From WikiChip
Difference between revisions of "intel/process-architecture-optimization"
< intel

(Roadmap)
(Roadmap)
 
(12 intermediate revisions by 7 users not shown)
Line 1: Line 1:
 
{{intel title|Process-Architecture-Optimization (PAO)}}
 
{{intel title|Process-Architecture-Optimization (PAO)}}
'''[[name::Process-Architecture-Optimization]]''' is a [[instance of::development model]] introduced by [[Intel]] for their mainstream microprocessors in [[2016]] following the phase-out of their {{intel|Tick-Tock}} model. The change is a result of the increase in cost and complexity of advancing lithography processes in the past decade. Under the new model the amount of time utilized for any given process technology is lengthened as [[Moore's Law]] increases in complexity with smaller [[technology node|nodes]].
+
'''[[name::Process-Architecture-Optimization]]''' was a temporary [[instance of::development model]] introduced by [[Intel]] for their mainstream microprocessors in [[2016]] following the phase-out of their {{intel|Tick-Tock}} model as a result of major delays and challenges involving their [[10 nm process]].
  
 
Under the Process-Architecture-Optimization Model:
 
Under the Process-Architecture-Optimization Model:
  
* '''Process''' -  With each process, Intel advances their manufacturing [[process technology]] in line with [[Moore's Law]]. Each new process introduces higher transistor density and a generally a plethora of other advantages such as higher performance and lower power consumption. During a "process", Intel retrofits their {{intel|microarchitectures|previous}} [[microarchitecture]] to the new process which inherently yields better performance and energy saving. During a "process", usually just a few features and improvements are introduced.
+
* '''Process''' -  With each process, Intel advances their manufacturing [[process technology]] in line with [[Moore's Law]]. Each new process introduces higher transistor density and generally a plethora of other advantages such as higher performance and lower power consumption. During a "process", Intel retrofits their {{intel|microarchitectures|previous}} [[microarchitecture]] to the new process which inherently yields better performance and energy saving. During a "process", usually, just a few features and improvements are introduced.
  
 
* '''Architecture''' - With each architecture, Intel uses the their latest manufacturing [[process technology]] from their "process" to manufacture a newly designed [[microarchitecture]]. The new microarchitecture is designed with the new process in mind and typically introduces Intel's newest big features and functionalities. New [[instruction set|instructions]] are often added during this cycle stage.
 
* '''Architecture''' - With each architecture, Intel uses the their latest manufacturing [[process technology]] from their "process" to manufacture a newly designed [[microarchitecture]]. The new microarchitecture is designed with the new process in mind and typically introduces Intel's newest big features and functionalities. New [[instruction set|instructions]] are often added during this cycle stage.
  
* '''Optimization''' - With each optimization, Intel improves upon their {{intel|microarchitectures|previous}} microarchitecture by introducing incremental improvements and enhancements without introducing any large charges. Additionally the process itself enjoys various refinements as it matures. (For example with {{intel|Kaby Lake|l=arch}}, an optimized process called "14 nm+" is used. The enhanced process had a number of transistor-level modifications done to it (e.g. taller fins) allowing for higher frequency at identical voltage levels.)
+
* '''Optimization''' - With each optimization, Intel improves upon their {{intel|microarchitectures|previous}} microarchitecture by introducing incremental improvements and enhancements without introducing any large charges. Additionally, the process itself enjoys various refinements as it matures. (For example with {{intel|Kaby Lake|l=arch}}, an optimized process called "14 nm+" is used. The enhanced process had a number of transistor-level modifications done to it (e.g. taller fins) allowing for higher frequency at identical voltage levels.)
  
