Intel CPU Microarchitecture

http://en.wikipedia.org/wiki/Intel_Tick_Tock

 

http://en.wikipedia.org/wiki/Core_(microarchitecture)

Intel Core (microarchitecture)

From Wikipedia, the free encyclopedia
  (Redirected from Core (microarchitecture))
 
Intel Core
Intel Core 2 Quad.png
L1 cache 64 kB per core
L2 cache 1 MB to 8 MB unified
L3 cache 8 MB to 16 MB shared (Xeon)
Predecessor Enhanced Pentium M
Successor Penryn (tick)
Nehalem (tock)
Socket(s)

The Intel Core microarchitecture (previously known as the Next-Generation Micro-Architecture) is a multi-core processor microarchitecture unveiled by Intelin Q1 2006. It is based on the Yonah processor design and can be considered an iteration of the P6 microarchitecture, introduced in 1995 with Pentium Pro. The high power consumption and heat intensity, the resulting inability to effectively increase clock speed, and other shortcomings such as the inefficient pipeline were the primary reasons for which Intel abandoned the NetBurst microarchitecture and switched to completely different architectural design, delivering high efficiency through a small pipeline rather than high clock speeds. It is worth noting that the Core microarchitecture never reached the clock speeds of the Netburst microarchitecture, even after moving to the 45 nm lithography.

The first processors that used this architecture were code-named MeromConroe, and Woodcrest; Merom is for mobile computing, Conroe is for desktop systems, and Woodcrest is for servers and workstations. While architecturally identical, the three processor lines differ in the socket used, bus speed, and power consumption. Mainstream Core-based processors are branded Pentium Dual-Core or Pentium and low end branded Celeron; server and workstation Core-based processors are branded Xeon, while desktop and mobile Core-based processors are branded as Core 2. Despite their names, processors sold as Core Solo/Core Duo and Core i3/i5/i7 do not actually use the Core microarchitecture and are based on the Enhanced Pentium M and newer Nehalem/Sandy Bridge/Haswellmicroarchitectures, respectively.

 

 

Features[edit]

The Core microarchitecture returned to lower clock rates and improved the usage of both available clock cycles and power when compared with the preceding NetBurst microarchitecture of the Pentium 4/D-branded CPUs.[1] The Core microarchitecture provides more efficient decoding stages, execution units, caches, and buses, reducing the power consumption of Core 2-branded CPUs while increasing their processing capacity. Intel's CPUs have varied widely in power consumption according to clock rate, architecture, and semiconductor process, shown in the CPU power dissipation tables.

Like the last NetBurst CPUs, Core based processors feature multiple cores and hardware virtualization support (marketed as Intel VT-x), as well as Intel 64 and SSSE3. However, Core-based processors do not have the Hyper-Threading Technology found in Pentium 4 processors. This is because the Core microarchitecture is a descendant of the P6 microarchitecture used by Pentium Pro, Pentium II, Pentium III, and Pentium M.

The L1 cache size was enlarged in the Core microarchitecture, from 32KB on Pentium II/III (16 KB L1 Data + 16 KB L1 Instruction) to 64 KB L1 cache/core (32 KB L1 Data + 32 KB L1 Instruction) on Pentium M and Core/Core 2. It also lacks an L3 Cache found in the Gallatin core of the Pentium 4 Extreme Edition, although an L3 Cache is present in high-end versions of Core-based Xeons. Both an L3 cache and Hyper-threading were reintroduced in the Nehalem microarchitecture.

Roadmap[edit]

Technology[edit]

While the Core microarchitecture is a major architectural revision it is based in part on the Pentium M processor family designed by Intel Israel.[2] The Penryn pipeline is 12–14 stages long[3] — less than half of Prescott's, a signature feature of wide order execution cores. Penryn's successor, Nehalem has 16 pipeline stages.[3] Core's execution unit is 4 issues wide, compared to the 3-issue cores of P6Pentium M, and 2-issue cores of NetBurst microarchitectures. The new architecture is a dual core design with linked L1 cache and shared L2 cache engineered for maximum performance per watt and improved scalability.

One new technology included in the design is Macro-Ops Fusion, which combines two x86 instructions into a single micro-operation. For example, a common code sequence like a compare followed by a conditional jump would become a single micro-op.

Other new technologies include 1 cycle throughput (2 cycles previously) of all 128-bit SSE instructions and a new power saving design. All components will run at minimum speed, ramping up speed dynamically as needed (similar to AMD's Cool'n'Quiet power-saving technology, as well as Intel's own SpeedStep technology from earlier mobile processors). This allows the chip to produce less heat, and consume as little power as possible.


Intel Core microarchitecture.

For most Woodcrest CPUs, the front side bus (FSB) runs at 1333 MT/s; however, this is scaled down to 1066 MT/s for lower end 1.60 and 1.86 GHz variants.[4][5] The Merom mobile variant was initially targeted to run at a FSB of 667 MT/s while the second wave of Meroms, supporting 800 MT/s FSB, were released as part of the Santa Rosa platform with a different socket in May 2007. The desktop-oriented Conroe began with models having an FSB of 800 MT/s or 1066 MT/s with a 1333 MT/s line officially launched on July 22, 2007.

The power consumption of these new processors is extremely low—average use energy consumption is to be in the 1–2 watt range in ultra low voltage variants, withthermal design powers (TDPs) of 65 watts for Conroe and most Woodcrests, 80 watts for the 3.0 GHz Woodcrest, and 40 watts for the low-voltage Woodcrest. In comparison, an AMD Opteron 875HE processor consumes 55 watts, while the energy efficient Socket AM2 line fits in the 35 watt thermal envelope (specified a different way so not directly comparable). Merom, the mobile variant, is listed at 35 watts TDP for standard versions and 5 watts TDP for Ultra Low Voltage (ULV) versions.[citation needed]

Previously, Intel announced that it would now focus on power efficiency, rather than raw performance. However, at IDF in the spring of 2006, Intel advertised both. Some of the promised numbers were:

  • 20% more performance for Merom at the same power level (compared to Core Duo)
  • 40% more performance for Conroe at 40% less power (compared to Pentium D)
  • 80% more performance for Woodcrest at 35% less power (compared to the original dual-core Xeon)

Processor cores[edit]

The processors of the Core microarchitecture can be categorized by number of cores, cache size, and socket; each combination of these has a unique code name and product code that is used across a number of brands. For instance, code name "Allendale" with product code 80557 has two cores, 2 MB L2 cache and uses the desktop socket 775, but has been marketed as Celeron, Pentium, Core 2 and Xeon, each with different sets of features enabled. Most of the mobile and desktop processors come in two variants that differ in the size of the L2 cache, but the specific amount of L2 cache in a product can also be reduced by disabling parts at production time. Wolfdale-DP and all quad-core processors except Dunnington QC are multi-chip modules combining two dies. For the 65 nm processors, the same product code can be shared by processors with different dies, but the specific information about which one is used can be derived from the stepping.

 fabcoresMobileDesktop, UP ServerCL ServerDP ServerMP Server
Single-Core 65 nm 65 nm 1 Merom-L
80537
  Conroe-L
80557
       
