(some rough page) |
(→Architecture) |
||
Line 42: | Line 42: | ||
# Reduce power consumption, | # Reduce power consumption, | ||
# while staying fully [[x86]]-compatible, | # while staying fully [[x86]]-compatible, | ||
− | # at acceptable performance | + | # at acceptable performance |
+ | Performance/Power new rule: +1% performance for at most +1% power consumption. | ||
+ | === Memory Hierarchy === | ||
+ | * Cache | ||
+ | ** Hardware prefetchers | ||
+ | ** L1 Cache: | ||
+ | *** 32 KB 8-way [[set associative]] instruction | ||
+ | **** 1 read and 1 write port | ||
+ | *** 24 KB 6-way set associative data | ||
+ | **** 1 read and 1 write port | ||
+ | *** Per core | ||
+ | ** L2 Cache: | ||
+ | *** 512 KB 8-way set associative | ||
+ | *** Per core | ||
+ | ** L3 Cache: | ||
+ | *** No level 3 cache | ||
+ | ** RAM | ||
+ | *** Maximum of 2 GB, 4 GB, and 8 GB | ||
+ | |||
+ | Note that the L1 cache for data and instructions were originally both 32 KB (8-way), however due to power restrictions, the L1d$ was later reduced to 24 KB. | ||
+ | === Functional Units === | ||
+ | The number of functional units were kept to minimum to cut on power consumption. | ||
+ | * 2 Integer [[ALU]]s (1 for jumps, 1 for shifts) | ||
+ | * 2 FP ALUs (1 adder, 1 for others) | ||
+ | * No Integer multiplier & divider | ||
+ | === Pipeline === | ||
+ | Much like other x86 microarchitectures, Bonnell converts the complex instructions into finer [[micro-ops]]. Bonnell has a 16-stage pipeline with a 13-stage miss penalty. Bonnell is a dual-issue [[superscalar]] but with in-order execution (in fact, first microarchitecture since {{intel|Pentium Pro}} to not feature a OoOE). The elimination of reordering logic allowed for lower power consumption and small die area. This does imply the overall MPU is less efficient in managing its own resources; memory accesses and FP operations also stall the whole pipeline. Bonnell employed [[Safe Instruction Recognition]] (SIR) and [[Simultaneous multithreading]] (SMT) to bring performance to acceptable level. Intel claimed sub-20% power consumption penalty while improving performance between 30% and 50%. | ||
== Cores == | == Cores == |
Revision as of 15:34, 7 April 2016
Edit Values | |
Bonnell µarch | |
General Info |
Bonnell was a microarchitecture for Intel's 45 nm ultra-low power microprocessors first introduced in 2008 for their then-new Atom family. Bonnell, which was named after the highest point in Austin - Mount Bonnell, was Intel's first x86-compatible microarchitecture designed to target the ultra-low power market.
Contents
Architecture
Bonnell's primary goals were:
- Reduce power consumption,
- while staying fully x86-compatible,
- at acceptable performance
Performance/Power new rule: +1% performance for at most +1% power consumption.
Memory Hierarchy
- Cache
- Hardware prefetchers
- L1 Cache:
- 32 KB 8-way set associative instruction
- 1 read and 1 write port
- 24 KB 6-way set associative data
- 1 read and 1 write port
- Per core
- 32 KB 8-way set associative instruction
- L2 Cache:
- 512 KB 8-way set associative
- Per core
- L3 Cache:
- No level 3 cache
- RAM
- Maximum of 2 GB, 4 GB, and 8 GB
Note that the L1 cache for data and instructions were originally both 32 KB (8-way), however due to power restrictions, the L1d$ was later reduced to 24 KB.
Functional Units
The number of functional units were kept to minimum to cut on power consumption.
- 2 Integer ALUs (1 for jumps, 1 for shifts)
- 2 FP ALUs (1 adder, 1 for others)
- No Integer multiplier & divider
Pipeline
Much like other x86 microarchitectures, Bonnell converts the complex instructions into finer micro-ops. Bonnell has a 16-stage pipeline with a 13-stage miss penalty. Bonnell is a dual-issue superscalar but with in-order execution (in fact, first microarchitecture since Pentium Pro to not feature a OoOE). The elimination of reordering logic allowed for lower power consumption and small die area. This does imply the overall MPU is less efficient in managing its own resources; memory accesses and FP operations also stall the whole pipeline. Bonnell employed Safe Instruction Recognition (SIR) and Simultaneous multithreading (SMT) to bring performance to acceptable level. Intel claimed sub-20% power consumption penalty while improving performance between 30% and 50%.
Cores
First generation of Bonnel-based microprocessors introduced 2 cores: Silverthorne for ultra-mobile PCs and mobile Internet devices (MIDs) and Diamondville for ultra cheap notebooks and desktops.
Silverthorne
- Main article: Silverthorne
Silverthorne was the codename for a series of ultra-mobile PCs introduced in 2008.
Key features:
- 32-bit
- 1 Core / 2 Threads
- FSB 400 MHz - 533 MHz
- Cache
- 32 KB L1i$ / 23 KB L1d$
- 512 KB L2$
- No L3$
Diamondville
- Main article: Diamondville
Diamondville was the codename for the series of ultra cheap notebooks and desktops introduced in 2008. Diamondville is very much a derivative of Silverthorne with faster FSB.
Key features:
- 32-bit
- 1 Core / 2 Threads
- FSB 533 MHz - 667 MHz
- Cache
- 32 KB L1i$ / 23 KB L1d$ (per core)
- 512 KB L2$ (per core)
- No L3$