 
== Roadmap ==
 
== Roadmap ==
Line 17: Line 17:
 
!Cycle !! [[technology node|Process]] !! Introduction !! Micro­archi­tecture
 
!Cycle !! [[technology node|Process]] !! Introduction !! Micro­archi­tecture
 
|-
 
|-
| Process || [[14 nm]] || 2014 || {{intel|microarchitectures/Broadwell|Broadwell}}
+
| Process || [[14 nm]] || 2014 || {{intel|Broadwell|l=arch}}
 
|-
 
|-
| Architecture || [[14 nm]] || 2015 || {{intel|microarchitectures/Skylake|Skylake}}
+
| Architecture || [[14 nm]] || 2015 || {{intel|Skylake (Client)|l=arch}}
 
|-
 
|-
| Optimization || [[14 nm|14 nm+]] || 2016 || {{intel|microarchitectures/Kaby Lake|Kaby Lake}}
+
| Optimization || [[14 nm|14 nm+]] || 2016 || {{intel|Kaby Lake|l=arch}}
 
|-
 
|-
| Optimization || [[14 nm|14 nm++]] || 2017 || {{intel|microarchitectures/Coffee Lake|Coffee Lake}}
+
| Optimization || [[14 nm|14 nm++]] || 2017 || {{intel|Coffee Lake|l=arch}}, {{intel|Skylake (Server)|l=arch}}
 
|-
 
|-
| Process || [[10 nm]] || 2017 || {{intel|microarchitectures/Cannonlake|Cannonlake}}
+
| Optimization || [[14 nm|14 nm++]] || 2018 || {{intel|Amber Lake|l=arch}} {{intel|Whiskey Lake|l=arch}}
 
|-
 
|-
| Architecture || [[10 nm]] || 2018 || {{intel|microarchitectures/Icelake|Icelake}}
+
| Optimization || [[14 nm|14 nm++]] || 2019 || {{intel|Cascade Lake|l=arch}}
 
|-
 
|-
| Optimization || [[10 nm|10 nm+]] || 2019 || {{intel|microarchitectures/Tigerlake|Tigerlake}}
+
| Optimization || [[14 nm|14 nm++]] || 2020 || {{intel|Cooper Lake|l=arch}}, {{intel|Comet Lake|l=arch}}
 
|-
 
|-
| Optimization || [[10 nm|10 nm++]] || 2020 ||  
+
| Optimization || [[14 nm|14 nm++]] || 2021 || {{intel|Rocket Lake|l=arch}}
|-
 
| Process || [[7 nm]] || 20?? ||
 
 
|}
 
|}
  
 
[[Category:intel]]
 
[[Category:intel]]

Latest revision as of 21:33, 8 November 2019

Process-Architecture-Optimization was a temporary development model introduced by Intel for their mainstream microprocessors in 2016 following the phase-out of their Tick-Tock model as a result of major delays and challenges involving their 10 nm process.

Under the Process-Architecture-Optimization Model:

  • Process - With each process, Intel advances their manufacturing process technology in line with Moore's Law. Each new process introduces higher transistor density and generally a plethora of other advantages such as higher performance and lower power consumption. During a "process", Intel retrofits their previous microarchitecture to the new process which inherently yields better performance and energy saving. During a "process", usually, just a few features and improvements are introduced.
  • Architecture - With each architecture, Intel uses the their latest manufacturing process technology from their "process" to manufacture a newly designed microarchitecture. The new microarchitecture is designed with the new process in mind and typically introduces Intel's newest big features and functionalities. New instructions are often added during this cycle stage.
  • Optimization - With each optimization, Intel improves upon their previous microarchitecture by introducing incremental improvements and enhancements without introducing any large charges. Additionally, the process itself enjoys various refinements as it matures. (For example with Kaby Lake, an optimized process called "14 nm+" is used. The enhanced process had a number of transistor-level modifications done to it (e.g. taller fins) allowing for higher frequency at identical voltage levels.)

Roadmap[edit]

pao.png
Intel PAO Roadmap
Cycle Process Introduction Micro­archi­tecture
Process 14 nm 2014 Broadwell
Architecture 14 nm 2015 Skylake (Client)
Optimization 14 nm+ 2016 Kaby Lake
Optimization 14 nm++ 2017 Coffee Lake, Skylake (Server)
Optimization 14 nm++ 2018 Amber Lake Whiskey Lake
Optimization 14 nm++ 2019 Cascade Lake
Optimization 14 nm++ 2020 Cooper Lake, Comet Lake
Optimization 14 nm++ 2021 Rocket Lake
instance ofdevelopment model +
nameProcess-Architecture-Optimization +