Single-Core 45 nm 45 nm 1   Penryn-L
80585
      Wolfdale-CL
80588
 
Dual-Core 65 nm 65 nm 2 Merom-2M
80537
Merom
80537
Allendale
80557
Conroe
80557
Conroe-CL
80556
Woodcrest
80556
Tigerton-DC
80564
Dual-Core 45 nm 45 nm 2 Penryn-3M
80577
Penryn
80576
Wolfdale-3M
80571
Wolfdale
80570
Wolfdale-CL
80588
Wolfdale-DP
80573
 
Quad-Core 65 nm 65 nm 4       Kentsfield
80562
  Clovertown
80563
Tigerton
80565
Quad-Core 45 nm 45 nm 4   Penryn-QC
80581
Yorkfield-6M
80580
Yorkfield
80569
Yorkfield-CL
80584
Harpertown
80574
Dunnington QC
80583
Six-Core 45 nm 45 nm 6             Dunnington
80582

Conroe/Merom (65 nm)[edit]

The original Core 2 processors are based around the same dies that can be identified as CPUID Family 6 Model 15. Depending on their configuration and packaging, their code names are Conroe (LGA 775, 4 MB L2 cache), Allendale (LGA 775, 2 MB L2 cache), Merom (Socket M, 4 MB L2 cache) and Kentsfield (Multi-chip module, LGA 775, 2x4MB L2 cache). Merom and Allendale processors with limited features can be found in Pentium Dual Core and Celeron processors, while Conroe, Allendale and Kentsfield also are sold as Xeon processors.

Additional code names for processors based on this model are Woodcrest (LGA 771, 4 MB L2 cache), Clovertown (MCM, LGA 771, 2×4MB L2 cache) and Tigerton (MCM, Socket 604, 2×4MB L2 cache), all of which are marketed only under the Xeon brand.

ProcessorBrand nameModel (list)CoresL2 CacheSocketTDP
Merom-2M Mobile Core 2 Duo U7xxx 2 2 MiB BGA479 10 W
Merom L7xxx 4 MiB 17 W
Merom
Merom-2M
T5xxx
T7xxx
2–4 MiB Socket M
Socket P
BGA479
35 W
Merom Mobile Core 2 Extreme X7xxx 2 4 MiB Socket P 44 W
Merom Celeron M 5x0 1 512 KiB Socket M
Socket P
30 W
Merom-2M 5x5 Socket P 31 W
Merom-2M Celeron Dual-Core T1xxx 2 512–1024 KiB Socket P 35 W
Merom-2M Pentium Dual-Core T2xxx
T3xxx
2 1 MiB Socket P 35 W
Allendale Xeon 3xxx 2 2 MB LGA 775 65 W
Conroe 3xxx 2–4 MB
Conroe and
Allendale
Core 2 Duo E4xxx 2 2 MB LGA 775 65 W
E6xx0 2–4 MB
Conroe-CL E6xx5 2–4 MB LGA 771
Conroe-XE Core 2 Extreme X6xxx 2 4 MB LGA 775 75 W
Allendale Pentium Dual-Core E2xxx 2 1 MB LGA 775 65 W
Allendale Celeron E1xxx 2 512 kB LGA 775 65 W
Kentsfield Xeon 32xx 4 2×4 MiB LGA 775 95–105 W
Kentsfield Core 2 Quad Q6xxx 4 2×4 MiB LGA 775 95–105 W
Kentsfield XE Core 2 Extreme QX6xxx 4 2×4 MiB LGA 775 130 W
Woodcrest Xeon 51xx 2 4 MB LGA 771 65–80 W
Clovertown L53xx 4 2×4 MB LGA 771 40–50 W
E53xx 80 W
X53xx 120–150 W
Tigerton-DC E72xx 2 2×4 MB Socket 604 80 W
Tigerton L73xx 4 50 W
E73xx 2×2–2×4 MB 80 W
X73xx 2×4 MB 130 W

Conroe-L/Merom-L[edit]

The Conroe-L and Merom-L processors are based around the same core as Conroe and Merom, but only contain a single core and 1 MB of L2 cache, significantly reducing production cost and power consumption of the processor at the expense of performance compared to the dual-core version. It is used only in ultra-low voltage Core 2 Solo U2xxx and in Celeron processors and is identified as CPUID family 6 model 22.

ProcessorBrand nameModel (list)CoresL2 CacheSocketTDP
Merom-L Mobile Core 2 Solo U2xxx 1 2 MiB BGA479 5.5 W
Merom-L Celeron M 5x0 1 512 KiB Socket M
Socket P
27 W
Merom-L 5x3 512–1024 KiB BGA479 5.5–10 W
Conroe-L Celeron M 4x0 1 512 KiB LGA 775 35 W
Conroe-CL 4x5 LGA 771 65 W

Penryn/Wolfdale (45 nm)[edit]

In Intel's Tick-Tock cycle, the 2007/2008 "Tick" was the shrink of the Core microarchitecture to 45 nanometers as CPUID model 23. In Core 2 processors, it is used with the code names Penryn (Socket P), Wolfdale (LGA 775) and Yorkfield (MCM, LGA 775), some of which are also sold as Celeron, Pentium and Xeon processors. In the Xeon brand, the Wolfdale-DP and Harpertown code names are used for LGA 771 based MCMs with two or four active Wolfdale cores.

The chips come in two sizes, with 6 MB and 3 MB L2 cache. The smaller version is commonly called Penryn-3M and Wolfdale-3M as well as Yorkfield-6M, respectively. The single-core version of Penryn, listed as Penryn-L here, is not a separate model like Merom-L but a version of the Penryn-3M model with only one active core.

ProcessorBrand nameModel (list)CoresL2 CacheSocketTDP
Penryn-L Core 2 Solo SU3xxx 1 3 MiB BGA956 5.5 W
Penryn-3M Core 2 Duo SU7xxx 2 3 MB BGA956 10 W
SU9xxx
Penryn SL9xxx 6 MiB 17 W
SP9xxx 25/28 W
Penryn-3M P7xxx 3 MiB Socket P
FCBGA6
25 W
P8xxx
Penryn P9xxx 6 MiB
Penryn-3M T6xxx 2 MiB 35 W
T8xxx 3 MiB
Penryn T9xxx 6 MiB
E8x35 6 MiB Socket P 35-55 W
Penryn-QC Core 2 Quad Q9xxx 4 2x3-2x6 MiB Socket P 45 W
Penryn XE Core 2 Extreme X9xxx 2 6 MiB Socket P 44 W
Penryn-QC QX9xxx 4 2x6 MiB 45 W
Penryn-3M Celeron T3xxx 2 1 MiB Socket P 35 W
SU2xxx µFC-BGA 956 10 W
Penryn-L 9x0 1 1 MiB Socket P 35 W
7x3 µFC-BGA 956 10 W
Penryn-3M Pentium T4xxx 2 1 MiB Socket P 35 W
SU4xxx 2 MiB µFC-BGA 956 10 W
Penryn-L SU2xxx 1 5.5 W
Wolfdale-3M
Celeron E3xxx 2 1 MB LGA 775 65 W
Pentium E2210
E5xxx 2 MB
E6xxx
Core 2 Duo E7xxx 3 MB
Wolfdale E8xxx 6 MB
Xeon 31x0 45-65 W
Wolfdale-CL 30x4 1 LGA 771 30 W
31x3 2 65 W
Yorkfield Xeon X33x0 4 2×3–2×6 MB LGA 775 65–95 W
Yorkfield-CL X33x3 LGA 771 80 W
Yorkfield-6M Core 2 Quad Q8xxx 2×2 MB LGA 775 65–95 W
Q9x0x 2×3 MB
Yorkfield Q9x5x 2×6 MB
Yorkfield XE Core 2 Extreme QX9xxx 2×6 MB 130–136 W
QX9xx5 LGA 771 150 W
Wolfdale-DP Xeon E52xx 2 6 MB LGA 771 65 W
L52xx 20-55 W
X52xx 80 W
Harpertown E54xx 4 2×6 MB LGA 771 80 W
L54xx 40-50 W
X54xx 120-150 W

Dunnington[edit]

The Xeon "Dunnington" processor (CPUID Family 6, model 30) is closely related to Wolfdale but comes with six cores and an on-chip L3 cache and is designed for servers with Socket 604, so it is marketed only as Xeon, not as Core 2.

