundamental implementation die design. They will be the Xeon E3 but Xeons none-the-less.
Xeon E5 v5 likely isn't coming late 2016. The Broadwell-E (which are based on same basic implementation design used for Xeon E5 ) are targeted on the chart as Q1 2016. It is unlikely Intel is going to drop two Xeon E5 class updates in a single calendar year. Intel stretching out the Xeon E5 updates is a pattern at this point starting with the original Xeon E5 models.
[ huge transistor budgets and optimization on that large budgets just takes more time. Plus AMD isn't exactly breathing down their necks either. The more mature the process tech is when they roll these out the bigger profits. It isn't a drag race. ]
Not just Skylake, but also the Braswell ( and the other Atoms also). If their volume cranks up the fab capacity will get soaked up also. Sliding out the bigger dies ( Xeon E5 (-E in Core i7 space) , big quads in Core i7 ) would not be surprising. AMD isn't doing anything major until late 2016 and even then it is questionable it will be fiercely competitive with Intel's offerings at that point in the high performance zone.
The Atom ( Braswell , Bay-Trail ) zone is where Intel is burning money right now. They deeply need stuff that doesn't require huge subsidies to get system vendors to buy. If the new process tech and more optimized designs for that new process tech tip the balance there, then they will get pretty high priority routing through the fabs. It is a bigger "bottom line" bang-for-the-buck to get rid of the subsidies.