Plan B was to soup up 14nm by enhancing it by as much as the rest of the industry is enhancing their 10nm processes compared to their 14nm/16nm processes, tweak the circuit design to allow for higher frequencies without increasing power, and add new media capabilities that have a serious impact on user experience.
In mobile, Kaby Lake is arguably a bigger improvement in CPU perf than either Broadwell or Skylake were. But you seem so set in your belief that Intel sucks that you ignore a lot of this goodness
dude you can keep defending Intel but just because process improvements happen does not mean architectural improvements cannot be done. AMD moved from Pilderiver (32nm SOI) to Steamroller (28nm SHP bulk) but it also had IPC improvements. This is not unheard of in the fabless ecosystem. Apple keeps improving IPC every year no matter even if there are process improvements. If AMD and Apple can why the heck Intel cannot. This tick tock process optimization rubbish is going to hurt them.
Bristol Ridge???
~10% higher clocks with no GPU inprovement. What goodness?In mobile, Kaby Lake is arguably a bigger improvement in CPU perf than either Broadwell or Skylake were. But you seem so set in your belief that Intel sucks that you ignore a lot of this goodness
Richland?Bristol Ridge???
So is Kaby Lake, what is your point?dude Bristol Ridge is a product which filled a gap in the product roadmap till Raven Ridge arrived.
Conroe was based on the good know old Pentium M. Netburst was a flop, dead end. Now, there is nothing like this. Intel was very self confident up to this point. They most likely have something really new in development but certainly won't be ready for 2018.remember 10-11 years ago when Intel had a long running arch, NetBurst, and faced fierce competition from AMD on the dual core front? AMD had a more efficient arch and covered most price segments, with a strong back-wind from the success of it's Athlon CPUs vs Inte's P4.
remember what happened then? Conroe happened, and started a decade of market domination.
if AMD pulls off something big with Ryzen, i foresee a Conroe 2.0 towards 2018. the improvements don't have to be architectural, but they SHOULD be.
Intel can't afford to lose now, they've gotten too big.
The answer is obvious: Intel IPC already is quite high and their frequency already is quite high. AMD was starting from a sh*tty CPU so gaining a lot of IPC was trivial. Apply isn't targeting >4 GHz (yet) so gaining IPC is easier.dThis is not unheard of in the fabless ecosystem. Apple keeps improving IPC every year no matter even if there are process improvements. If AMD and Apple can why the heck Intel cannot.
I don't think you will get much sleep.The answer is obvious: Intel IPC already is quite high and their frequency already is quite high. AMD was starting from a sh*tty CPU so gaining a lot of IPC was trivial. Apply isn't targeting >4 GHz (yet) so gaining IPC is easier.
Wake me when either AMD or Apple has the edge on single thread performance over Intel.
The answer is obvious: Intel IPC already is quite high and their frequency already is quite high. AMD was starting from a sh*tty CPU so gaining a lot of IPC was trivial. Apply isn't targeting >4 GHz (yet) so gaining IPC is easier.
Wake me when either AMD or Apple has the edge on single thread performance over Intel.
You forgot about 3 tocks - Piledriver(Trinity) , Steamroller (Kaveri), Excavator (Carrizo) and you picked on the lone Bristol Ridge example.
It has been known from the start that kaby lake is the exact same core as skylake. Improvements come from a refined process (14nm+).Followup article by CPCHardware:
https://www.cpchardware.com/core-i7-7740k-et-i5-7640k-kaby-lake-x-lga2066/
They got the i5 7640K sample and it's a LGA2066 but they don't have a compatible MB yet. The CPU advertises itself as LGA1151 via CPUID (tested by sample source). Also, the i5 seems to have no Turbo boost enabled.
They then go on and that Kaby Lake is in fact "just" a Skylake stepping and not a new architecture, same thing for Coffee Lake.
Interesting read it seems to show some kind of disarray at Intel.
Followup article by CPCHardware:
https://www.cpchardware.com/core-i7-7740k-et-i5-7640k-kaby-lake-x-lga2066/
They got the i5 7640K sample and it's a LGA2066 but they don't have a compatible MB yet.
I really hope I'll get little sleep, I'm a big fan of sane competition and I'm tired with Intel behavior.I don't think you will get much sleep.
They got the i5 7640K sample and it's a LGA2066 but they don't have a compatible MB yet. The CPU advertises itself as LGA1151 via CPUID (tested by sample source).
I'll stick with my prediciton: Intel is scrambling to get SKL-X out ASAP.So to get a 100 MHz boost in clocks (along with a 20W "boost" in power consumption), an entirely new motherboard is required.
Any discernible reason for this?
Isn't KBL-X dual channel?The reason is the migration to the new quad-channel platform ... and that's about it!
There is also a difference in AVFS implementation: http://ieeexplore.ieee.org/document/7417913/ It's paywalled, but in short, Carrizo AVFS gives consensus to a turbo P-state, given by active core numbers, if the current status (temperature etc) is good enough. On Bristol Ridge, the AVFS chooses the maximum P-state attainable, regardless the number of cores active. It is a precursor of current XFR: it's similar, but limited to the max boost supported by the CPU. XFR, AFAIK, is not limited, moreover has 25MHz granularity and it's not bound to p-states...Carrizo = Bristol Ridge. They're the same thing, only difference is Carrizo uses DDR3 while BR uses DDR4.
Isn't KBL-X dual channel?
But do you really think AMD or Apple will come with a chip with better single thread performance than Intel within 2 or 3 years? I don't.
There is also a difference in AVFS implementation: http://ieeexplore.ieee.org/document/7417913/ It's paywalled, but in short, Carrizo AVFS gives consensus to a turbo P-state, given by active core numbers, if the current status (temperature etc) is good enough. On Bristol Ridge, the AVFS chooses the maximum P-state attainable, regardless the number of cores active. It is a precursor of current XFR: it's similar, but limited to the max boost supported by the CPU. XFR, AFAIK, is not limited, moreover has 25MHz granularity and it's not bound to p-states...
EDIT: if you know italian, here it is my full post: http://www.hwupgrade.it/forum/showpost.php?p=44459489&postcount=15611