ProcessorBrand nameModel (list)CoresL3 CacheSocketTDP
Dunnington Xeon E74xx 4-6 8-16 MB Socket 604 90 W
L74xx 4-6 12 MB 50-65 W
X7460 6 16 MB 130 W

Steppings[edit]

The Core microarchitecture uses a number of steppings, which unlike previous microarchitectures not only represent incremental improvements but also different sets of features like cache size and low power modes. Most of these steppings are used across brands, typically by disabling some of the features and limiting clock frequencies on low-end chips.

Steppings with a reduced cache size use a separate naming scheme, which means that the releases are no longer in alphabetic order. Additional steppings have been used in internal and engineering samples, but are not listed in the tables.

Many of the high-end Core 2 and Xeon processors use Multi-Chip Modules of two or three chips in order to get larger cache sizes or more than two cores.

Steppings using 65 nm process[edit]

  Mobile (Merom)Desktop (Conroe)Desktop (Kentsfield)Server (WoodcrestClovertown,Tigerton)
SteppingReleasedAreaCPUIDL2 cacheMax. clockCeleronPentiumCore 2CeleronPentiumCore 2XeonCore 2XeonXeon
B2 Jul 2006 143 mm² 06F6 4 MiB 2.93 GHz M5xx   T5000 T7000L7000     E6000X6000 3000     5100
B3 Nov 2006 143 mm² 06F7 4 MiB 3.00 GHz               Q6000QX6000 3200 5300
L2 Jan 2007 111 mm² 06F2 2 MiB 2.13 GHz     T5000 U7000   E2000 E4000E6000 3000      
E1 May 2007 143 mm² 06FA 4 MiB 2.80 GHz M5xx   T7000 L7000X7000              
G0 Apr 2007 143 mm² 06FB 4 MiB 3.00 GHz M5xx   T7000 L7000X7000   E2000 E4000E6000 3000 Q6000QX6000 3200 5100 5300 7200 7300
G2 Mar 2009 143 mm² 06FB 4 MiB 2.16 GHz M5xx   T5000 T7000L7000              
M0 Jul 2007 111 mm² 06FD 2 MiB 2.40 GHz 5xxT1000 T2000T3000 T5000 T7000U7000 E1000 E2000 E4000        
A1 Jun 2007 81 mm² 10661 1 MiB 2.20 GHz M5xx   U2000 220 4x0            

Steppings B2/B3, E1 and G0 of model 15 (cpuid 06fx) processors are evolutionary steps of the standard Merom/Conroe die with 4 MiB L2 cache, with the short-lived E1 stepping only being used in mobile processors. Stepping L2 and M0 are the "Allendale" chips with just 2 MiB L2 cache, reducing production cost and power consumption for low-end processors.

The G0 and M0 steppings improve idle power consumption in C1E state and add the C2E state in desktop processors. In mobile processors, all of which support C1 through C4 idle states, steppings E1, G0, and M0 add support for the Mobile Intel 965 Express (Santa Rosa) platform with Socket P, while the earlier B2 and L2 steppings only appear for the Socket M based Mobile Intel 945 Express (Napa refresh) platform.

The model 22 stepping A1 (cpuid 10661h) marks a significant design change, with just a single core and 1 MiB L2 cache further reducing the power consumption and manufacturing cost for the low-end. Like the earlier steppings, A1 is not used with the Mobile Intel 965 Express platform.

Steppings G0, M0 and A1 mostly replaced all older steppings in 2008. In 2009, a new stepping G2 was introduced to replace the original stepping B2.[6]

Steppings using 45 nm process[edit]

  Mobile (Penryn)Desktop (Wolfdale)Desktop (Yorkfield)Server (Wolfdale-DPHarpertown,Dunnington)
SteppingReleasedAreaCPUIDL2 cacheMax. clockCeleronPentiumCore 2CeleronPentiumCore 2XeonCore 2XeonXeon
C0 Nov 2007 107 mm² 10676 6 MiB 3.00 GHz     E8000 P7000 T8000 T9000 P9000SP9000 SL9000 X9000     E8000 3100 QX9000   5200 5400
M0 Mar 2008 82 mm² 10676 3 MiB 2.40 GHz 7xx   SU3000 P7000 P8000 T8000SU9000   E5000E2000 E7000        
C1 Mar 2008 107 mm² 10677 6 MiB 3.20 GHz               Q9000 QX9000 3300  
M1 Mar 2008 82 mm² 10677 3 MiB 2.50 GHz               Q8000 Q9000 3300  
E0 Aug 2008 107 mm² 1067A 6 MiB 3.33 GHz     T9000 P9000 SP9000 SL9000Q9000 QX9000     E8000 3100 Q9000 Q9000SQX9000 3300 5200 5400
R0 Aug 2008 82 mm² 1067A 3 MiB 2.93 GHz 7xx 900SU2000T3000 T4000SU2000SU4000 SU3000 T6000 SU7000 P8000SU9000 E3000 E5000E6000 E7000   Q8000 Q8000SQ9000 Q9000S 3300  
A1 Sep 2008 503 mm² 106D1 3 MiB 2.67 GHz                   7400

In the model 23 (cpuid 01067xh), Intel started marketing stepping with full (6 MiB) and reduced (3 MiB) L2 cache at the same time, and giving them identical cpuid values. All steppings have the new SSE4.1instructions. Stepping C1/M1 was a bug fix version of C0/M0 specifically for quad core processors and only used in those. Stepping E0/R0 adds two new instructions (XSAVE/XRSTOR) and replaces all earlier steppings.

In mobile processors, stepping C0/M0 is only used in the Intel Mobile 965 Express (Santa Rosa refresh) platform, whereas stepping E0/R0 supports the later Intel Mobile 4 Express (Montevina) platform.

Model 30 stepping A1 (cpuid 106d1h) adds an L3 cache as well as six instead of the usual two cores, which leads to an unusually large die size of 503 mm².[7] As of February 2008, it has only found its way into the very high-end Xeon 7400 series (Dunnington).

System requirements[edit]

Motherboard compatibility[edit]

Conroe, Conroe XE and Allendale all use Socket LGA 775; however, not every motherboard is compatible with these processors.

