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

(Architecture)
Line 86: Line 86:
  
 
== Architecture ==
 
== Architecture ==
{{future information}}
+
Tremont is designed with significant single-thread performance in mind while focusing on low-power small silicon area cores.
 
=== Key changes from {{\\|Goldmont Plus}} ===
 
=== Key changes from {{\\|Goldmont Plus}} ===
* Graphics
+
* Significant [[IPC]] uplift ([[Intel]] self-reported average 32% IPC accross proxy benchmarks such as [[SPEC CPU2006]]/[[SPEC CPU2017]])
** {{intel|Gen 11|l=arch}}
+
* Front-end
 +
** Redesigned front-end
 +
*** New dual symmetric decode cluster
 +
**** Out-of-order decode
 +
**** 6-wide decode
 +
***** 3-way decode per cluster
 +
** Smarter [[prefetchers]]
 +
** Improved [[branch predictor]]
 +
*** Big-core level of performance
 +
* Back-end
 +
** larger ROB
 +
** wide issue (10-wide)
 +
* Execution Engine
 +
** 2x store data ports (up from 1)
 +
 
  
 
====New instructions ====
 
====New instructions ====
Line 104: Line 118:
 
=== Block Diagram ===
 
=== Block Diagram ===
 
==== Individual Core ====
 
==== Individual Core ====
 +
:[[File:tremont block diagram.svg|850px]]

Revision as of 12:26, 24 October 2019

Edit Values
Tremont µarch
General Info
Arch TypeCPU
DesignerIntel
ManufacturerIntel
Introduction2019
Process10 nm
Pipeline
TypeSuperscalar
OoOEYes
SpeculativeYes
Reg RenamingYes
Instructions
ISAx86-64
ExtensionsMOVBE, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND, XSAVE, XSAVEOPT, FSGSBASE, PTWRITE, RDPID, SGX, UMIP, GFNI-SSE, CLWB, ENCLV, SHA
Cores
Core NamesElkhart Lake,
Skyhawk Lake
Succession

Tremont is Intel's successor to Goldmont Plus, a 10 nm microarchitecture for ultra-low power devices and microservers.

Codenames

Platform Core Name PCH
Skyhawk Lake
Jacobsville Elkhart Lake Mule Creek Canyon

Brands

New text document.svg This section is empty; you can help add the missing info by editing this page.

Release Dates

Tremont was released in a number of products in late 2019.

Technology

Tremont uses Intel's 10 nm process.

Compiler support

Compiler Arch-Specific Arch-Favorable
ICC -march=tremont -mtune=tremont
GCC -march=tremont -mtune=tremont
LLVM -march=tremont -mtune=tremont
Visual Studio /arch:? /tune:?

CPUID

Core Extended
Family
Family Extended
Model
Model
 ? 0 0x6 0x8 0x6
Family 6 Model 134

Architecture

Tremont is designed with significant single-thread performance in mind while focusing on low-power small silicon area cores.

Key changes from Goldmont Plus

  • Significant IPC uplift (Intel self-reported average 32% IPC accross proxy benchmarks such as SPEC CPU2006/SPEC CPU2017)
  • Front-end
    • Redesigned front-end
      • New dual symmetric decode cluster
        • Out-of-order decode
        • 6-wide decode
          • 3-way decode per cluster
    • Smarter prefetchers
    • Improved branch predictor
      • Big-core level of performance
  • Back-end
    • larger ROB
    • wide issue (10-wide)
  • Execution Engine
    • 2x store data ports (up from 1)


New instructions

Tremont introduced a number of new instructions:

  • CLWB - Force cache line write-back without flush
  • ENCLV - SGX oversubscription instructions
  • CLDEMOTE - Cache line demote instruction
  • SSE_GFNI - SSE-based Galois Field New Instructions
  • Direct store instructions: MOVDIRI, MOVDIR64B
  • User wait instructions: TPAUSE, UMONITOR, UMWAIT
  • Split Lock Detection - detection and cause an exception for split locks

Block Diagram

Individual Core

tremont block diagram.svg
codenameTremont +
designerIntel +
first launched2019 +
full page nameintel/microarchitectures/tremont +
instance ofmicroarchitecture +
instruction set architecturex86-64 +
manufacturerIntel +
microarchitecture typeCPU +
nameTremont +
process10 nm (0.01 μm, 1.0e-5 mm) +