Dezső Sima Evolution of Intel’s Basic Microarchitectures - 2 October 2012 Vers. 3.2
Jan 05, 2016
Dezső Sima
Evolution of Intel’s Basic Microarchitectures - 2
October 2012
Vers. 3.2
Contents
1. Introduction•
2. Core 2•
3. Penryn•
4. Nehalem•
7. Westmere-EX •
5. Nehalem-EX •
6. Westmere•
Contents
9. Sandy Bridge Extreme Edition•
10. Ivy Bridge•
11. Overview of the evolution•
8. Sandy Bridge•
11. Haswell•
8. Sandy Bridge
8.1 Introduction•
8.2 Advanced Vector Extension (AVX)•
8.3 On-die ring interconnect bus•
8.4 On-die integrated graphics unit•
8.5 Enhanced turbo boost technology•
8.1 Introduction (1)
• Sandy Bridge is Intel’s new microarchitecture using 32 nm line width.
• First delivered in 1/2011
8.1 Introduction
32K L1D (3 clk)AVX 256 bit4 Operands
256 KB L2(9 clk)
HyperthreadingAES Instr.
VMX Unrestrict.20 nm2 / Core
256 KB L2(9 clk)
256 KB L2(9 clk)
256 KB L2(9 clk)
256 KB L2(9 clk)
256 KB L2(9 clk)
256 KB L2(9 clk)
PCIe 2.0
@ 1.0 1.4 GHz(to L3 connected)
256 b/cycle Ring Architecture
(25 clk)
DDR3-1600 25.6 GB/s
Main functional units of Sandy Bridge [143] Part 4
32 nm process / ~225 nm2 die size / 85W TDP
8.1 Introduction (2)
Key features and benefits of the Sandy Bridge line vs the 1. generation Nehalem line [61]
8.1 Introduction (3)
Desktops
Servers
DP-Servers
E5 2xxx, Sandy Bridge-EP, up to 8C, Q4/2011
UP-Servers
E3 12xx, 4C, Sandy Bridge-H2, 4C, 3/2011
Mobiles
Core i3-23xxM, 2C, 2/2011 Core i5-24xxM//25xxM, 2C, 2/2011Core i7-26xxQM/27xxQM/28xxQM, 4C, 1/2011 Core i7 Extreme-29xxXM , 4C, Q1 2011
Core i3-21xx, 2C,no HT, no vPro, 2/2011Core i5-23xx 4C+G, no HT no VPro, 1/2011Core i5/24xx/25xx, 4C+G, no HT, vPro, 1/2011Core i7-26xx, 4C+G, HT, vPro, 1/2011Core i7-2700K, 4C+G, HT, no vPro, 10/2011
MP-Servers
E5 4xxx, Sandy Bridge-EX, up to 8C, Q1/2012
Overview of the Sandy Bridge based processor lines
Based on [62] and [63]
8.1 Introduction (4)
Core i7-3960X, 6C, HT, vPro??, 11/2011Core i7-3930K, 6C, HT, vPro??, 11/2011
Desktops
Sandy Bridge Sandy Bridge-ESection 9)
8.2 Advanced Vector Extension (AVX) (1)
Figure: Evolution of the SIMDprocessing width [18] BMA-ból
8.2 Advanced Vector Extension (AVX)
Sandy Bridge
Introduction of AVX
NorhwoodNorthwood (Pentium4)Northwood (Pentium4)
8 MM registers (64-bit), aliased on the FP Stack registers
8 XMM registers (128-bit)
16 XMM registers (128-bit)
16 YMM registers (256-bit)
Figure: Intel’s x86 ISA extensions - the SIMD register space (based on [18]) BMA
Ivy Bridge
8.2 Advanced Vector Extension (AVX) (2)
8.3 On-die ring interconnect bus (1)
8.4 The on die ring interconnect bus of Sandy Bridge [66]
Six bus agents.
The four cores and theL3 slices share interfaces.
8.4 On-die integrated graphics unit (1)
8.5 Sandy Bridge’s integrated graphics unit [102] Part4
12 EUs
Specification data of the HD 2000 and HD 3000 graphics [125] Part 4
-
8.4 On-die integrated graphics unit (2)
frames per sec
i5/i7 2xxx/3xxx:Sandy Bridge
i5 6xxArrandale
HD5570400 ALUs
Performance comparison: gaming [126] part 4
8.4 On-die integrated graphics unit (3)
8.5 Enhanced turbo boost technology (1)
Cooler
Innovative concept of the 2.0 generation Turbo Boost technology
Thermal capacitance
The concept utilizes the real temperature response of processors to power changes in order to increase the extent of overclocking [64]
8.5 Enhanced turbo boost technology [64]
Concept: Use thermal energy budget accumulated during idle periods to push the corebeyond the TDP for short periods of time (e.g. for 20 sec).