Supporting chipsets are:

See also: List of Intel chipsets

The currently released Yorkfield XE model QX9770 (45 nm with 1600FSB) currently has limited chipset compatibility - with only X38, P35 (With Overclocking) and some high-performance X48 and P45 motherboards being compatible. BIOS updates are gradually being released to provide support for the new Penryn technology, and the new QX9775 is only compatible with D5400XS. The Wolfdale-3M model E7200 also has limited compatibility (at least the Xpress 200 chipset is incompatible)[citation needed].

Although a motherboard may have the required chipset to support Conroe, some motherboards based on the above mentioned chipsets do not support Conroe. This is because all Conroe-based processors require a new power delivery feature set specified in Voltage Regulator-Down (VRD) 11.0. This requirement is a result of Conroe's significantly lower power consumption, compared to the Pentium 4/D CPUs it is replacing. A motherboard that has both a supporting chipset and VRD 11 supports Conroe processors, but even then some boards will need an updated BIOS to recognize Conroe's FID (Frequency ID) and VID (Voltage ID).

Synchronous memory modules[edit]

Unlike the previous Pentium 4 and Pentium D design, the Core 2 technology sees a greater benefit from memory running synchronously with the Front Side Bus (FSB). This means that for the Conroe CPUs with FSB of 1066 MT/s, the ideal memory performance for DDR2 is PC2-8500. In a few configurations, using PC2-5300 instead of PC2-4200 can actually decrease performance. Only when going to PC2-6400is there a significant performance increase. While DDR2 memory models with tighter timing specifications do improve performance, the difference in real world games and applications is often negligible.[8]

Optimally, the memory bandwidth afforded should match the bandwidth of the FSB, that is to say that a CPU with a 533 MT/s rated bus speed should be paired with RAM matching the same rated speed, for example DDR2 533, or PC2-4200. A common myth[citation needed] is that installing interleaved RAM will offer double the bandwidth. However, at most the increase in bandwidth by installing interleaved RAM is roughly 5–10%. The AGTL+ PSB[dead link] used by all NetBurst processors as well as current and medium-term (pre-QuickPath) Core 2 processors provide a 64-bit data path. Current chipsets provide for a couple of either DDR2 or DDR3 channels.

Matched processor and RAM ratings
Processor modelFront side busMatched memory and maximum bandwidth
single channel / dual channel
DDRDDR2DDR3
mobile: T5200, T5300, U2n00, U7n00 533 MT/s PC-3200 (DDR-400)
3.2 GB/s
PC2-4200 (DDR2-533)
4.264 GB/s
PC2-8500 (DDR2-1066)
8.532 GB/s
PC3-8500 (DDR3-1066)
8.530 GB/s
desktop: E6n00, E6n20, X6n00, E7n00, Q6n00 and QX6n00
mobile: T9400, T9550, T9600, P7350, P7450, P8400, P8600, P8700, P9500, P9600, SP9300, SP9400, X9100
1066 MT/s
mobile: T5n00, T5n50, T7n00 (Socket M), L7200, L7400 667 MT/s PC-3200 (DDR-400)
3.2 GB/s
PC2-5300 (DDR2-667)
5.336 GB/s
PC3-10600 (DDR3-1333)
10.670 GB/s
desktop: E6n40, E6n50, E8nn0, Q9nn0, QX6n50, QX9650 1333 MT/s
mobile: T5n70, T6400, T7n00 (Socket P), L7300, L7500, X7n00, T8n00, T9300, T9500, X9000
desktop: E4n00, Pentium E2nn0, Pentium E5nn0, Celeron 4n0, E3n00
800 MT/s PC-3200 (DDR-400)
3.2 GB/s
PC-3200 (DDR-400)
3.2 GB/s
PC2-6400 (DDR2-800)
6.400 GB/s
PC2-8500 (DDR2-1066)
8.532 GB/s
PC3-6400 (DDR3-800)
6.400 GB/s
PC3-12800 (DDR3-1600)
12.800 GB/s
desktop: QX9770, QX9775 1600 MT/s

On jobs requiring large amounts of memory access, the quad-core Core 2 processors can benefit significantly[9] from using a PC2-8500 memory, which runs exactly the same speed as the CPU's FSB; this is not an officially supported configuration, but a number of motherboards offer it.

The Core 2 processor does not require the use of DDR2. While the Intel 975X and P965 chipsets require this memory, some motherboards and chipsets support both the Core 2 and DDR memory. When using DDR memory, performance may be reduced because of the lower available memory bandwidth.

Chip errata[edit]

The Core 2 memory management unit (MMU) in X6800, E6000 and E4000 processors does not operate to previous specifications implemented in previous generations of x86 hardware. This may cause problems, many of them serious security and stability issues, with existing operating system software. Intel's documentation states that their programming manuals will be updated "in the coming months" with information on recommended methods of managing the translation lookaside buffer (TLB) for Core 2 to avoid issues, and admits that, "in rare instances, improper TLB invalidation may result in unpredictable system behavior, such as hangs or incorrect data."[10]

Among the issues noted:

  • Non-execute bit is shared across the cores.
  • Floating point instruction non-coherencies.
  • Allowed memory corruptions outside of the range of permitted writing for a process by running common instruction sequences.

Intel errata Ax39, Ax43, Ax65, Ax79, Ax90, Ax99 are said to be particularly serious.[11] 39, 43, 79, which can cause unpredictable behavior or system hang, have been fixed in recent steppings.

Among those who have noted the errata to be particularly serious are OpenBSD's Theo de Raadt[12] and DragonFly BSD's Matthew Dillon.[13] Taking a contrasting view was Linus Torvalds, calling the TLB issue "totally insignificant", adding, "The biggest problem is that Intel should just have documented the TLB behavior better."[14]

Microsoft has issued update KB936357 to address the errata by microcode update,[15] with no performance penalty. BIOS updates are also available to fix the issue.

Key Terms[edit]

MT/s. Millions of transfers/second, each transfer on the Intel Core architecture is 32-bits.

See also[edit]

 

 

 

http://en.wikipedia.org/wiki/Penryn_(microarchitecture)

Penryn (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Penryn
Intel Core2 Qu2009.png
L1 cache 64 KB per core
L2 cache 3 MB to 12 MB unified
L3 cache 8 MB to 16 MB shared (Xeon)
Predecessor Core
Successor Nehalem
Socket(s)

In Intel's Tick-Tock cycle, the 2007/2008 "Tick" was the shrink of the Core microarchitecture to 45 nanometers as CPUID model 23. In Core 2 processors, it is used with the code names Penryn (Socket P), Wolfdale (LGA 775) and Yorkfield (MCM, LGA 775), some of which are also sold as Celeron, Pentium and Xeon processors. In the Xeon brand, the Wolfdale-DP and Harpertown code names are used for LGA 771 based MCMs with two or four active Wolfdale cores.

The chips come in two sizes, with 6 MB and 3 MB L2 cache. The smaller version is commonly called Penryn-3M and Wolfdale-3M as well as Yorkfield-6M, respectively. The single-core version of Penryn, listed as Penryn-L here, is not a separate model like Merom-L but a version of the Penryn-3M model with only one active core.

