From WikiChip
Difference between revisions of "intel/microarchitectures/golden cove"
< intel‎ | microarchitectures

(Redwood Cove in Meteor Lake is the successor)
 
(12 intermediate revisions by 9 users not shown)
Line 5: Line 5:
 
|designer=Intel
 
|designer=Intel
 
|manufacturer=Intel
 
|manufacturer=Intel
|introduction=2H 2021
+
|introduction=November 4, 2021
 
|process=10nm
 
|process=10nm
 
|isa=x86-64
 
|isa=x86-64
|core name=Alder Lake
+
|core name=Alder Lake, Sapphire Rapids
 
|predecessor=Willow Cove
 
|predecessor=Willow Cove
 
|predecessor link=intel/microarchitectures/willow cove
 
|predecessor link=intel/microarchitectures/willow cove
|successor=Ocean Cove
+
|predecessor 2=Cypress Cove
|successor link=intel/microarchitectures/ocean cove
+
|predecessor 2 link=intel/microarchitectures/rocket lake
 +
|successor=Redwood Cove
 +
|successor link=intel/microarchitectures/redwood cove
 
}}
 
}}
'''Golden Cove''' is the successor to {{\\|Willow Cove}}, a high-performance [[10 nm]] [[x86]] core microarchitecture designed by [[Intel]] for an array of server and client products, including Granite Rapids(server) and Alder Lake(client).  
+
'''Golden Cove''' is the successor to {{\\|Willow Cove}}, a high-performance [[10 nm]] [[x86]] core microarchitecture designed by [[Intel]] for an array of server and client products, including Sapphire Rapids (server) and Alder Lake (client).  
  
 
== History ==
 
== History ==
Line 21: Line 23:
  
 
== Process Technology ==
 
== Process Technology ==
Intel has confirmed that the Golden Cove architecture will be fabricated on their [[10 nm process]].
+
Intel has confirmed that the Golden Cove architecture will be fabricated on their Intel 7 process (previously 10 nm Enhanced SuperFin (ESF)).
  
 
== Architecture ==
 
== Architecture ==
 
=== Key changes from {{\\|Willow Cove}}===
 
=== Key changes from {{\\|Willow Cove}}===
 
* Performance improvements
 
* Performance improvements
** Strong IPC improvement
+
** Strong IPC improvement (19%)
** AI workload improvement
+
** AI workload improvement (AMX)
 
** Network/5G performance improvements
 
** Network/5G performance improvements
 
* New security features
 
* New security features
 +
* Front-End
 +
** Add 2 simple decoders from 5 but canceled the complex decoder design as it is no longer practical and complex instructions can now be handled by macro uop fusion independently in instruction sequencer, total is now 6 simple decoders
 +
** x2.5 BTB at 12K entries
 +
** 2x pages 4k
 +
** Add more 256 and 32 pages of 2M and 4MB respectively
 +
* Back-End
 +
** Increased ROB 512 (from 352 Sunny Cove)
 +
** Add 2 execution port for a total 12
 +
* Execution Engine
 +
** Add one ALU and LEA for a total 5
 
{{expand list}}
 
{{expand list}}
  
 
== Bibliography ==
 
== Bibliography ==
 
* Intel Architecture Day 2018, December 11, 2018
 
* Intel Architecture Day 2018, December 11, 2018
 +
* Intel Details Golden Cove: Next-Generation Big Core For Client and Server SoCs

Latest revision as of 19:57, 12 June 2024

Edit Values
Golden Cove µarch
General Info
Arch TypeCPU
DesignerIntel
ManufacturerIntel
IntroductionNovember 4, 2021
Process10nm
Instructions
ISAx86-64
Cores
Core NamesAlder Lake, Sapphire Rapids
Succession

Golden Cove is the successor to Willow Cove, a high-performance 10 nm x86 core microarchitecture designed by Intel for an array of server and client products, including Sapphire Rapids (server) and Alder Lake (client).

History[edit]

Intel Core roadmap

Golden Cove was originally unveiled by Intel at their 2018 architecture day. Golden Cove is intended to succeed Willow Cove in the 2021 timeframe.

Process Technology[edit]

Intel has confirmed that the Golden Cove architecture will be fabricated on their Intel 7 process (previously 10 nm Enhanced SuperFin (ESF)).

Architecture[edit]

Key changes from Willow Cove[edit]

  • Performance improvements
    • Strong IPC improvement (19%)
    • AI workload improvement (AMX)
    • Network/5G performance improvements
  • New security features
  • Front-End
    • Add 2 simple decoders from 5 but canceled the complex decoder design as it is no longer practical and complex instructions can now be handled by macro uop fusion independently in instruction sequencer, total is now 6 simple decoders
    • x2.5 BTB at 12K entries
    • 2x pages 4k
    • Add more 256 and 32 pages of 2M and 4MB respectively
  • Back-End
    • Increased ROB 512 (from 352 Sunny Cove)
    • Add 2 execution port for a total 12
  • Execution Engine
    • Add one ALU and LEA for a total 5

This list is incomplete; you can help by expanding it.

Bibliography[edit]

  • Intel Architecture Day 2018, December 11, 2018
  • Intel Details Golden Cove: Next-Generation Big Core For Client and Server SoCs
codenameGolden Cove +
designerIntel +
first launched0002 JL +
full page nameintel/microarchitectures/golden cove +
instance ofmicroarchitecture +
instruction set architecturex86-64 +
manufacturerIntel +
microarchitecture typeCPU +
nameGolden Cove +
process10 nm (0.01 μm, 1.0e-5 mm) +