Multiple algorithms manage in parallel current, power and die temperature. [64]
8.5 Enhanced turbo boost technology (2)
Intelligent power sharing between the cores and the integrated graphics [64]
8.5 Enhanced turbo boost technology (3)
Intelligent power sharing between the cores and the integrated graphics [68]
8.5 Enhanced turbo boost technology (4)
[61]
WSM/M
WSM/D
NHM/M
NHM/D
8.5 Enhanced turbo boost technology (5)
Remark
8.5 Enhanced turbo boost technology (6)
• Individual cores may run at different frequencies but all cores share the same power plane.
• Individual cores may be shut down if idle by power gates.
9. The Sandy Bridge-E line
9. The Sandy Bridge-E line (1)
9 The Sandy Bridge-E line of processors (2. gen. Core i7 processors)
Introduced in 11/2011 as a “precursor” of the upcoming DP/MP server lines.
Key features vs the original Sandy Bridge line (1)
a) 6 cores with 2 cores disabled but does not incorporate graphics [76].
435 mm2 2.27 B trs15 MB L3
216 mm2 995 mtrs8 MB L2
[61][76]
9. The Sandy Bridge-E line (2)
Sandy Bridge (2x)Sandy Bridge E
CPU Specification Comparison
CPUManufacturin
gProcess
Cores
Transistor Count
Die Size
AMD Bulldozer 8C 32nm 8 ~2B 315mm2
AMD Thuban 6C 45nm 6 904M 346mm2
AMD Deneb 4C 45nm 4 758M 258mm2
Intel Gulftown 6C 32nm 6 1.17B 240mm2
Intel Sandy Bridge E (6C) 32nm 6 2.27B 435mm2
Intel Nehalem/Bloomfield 4C 45nm 4 731M 263mm2
Intel Sandy Bridge 4C 32nm 4 995M 216mm2
Intel Lynnfield 4C 45nm 4 774M 296mm2
Intel Clarkdale 2C 32nm 2 384M 81mm2
Intel Sandy Bridge 2C (GT1) 32nm 2 504M 131mm2
Intel Sandy Bridge 2C (GT2) 32nm 2 624M 149mm2
Comparison of die parameters of recent DT processors [77]
9. The Sandy Bridge-E line (3)
L1 L2 L3Main
Memory
AMD FX-8150 (3.6GHz)
4 21 65 195
AMD Phenom II X4 975 BE (3.6GHz)
3 15 59 182
AMD Phenom II X6 1100T (3.3GHz)
3 14 55 157
Intel Core i5 2500K (3.3GHz)
4 11 25 148
Intel Core i7 3960X (3.3GHz)
4 11 30 167
Cache/memory latencies of recent DT processors [77]
9. The Sandy Bridge-E line (4)
b) 4 parallel memory channels (inherited from the server side) instead of 2 of the previous lines. Support of DDR3 of up to 1600 MT/s. A single DDR3-1600 DIMM per channel or 2 DDR3-1333 DIMMs per channel [78].
9. The Sandy Bridge-E line (5)
c) 40 PCIe 2. gen. lanes to connect graphics cards directly to the processor instead of 16 to 32 of the previous generation Sandy Bridge [78].d) LGA-2011 socket instead of the LGA-1155 used in the pervious generation Sandy Bridge due to the increased number of memory channels connected to the processor..
9. The Sandy Bridge-E line (6)
ProcessorCore Clock
Cores / Threads
L3 CacheMax
Turbo
Max Overclock Multiplier
TDP Price
Intel Core i7 3960X 3.3GHz 6 / 12 15MB 3.9GHz 57x 130W $990
Intel Core i7 3930K 3.2GHz 6 / 12 12MB 3.8GHz 57x 130W $555
Intel Core i7 3820 3.6GHz 4 / 8 10MB 3.9GHz 43x 130W TBD
Intel Core i7 2700K 3.5GHz 4 / 8 8MB 3.9GHz 57x 95W $332
Intel Core i7 2600K 3.4GHz 4 / 8 8MB 3.8GHz 57x 95W $317
Intel Core i7 2600 3.4GHz 4 / 8 8MB 3.8GHz 42x 95W $294
Intel Core i5 2500K 3.3GHz 4 / 4 6MB 3.7GHz 57x 95W $216
Intel Core i5 2500 3.3GHz 4 / 4 6MB 3.7GHz 41x 95W $205
Main features of the Sandy Bridge-E line vs the Sandy Bridge line [77]
9. The Sandy Bridge-E line (7)
Intel’s DT board for Sandy Bridge-E processors [79](The DX79SI ATX board for the Core i7-3960X /3930K)
Max. memory: 8 x 8 GB
9. The Sandy Bridge-E line (8)
10. The Ivy Bridge line
10. Te Ivy Bridge line – 10.1 Introduction (1)
Introduced: 4/2013
Figure 10.1: Intel’s Tick-Tock development model [Based on 1]
Tick-Tock Development Model
Merom1
NEWMicroarchitecture
65nm
Penryn
NEWProcess
45nm
Nehalem
NEWMicroarchitecture
45nm
Westmere
NEWProcess
32nm
SandyBridge
NEWMicroarchitecture
32nm
IvyBridge
NEWProcess
22nm
Haswell
NEWMicroarchitecture
22nm
TOCK TOCKTICKTOCKTICKTOCKTICK
10. The Ivy Bridge line
11.1 Introduction
The Ivy Bridge is termed also as the 3. gen. Intel Core processors.