 

 

CPU List[edit]

ProcessorBrand nameModel (list)CoresL2 CacheSocketTDP
Penryn-L Core 2 Solo SU3xxx 1 3 MiB BGA956 5.5 W
Penryn-3M Core 2 Duo SU7xxx 2 3 MB BGA956 10 W
SU9xxx
Penryn SL9xxx 6 MiB 17 W
SP9xxx 25/28 W
Penryn-3M P7xxx 3 MiB Socket P
FCBGA6
25 W
P8xxx
Penryn P9xxx 6 MiB
Penryn-3M T6xxx 2 MiB 35 W
T8xxx 3 MiB
Penryn T9xxx 6 MiB
E8x35 6 MiB Socket P 35-55 W
Penryn-QC Core 2 Quad Q9xxx 4 2x3-2x6 MiB Socket P 45 W
Penryn XE Core 2 Extreme X9xxx 2 6 MiB Socket P 44 W
Penryn-QC QX9xxx 4 2x6 MiB 45 W
Penryn-3M Celeron T3xxx 2 1 MiB Socket P 35 W
SU2xxx µFC-BGA 956 10 W
Penryn-L 9x0 1 1 MiB Socket P 35 W
7x3 µFC-BGA 956 10 W
Penryn-3M Pentium T4xxx 2 1 MiB Socket P 35 W
SU4xxx 2 MiB µFC-BGA 956 10 W
Penryn-L SU2xxx 1 5.5 W
Wolfdale-3M
Celeron E3xxx 2 1 MB LGA 775 65 W
Pentium E2210
E5xxx 2 MB
E6xxx
Core 2 Duo E7xxx 3 MB
Wolfdale E8xxx 6 MB
Xeon 31x0 45-65 W
Wolfdale-CL 30x4 1 LGA 771 30 W
31x3 2 65 W
Yorkfield Xeon X33x0 4 2×3–2×6 MB LGA 775 65–95 W
Yorkfield-CL X33x3 LGA 771 80 W
Yorkfield-6M Core 2 Quad Q8xxx 2×2 MB LGA 775 65–95 W
Q9x0x 2×3 MB
Yorkfield Q9x5x 2×6 MB
Yorkfield XE Core 2 Extreme QX9xxx 2×6 MB 130–136 W
QX9xx5 LGA 771 150 W
Wolfdale-DP Xeon E52xx 2 6 MB LGA 771 65 W
L52xx 20-55 W
X52xx 80 W
Harpertown E54xx 4 2×6 MB LGA 771 80 W
L54xx 40-50 W
X54xx 120-150 W

Processor cores[edit]

The processors of the Core microarchitecture can be categorized by number of cores, cache size, and socket; each combination of these has a unique code name and product code that is used across a number of brands. For instance, code name "Allendale" with product code 80557 has two cores, 2 MB L2 cache and uses the desktop socket 775, but has been marketed as Celeron, Pentium, Core 2 and Xeon, each with different sets of features enabled. Most of the mobile and desktop processors come in two variants that differ in the size of the L2 cache, but the specific amount of L2 cache in a product can also be reduced by disabling parts at production time. Wolfdale-DP and all quad-core processors except Dunnington QC are multi-chip modules combining two dies. For the 65 nm processors, the same product code can be shared by processors with different dies, but the specific information about which one is used can be derived from the stepping.

 fabcoresMobileDesktop, UP ServerCL ServerDP ServerMP Server
Single-Core 45 nm 45 nm 1   Penryn-L
80585
      Wolfdale-CL
80588
 
Dual-Core 45 nm 45 nm 2 Penryn-3M
80577
Penryn
80576
Wolfdale-3M
80571
Wolfdale
80570
Wolfdale-CL
80588
Wolfdale-DP
80573
 
Quad-Core 45 nm 45 nm 4   Penryn-QC
80581
Yorkfield-6M
80580
Yorkfield
80569
Yorkfield-CL
80584
Harpertown
80574
Dunnington QC
80583
Six-Core 45 nm 45 nm 6             Dunnington
80582

Steppings using 45 nm process[edit]

  Mobile (Penryn)Desktop (Wolfdale)Desktop (Yorkfield)Server (Wolfdale-DP,HarpertownDunnington)
SteppingReleasedAreaCPUIDL2 cacheMax. clockCeleronPentiumCore 2CeleronPentiumCore 2XeonCore 2XeonXeon
C0 Nov 2007 107 mm² 10676 6 MiB 3.00 GHz     E8000 P7000 T8000 T9000 P9000 SP9000SL9000 X9000     E8000 3100 QX9000   5200 5400
M0 Mar 2008 82 mm² 10676 3 MiB 2.40 GHz 7xx   SU3000 P7000 P8000 T8000 SU9000   E5000E2000 E7000        
C1 Mar 2008 107 mm² 10677 6 MiB 3.20 GHz               Q9000 QX9000 3300  
M1 Mar 2008 82 mm² 10677 3 MiB 2.50 GHz               Q8000 Q9000 3300  
E0 Aug 2008 107 mm² 1067A 6 MiB 3.33 GHz     T9000 P9000 SP9000 SL9000 Q9000QX9000     E8000 3100 Q9000 Q9000SQX9000 3300 5200 5400
R0 Aug 2008 82 mm² 1067A 3 MiB 2.93 GHz 7xx 900 SU2000T3000 T4000 SU2000SU4000 SU3000 T6000 SU7000 P8000 SU9000 E3000 E5000E6000 E7000   Q8000 Q8000S Q9000Q9000S 3300  
A1 Sep 2008 503 mm² 106D1 3 MiB 2.67 GHz                   7400

In the model 23 (cpuid 01067xh), Intel started marketing stepping with full (6 MiB) and reduced (3 MiB) L2 cache at the same time, and giving them identical cpuid values. All steppings have the new SSE4.1instructions. Stepping C1/M1 was a bug fix version of C0/M0 specifically for quad core processors and only used in those. Stepping E0/R0 adds two new instructions (XSAVE/XRSTOR) and replaces all earlier steppings.

In mobile processors, stepping C0/M0 is only used in the Intel Mobile 965 Express (Santa Rosa refresh) platform, whereas stepping E0/R0 supports the later Intel Mobile 4 Express (Montevina) platform.

Model 30 stepping A1 (cpuid 106d1h) adds an L3 cache as well as six instead of the usual two cores, which leads to an unusually large die size of 503 mm².[1] As of February 2008, it has only found its way into the very high-end Xeon 7400 series (Dunnington).

Roadmap[edit]

See also[edit]

 

 

 

http://en.wikipedia.org/wiki/Nehalem_(microarchitecture)

Nehalem (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Nehalem
Intel Nehalem.jpg
L1 cache 64 KB per core
L2 cache 256 KB per core
L3 cache 4 MB to 12 MB shared
Predecessor Core (tock)
Penryn (tick)
Successor Westmere (tick)
Sandy Bridge (tock)
Socket(s)

Nehalem /nəˈhləm/[1] is the codename for an Intel processor microarchitecture, successor to the Core microarchitecture.[2] Nehalem processors use the 45 nmprocess. A preview system with two Nehalem processors was shown at Intel Developer Forum in 2007. The first processor released with the Nehalem architecture was the desktop Core i7,[3] which was released in November 2008.

Nehalem, a recycled Intel code name, refers to an architecture that differs radically from Netburst, while retaining some of the latter's minor features. Nehalem-based microprocessors use higher clock speeds and are more energy-efficient than Penryn microprocessors. Hyper-threading is reintroduced, along with a reduction in L2 cache size, as well as an enlarged L3 cache that is shared by all cores. Nehalem was replaced with the Sandy Bridge microarchitecture, released in January 2011.

