A deep dive into energy efficient multi core processor
1. A Deep Dive into Energy Efficient Multi-
core Processor
Speaker:呂宗螢
Adviser:梁文耀 老師
Date:2006/11/1
2. 嵌入式及平行系統實驗室
Hardware Software Cooperation for Best
Performance-Power Optimization
ACPI
Schedule appropriate idle
state, either C1, C2, ….Cn
OS schedule and decide
designated idle state
Execute BIOS code
sequence
Enter idle state and
periodically re-evaluate
Quickly back to active
state when the break
event occurs
Utilize idle state effectively
3. 嵌入式及平行系統實驗室
Power States
C0 is core active power state
C1 is auto halt state
Instruction execution halted
Core clock is off
Quick transition and switch on/off almost immediately
C2 is stop clock state
Core and bus clock off
The front side bus can placed in a lower power state
Chipset can initiate power-saving measure
C3 is deep sleep state
Clock generator off
Processor disables internal Phase Locked Loops
C4 is deeper sleep state
Reduce VCC
DC4 is further deeper sleep state
Further reduce VCC
DC4 has eliminating most static power but taking a
long time the enter DC4 and back to C0
4. 嵌入式及平行系統實驗室
Multi-core Power States
Multi-core power state
Each core has its own power state
Multi-core shared resource has power states for all cores
• Maximum performance benefit
• Minimum power consumption
Multi-core power control partion
Typical individual core units:execution units , core scheduler, L1 cache, core
clocking control
Typical shared resources:APIC, global power control, L2 cache, bus
5. 嵌入式及平行系統實驗室
Intel Core™ 2 Duo Processor
A core reduce the frequency , the other
core still runt at full speed
Three domain
Each core have power management
domain
Share reources a power management
domain
All domain share a single power plane
and a single-core PLL(Phase Locked
Loops)
The Coordination mechanism serves as a
transparent layer between the individuall
controlled cores and the shared resource
on die and on the platform
Determines the required CPU C-state
Base on both cores’ individual requests
Controls the state of the shared
resources
6. 嵌入式及平行系統實驗室
Smart Cache Optimization
Smart Cache size shrink
Less accesses than threshold
Multi-core condition permits
Stop cach size shrink
Other core active
Not meet multi-core shrinking
condition
Interrupt pending
Expand cache size
Access frequency over threshold
Other core sleep state change
Multi-core trend to active