10.1 Introduction (2)
32 nm
216 mm2
995 mtrs
22 nm
160 mm2
1480 mtrs(Resized to
32 nm feature size)
Figure 10.2: Contrasting the Sandy Bridge and Ivy Bridge dies [81]
Sandy Bridge
Ivy Bridge
10.1 Introduction (3)
[84]
10.1 Introduction (4)
Major innovations of Ivy Bridge [80]
11.2 The new 22 nm tri-gate process technology (1)
11.2 The new 22 nm tri-gate process technology [82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
Figure: Contrasting 32 nm planar and 22 nm tri-gate transistors [82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
[82]
10.2 The new 22 nm tri-gate process technology (1)
Figure: Ivy Bridge chips on a 300 mm wafer
10.2 The new 22 nm tri-gate process technology (1)
Processor Feature size No. of coresL2 + L3
sizeNo. of transistor Die size
Ivy Bridge 22 nm Tri-Gate 4 (+ IGP) 9 MB 1,48 milliárd 160 mm2
Sandy Bridge 32 nm HKMG 4 (+ IGP) 9 MB 995 millió 216 mm2
Sandy Bridge-E 32 nm HKMG 6 16,5 MB 2,27 milliárd 435 mm2
Gulftown 32 nm HKMG 6 13,5 MB 1,17 milliárd 240 mm2
Lynnfield 45 nm HKMG 4 9 MB 774 millió 296 mm2
Bloomfield 45 nm HKMG 4 9 MB 731 millió 263 mm2
Orochi (Bulldozer) 32 nm HKMG SOI 8 (4 modul) 16 MB ~1,2 milliárd 315 mm2
Llano 32 nm HKMG SOI 4 (+ IGP) 4 MB 1,45 milliárd 228 mm2
Thuban 45 nm SOI 6 9 MB 904 millió 346 mm2
Deneb 45 nm SOI 4 8 MB 758 millió 258 mm2
Table: Main implementation parameters of recent processors [81]
10.3 Supervisory Mode Execution Protection (SMEP)
[83]
10.4 System architecture (1)
[81]
10.4 System architecture (2)
[81]
10.5 Performance (1)
[81]
10.5 Performance (2)
[81]
11. The Haswell line
11. The Haswell line of processors (1)
Expected date of introduction: 4/2013
Figure 1.1: Intel’s Tick-Tock development model [Based on 1]
Tick-Tock Development Model
Merom1
NEWMicroarchitecture
65nm
Penryn
NEWProcess
45nm
Nehalem
NEWMicroarchitecture
45nm
Westmere
NEWProcess
32nm
SandyBridge
NEWMicroarchitecture
32nm
IvyBridge
NEWProcess
22nm
Haswell
NEWMicroarchitecture
22nm
TOCK TOCKTICKTOCKTICKTOCKTICK
11. The Haswell line of processors
11. The Haswell line of processors (2)
The Haswell die [85]
11. The Haswell line of processors (3)
Haswell’s system architecture [86]
11. The Haswell line of processors (4)
[80]
11. The Haswell line of processors (5)
[80][80]
11. The Haswell line of processors (6)
[80]
11. The Haswell line of processors (7)
[80]
11. The Haswell line of processors (8)
[80]
To 12. – Additional references
[81]: Olivera, A régóta várt Intel Ivy Bridge tesztje, Prohardware, 2012-04-13, http://prohardver.hu/teszt/intel_ivy_bridge_teszt/az_ivy_bridge.html
[80]: Chappell R., Toll B., Singhal R.: Intel Next Generation Microarchitecture Codename Haswell: New Processor Innovations, IDF 2012
[82]: Bohr M., Mistry K.: Intel’s Revolutionary 22 nm transistor technology, May 2011, http://download.intel.com/newsroom/kits/22nm/pdfs/22nm-Details_Presentation.pdf
[83]: George V., Piazza T.,Jiang H.: Technology Insight: Intel Next Generation Microarchitecture Codename Ivy Bridge, IDF 2011
[84] 3rd Generation Intel Core Processor Family Quad Core Launch Product Information, April 23, 2012 http://download.intel.com/newsroom/kits/core/3rdgen/pdfs/3rd_Generation _Intel_Core_Product_Information.pdf
[85] Ivy Bridge and Haswell die configurations (estimates included), Anandtech, 03-21-2012, http://forums.anandtech.com/showthread.php?t=2234017
[86]: Piazza T.,Jiang H., Hammerlund P., Singhal R.: Technology Insight: Intel Next Generation Microarchitecture Codename Haswell, IDF 2012 SPCS001