 

 

Technology[edit]


Microarchitecture of a processor core in the quad-core implementation


TLB Sizes[5]

CachePage Size
NameLevel4 kB2 MB
DTLB 1st 64 32
ITLB 1st 128 7/logical core
STLB 2nd 512 none

Performance and power improvements[edit]

It has been reported that Nehalem has a focus on performance, thus the increased core size.[6] Compared to Penryn, Nehalem has:

  • 10-25% more single-threaded performance / 20-100% more multithreaded performance at the same power level
  • 30% lower power consumption for the same performance
  • Nehalem provides a 15–20% clock-for-clock increase in performance per core(average)

Overclocking is possible with Bloomfield processors and the X58 chipset. Lynnfield processors use a PCH removing the need for a northbridge chipset.[7]

Nehalem processors incorporate SSE 4.2 SIMD instructions, adding 7 new instructions to the SSE 4.1 set in the Core 2 series. The Nehalem architecture reduces atomic operation latency by 50% in an attempt to eliminate atomic overhead .[8]

Variants Overview[edit]

Processing Cores (interface)ProcessDie SizeCPUIDModelSteppingMobileDesktop, UP ServerDP ServerMP Server
Eight-Core (Quad-Channel) 45 nm 684 mm² 206E6 46 D0       Beckton (80604)
Quad-Core (Triple-Channel) 45 nm 263 mm² 106A4
106A5
26 C0
D0
  Bloomfield (80601) Gainestown (80602)  
Quad-Core (Dual-Channel, PCIe) 45 nm 296 mm² 106E4
106E5
30 B0
B1
Clarksfield (80607) Lynnfield (80605) Jasper Forest (80612)  
Dual-Core (Dual-Channel, PCIe, Graphics Core) 45 nm         Auburndale (canceled) Havendale (canceled)    

 

  • Lynnfield processors feature integrated PCIe 1 x16 or 2 x8.
  • 1 6500 series scalable up to 2 sockets, 7500 series scalable up to 4/8 sockets.[9]

Server and desktop processors[edit]

CodenameMarketCores /
Threads
SocketProcessor
Branding & Model
CPU
Clock rate
TurboTDPInterfacesL3
cache
Release DatePrice for
1k Unit
ChipsetMemory
Beckton1 MP Server /
DP Server
8 (16) LGA
1567
Xeon [10] X7560 2.26 GHz Yes 130 W 4× QPI 6.4 GT/s DDR3-800 /
1066
(Up to 4x with
SMB-Ready
Motherboard)
24 MB 2010-03-30[11] $3692
X7550 2.0 GHz 18 MB $2837
X6550 $2461
L7555 1.86 GHz 95 W 4× QPI 5.86 GT/s 24 MB $3157
6 (12) E7540 2.0 GHz 105 W 4× QPI 6.4 GT/s 18 MB $1980
E6540 12 MB $1712
E7530 1.86 GHz 4× QPI 5.86 GT/s $1391
L7545 18 MB $2087
6 (6) X7542 2.66 GHz 130 W $1980
4 (8) E7520 1.86 GHz No 105 W 4× QPI 4.8 GT/s $856
E6510 1.73 GHz 12 MB $744
Gainestown DP Server[12] 4 (8) LGA
1366
Xeon[13] W5590 3.33 GHz Yes 130 W 2× QPI 6.4 GT/s 3× DDR3-13331 8 MB 2009-08-09 $1600
W5580 3.2 GHz 2009-03-29[14] $1500
X5570 2.93 GHz 95 W $1286
X5560 2.8 GHz $1072
X5550 2.66 GHz $858
E5540 2.53 GHz 80 W 2× 5.86 GT/s 3× DDR3-10661 $744
E5530 2.4 GHz $530
E5520 2.26 GHz $373
L5530 2.4 GHz 60 W 2009-08-09 $744
L5520 2.26 GHz 2009-03-30 $530
L5518 2.13 GHz $
4 (4) E5507 2.26 GHz No 80 W 2× 4.8 GT/s 3× DDR3-8001 4 MB 2010-03-16 $266
E5506 2.13 GHz 2009-03-29
L5506 2.13 GHz 60 W $423
E5504 2.0 GHz 80 W $224
2 (4) L5508 2.0 GHz Yes 38 W 2× 5.86 GT/s 3× DDR3-1066 8 MB $
2 (2) E5503 2.0 GHz No 80 W 2× 4.8 GT/s 3× DDR3-800 4 MB 2010-03-16 $224
E5502 1.86 GHz 2009-03-29 $188
Bloomfield UP Server[15] 4 (8) Xeon[16] W3580 3.33 GHz Yes 130 W 1× QPI 6.4 GT/s 3× DDR3-1333 8 MB 2009-08-09 $999
W3570 3.2 GHz 2009-03-29[16]
W3565 3.2 GHz 1× QPI 4.8 GT/s 3× DDR3-1066 2009-11-01 $562
W3550 3.06 GHz 2009-08-09
W3540 2.93 GHz 2009-03-29[16]
W3530 2.8 GHz 2010-03-16 $294
W3520 2.66 GHz 2009-03-29[16] $284
2(2) W3505 2.53 GHz No 4 MB $
W3503 2.4 GHz $
Lynnfield 4 (8) LGA
1156
X3480 3.06 GHz Yes 95 W DMI 2× DDR3-1333 8 MB 2010-05-30 $612
X3470 2.93 GHz 2009-09-08 $589
X3460 2.8 GHz $316
X3450 2.66 GHz $241
X3440 2.53 GHz $215
L3426 1.86 GHz 45 W $284
4 (4) X3430 2.4 GHz 95 W $189
Bloomfield Enthusiast
Desktop[17]
4 (8) LGA
1366
Core i7
Extreme
975[18] 3.33 GHz Yes 130 W 1× QPI 6.4 GT/s 3× DDR3-1066 2009-05-31 $999
965 3.2 GHz 2008-11-17
Core i7 960[19] 3.2 GHz 1× QPI 4.8 GT/s 2009-10-20 $562
950[18] 3.06 GHz 2009-05-31
940 2.93 GHz 2008-11-17
930 2.8 GHz 2010-02-28 $294
920 2.66 GHz 2008-11-17 $284
Lynnfield Performance
Desktop
LGA
1156
880 3.06 GHz Yes 95 W DMI 2× DDR3-1333 2010-05-30 $583
875K 2.93 GHz $342
870[20] 2009-09-08 $562
870S 2.66 GHz 82 W 2010-07-19 $351
860 2.8 GHz 95 W 2009-09-08 $284
860S 2.53 GHz 82 W 2010-01-07 $337
4 (4) Core i5 760 2.8 GHz 95 W 2010-07-17 $209
750[21] 2.66 GHz 95 W 2009-09-08 $196
750S 2.4 GHz 82 W 2010-01-07 $259
  • Intel states the Gainestown processors have six memory channels. Gainestown processors have dual QPI links and have a separate set of memory registers for each link in effect, a multiplexed six-channel system.[22][23]

Mobile processors[edit]

CodenameMarketCores /
Threads
SocketProcessor
Branding & Model
Core
Clock rate
TurboTDPL3
cache
InterfaceRelease DatePrice for
1k Unit
Clarksfield Extreme /
Performance
Mobile
4 (8) µPGA
988
Core i7
Extreme
940XM 2.13 GHz Yes 55 W 8 MB * DMI
* 2x DDR3-1333
PCIe 1 x16 / 2 x8
2010-06-21 $1096
920XM 2.0 GHz 2009-09-23 $1054
Core i7 840QM 1.86 GHz 45 W 2010-06-21 $568
820QM 1.73 GHz 2009-09-23 $546
740QM 6 MB 2010-06-21 $378
720QM 1.6 GHz 2009-09-23 $364

Roadmap[edit]

The successor to Nehalem and Westmere is Sandy Bridge.

See also[edit]

 

 

 

http://en.wikipedia.org/wiki/Westmere_(microarchitecture)

Westmere (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Westmere
L1 cache 64KB per core
L2 cache 256KB per core
L3 cache 4MB to 30MB shared
GPU 533 MHz to 900 MHz
177M 45nm (K0)
Predecessor Nehalem
Successor Sandy Bridge
Socket(s)
Connection of the GPU inside the Westemere microarchitecture

Westmere (formerly Nehalem-C) is the name given to the 32 nm die shrink of Nehalem. The first Westmere-based processors were launched on January 7, 2010 by Intel Corporation.

Westmere's feature improvements from Nehalem as reported:

  • Native six-core (Gulftown) and ten-core (Westmere-EX) processors.[1]
  • A new set of instructions that gives over 3x the encryption and decryption rate of Advanced Encryption Standard (AES) processes compared to before.[2]
    • Delivers seven new instructions (AES instruction set or AES-NI) that will be used by the AES algorithm. Also an instruction called PCLMULQDQ (seeCLMUL instruction set) that will perform carry-less multiplication for use in cryptography.[3] These instructions will allow the processor to perform hardware-accelerated encryption, not only resulting in faster execution but also protecting against software targeted attacks.
  • Integrated graphics, added into the processor package (dual core Arrandale and Clarkdale only).
  • Improved virtualization latency.[4]
  • New virtualization capability: "VMX Unrestricted mode support," which allows 16-bit guests to run (real mode and big real mode).
  • Support for "Huge Pages" of 1 GB in size.


TLB Sizes[5]

CachePage Size
NameLevel4 kB2 MB1 GB
DTLB 1st 64 32  ?
ITLB 1st 128 7/logical core  ?
STLB 2nd 512 none none

 

 

CPU Variants[edit]

Processing Cores (interface)ProcessDie SizeCPUIDModelSteppingMobileDesktop, UP ServerDP ServerMP Server
Ten-Core (Quad-channel)[6] 32 nm 513 mm² 206F2 47 A2       Westmere-EX (80615)
Six-Core (Triple-Channel) 32 nm 248 mm² 206C2 44 B1   Gulftown (80613) Westmere-EP (80614)  
Dual-Core (Dual-Channel, PCIe, Graphics Core) 32 nm
45 nm
81+114 mm² 20652
20655
37 C2
K0
Arrandale (80617) Clarkdale (80616)    

Westmere CPUs[edit]

  • TDP includes the integrated GPU, if present.
  • Clarkdale processors feature an integrated PCIe 1 ×16.
  • Clarkdale and Arrandale contain the 32 nm dual core processor Hillel and the 45 nm integrated graphics device Ironlake, and support switchable graphics.[7][8]

Server / Desktop Processors[edit]

CodenameMarketCores /
Threads
SocketProcessor
Branding & model
Clock rateTurboTDPInterfacesL3
cache
Release
Date
Price
CoreGPUChipsetMemory
Westmere-EX
[9]
MP Server 10 (20) LGA
1567
Xeon E7-8870 2.4 GHz N/A Yes 130 W 4× QPI 6.4 GT/s 4× DDR3-1066 30 MB 2011-04-05
[10]
$4616
E7-4870 $4394
E7-2870 $4227
E7-8867L 2.13 GHz 105 W $4172
E7-8860 2.26 GHz 130 W 24 MB $4061
E7-4860 $3838
E7-2860 $3670
E7-8850 2 GHz $3059
E7-4850 $2837
E7-2850 $2558
8 (8) E7-8837 2.66 GHz $2280
8 (16) E7-8830 2.13 GHz 105 W
E7-4830 $2059
E7-2830 $1779
E7-4820 2 GHz 4× QPI 5.86 GT/s 18 MB $1446
E7-2820 $1334
6 (12) E7-4807 1.86 GHz No 95 W 4× QPI 4.8 GT/s 4× DDR3-800 $890
E7-2803 1.73 GHz 105 W $774
Gulftown /
Westmere-EP

[11]
DP Server 6 (12) LGA
1366
Xeon X5690 3.46 GHz N/A Yes 130 W 2× QPI 6.4 GT/s 3× DDR3-1333 12 MB 2011-02-13 $1663
X5680 3.33 GHz 2010-03-16
X5675 3.06 GHz 95 W 2011-02-05 $1440
X5670 2.93 GHz 2010-03-16
X5660 2.8 GHz $1219
X5650 2.66 GHz $996
E5645 2.4 GHz 80 W 2× QPI 5.86 GT/s $958
L5640 2.26 GHz 60 W $996
L5638 2.0 GHz $958
4 (8) X5677 3.46 GHz 130 W 2× QPI 6.4 GT/s $1663
X5667 3.06 GHz 95 W $1440
E5640 2.66 GHz 80 W 2× QPI 5.86 GT/s 3× DDR3-1066 $774
E5630 2.53 GHz $551
E5620 2.4 GHz $387
L5630 2.13 GHz 40 W $551
L5618 1.86 GHz $530
4 (4) L5609 1.86 GHz No 2× QPI 4.8 GT/s $440
L5607 2.26 GHz 80 W 8 MB 2011-02-13 $276
L5606 2.13 GHz $219
L5603 1.6 GHz 4 MB $188
UP Server 6 (12) Xeon W3690 3.46 GHz N/A Yes 130 W 1× QPI 6.4 GT/s 3× DDR3-1333 12 MB 2011-02-13[12] $999
W3680 3.33 GHz 2010-03-16[13] $999
W3670 3.20 GHz 1× QPI 4.8 GT/s 3× DDR3-1066 2010-08-29 $885
Extreme /
Performance
Desktop
Core i7
Extreme
990X 3.46 GHz 1× QPI 6.4 GT/s 2011-02-13 $999
980X 3.33 GHz 2010-03-16
Core i7 980 1× QPI 4.8 GT/s 2011-06-26 $583
970 3.20 GHz 2010-07-17 $583
Clarkdale[14] UP Server 2 (4) LGA
1156
Xeon L3406 2.26 GHz N/A Yes 30 W DMI 2× DDR3-1066 4 MB 2010-03-16 $189
L3403 2.0 GHz 2010-10 $
Mainstream /
Value
Desktop
Core i5 680 3.6 GHz 733 MHz 73 W 2× DDR3-1333 2010-04-18 $294
670 3.46 GHz 2010-01-07 $284
661 3.33 GHz 900 MHz 87 W $196
660 733 MHz 73 W
655K 3.2 GHz 2010-05-30 $216
650 2010-01-07 $176
Core i3 560 3.33 GHz No 2010-08-29 $138
550 3.20 GHz 2010-05-30
540 3.06 GHz 2010-01-07 $133
530 2.93 GHz $113
2 (2) Pentium G6960 533 MHz 2× DDR3-1066 3 MB 2011-01-09 $89
G6951 2.8 GHz Q3 2010 OEM
G6950 2010-01-07 $87
Celeron G1101 2.26 GHz 2 MB OEM

Mobile Processors[edit]

CodenameMarketCores /
Threads
Processor
Branding & Model
CPU Clock rateGPU Clock rateTurboTDPMemoryL3
cache
InterfaceRelease
Date
Price
StandardTurbo
(1C/2C active cores )
Arrandale Mainstream /

Value Mobile

2 (4) Core i7 640M 2.8 GHz 3.46/3.2 GHz 766 MHz Yes 35 W 2× DDR3-1066 4 MB * DMI
PCIe 1 x16
* Socket:
µPGA-988 /
BGA-1288
2010-09-26 $346
620M 2.66 GHz 3.33/3.2 GHz 2010-01-07 $332
610E 2.53 GHz 3.2/2.93 GHz
660LM 2.26 GHz 3.06/2.8 GHz 566 MHz 25 W 2010-09-26 $346
640LM 2.13 GHz 2.93/2.66 GHz 2010-01-07 $332
620LM / 620LE 2.0 GHz 2.8/2.53 GHz $300
680UM 1.46 GHz 2.53/2.16 GHz 500 MHz 18 W 2× DDR3-800 2010-09-26 $317
660UM / 660UE 1.33 GHz 2.4/2.0 GHz 2010-05-25
640UM 1.2 GHz 2.26/1.86 GHz 2010-01-07 $305
620UM / 620UE 1.06 GHz 2.13/1.76 GHz $278
Core i5 580M 2.66 GHz 3.33/2.93 GHz 766 MHz 35 W 2× DDR3-1066 3 MB 2010-09-26 $266
560M 3.2/2.93 GHz $225
540M 2.53 GHz 3.06/2.8 GHz 2010-01-07 $257
520M / 520E 2.4 GHz 2.93/2.66 GHz $225
560UM 1.33 GHz 2.13/1.86 GHz 500 MHz 18 W 2× DDR3-800 2010-09-26 $250
540UM 1.2 GHz 2.0/1.73 GHz 2010-05-25
520UM 1.06 GHz 1.86/1.6 GHz 2010-01-07 $241
480M 2.66 GHz 2.93/2.93 GHz 766 MHz 35 W 2× DDR3-1066 2011-01-09 OEM
460M 2.53 GHz 2.8/2.8 GHz 2010-09-26
450M 2.4 GHz 2.66/2.66 GHz 2010-06-26
430M 2.26 GHz 2.53/2.53 GHz 2010-01-07
470UM 1.33 GHz 1.86/1.6 GHz 500 MHz 18 W 2× DDR3-800 2010-10-01
430UM 1.2 GHz 1.73/1.46 GHz 2010-05-25
Core i3 390 2.66 GHz n/a 667 MHz No 35 W 2× DDR3-1066 2011-01-09
380M 2.53 GHz 2010-09-26
370M 2.4 GHz 2010-06-20
350M 2.26 GHz 2010-01-07
330M / 330E 2.13 GHz
380UM 1.33 GHz 500 MHz 18 W 2× DDR3-800 2010-10-01
330UM 1.2 GHz 2010-05-25
2 (2) Pentium P6300 2.26 GHz 667 MHz 35 W 2× DDR3-1066 2011-01-09
P6200 2.13 GHz 2010-09-26
P6100 2.0 GHz
P6000 1.86 GHz 2010-06-20
U5600 1.33 GHz 500 MHz 18 W 2× DDR3-800 2011-01-09
U5400 1.2 GHz 2010-05-25
Celeron P4600 2.0 GHz 667 MHz 35 W 2× DDR3-1066 2 MB 2010-09-26 $86
P4500 / P4505 1.86 GHz 2010-03-28 OEM
U3600 1.2 GHz 500 MHz 18 W 2× DDR3-800 2011-01-09 $134
U3400 / U3405 1.06 GHz 2× DDR3-800 / 1066 2010-05-25 OEM

Roadmap[edit]

The successor to Nehalem and Westmere is Sandy Bridge.

See also[edit]

 

 

 

http://en.wikipedia.org/wiki/Sandy_Bridge_(microarchitecture)

Sandy Bridge

From Wikipedia, the free encyclopedia
  (Redirected from Sandy Bridge (microarchitecture))
 
Sandy Bridge
Max. CPU clock rate 1.60 GHz to 3.60 GHz
Product code 80623 (desktop)
L1 cache 64 KB per core
L2 cache 256 KB per core
L3 cache 1 MB to 8 MB shared
10 MB to 15 MB (Extreme)
3 MB to 20 MB (Xeon)
GPU HD Graphics 2000
650 MHz to 1250 MHz
Predecessor Nehalem (tock)
Westmere (tick)
Successor Ivy Bridge (tick)
Haswell (tock)
Socket(s)
Bottom view of a Sandy Bridge i7-2600k.

Sandy Bridge is the codename for a microarchitecture developed by Intel beginning in 2005 for central processing units in computers to replace the Nehalemmicroarchitecture. Intel demonstrated a Sandy Bridge processor in 2009, and released first products based on the architecture in January 2011 under the Corebrand.[1][2]

Sandy Bridge implementations targeted a 32 nanometer manufacturing process based on planar double-gate transistors.[3] Intel's subsequent product, codenamedIvy Bridge, uses a 22 nanometer process. The Ivy Bridge die shrink, known in the Intel Tick-Tock model as the "tick", is based on FinFET (non-planar, "3D") tri-gate transistors. Intel demonstrated the Ivy Bridge processors in 2011.[4]

 

 

Technology[edit]

Developed primarily by the Israel branch of Intel, the codename was originally "Gesher" (meaning "bridge" in Hebrew). The name was changed to avoid being associated with the defunct Gesher political party;[5] the decision was led by Ron Friedman, vice president of Intel managing the group at the time.[1] Intel demonstrated a Sandy Bridge processor with A1 stepping at 2 GHz during the Intel Developer Forum in September 2009.[6]

Upgraded features from Nehalem include:

  • 32 KB data + 32 KB instruction L1 cache (3 clocks) and 256 KB L2 cache (8 clocks) per core.
  • Shared L3 cache includes the processor graphics (LGA 1155).
  • 64-byte cache line size.
  • Two load/store operations per CPU cycle for each memory channel.
  • Decoded micro-operation cache (uop cache) and enlarged, optimized branch predictor.
  • Improved performance for transcendental mathematicsAES encryption (AES instruction set), and SHA-1 hashing.
  • 256-bit/cycle ring bus interconnect between cores, graphics, cache and System Agent Domain.
  • Advanced Vector Extensions (AVX) 256-bit instruction set with wider vectors, new extensible syntax and rich functionality.
  • Intel Quick Sync Video, hardware support for video encoding and decoding.
  • Up to 8 physical cores or 16 logical cores through Hyper-threading.
  • Integration of the GMCH (integrated graphics and memory controller) and processor into a single die inside the processor package. In contrast, Sandy Bridge's predecessor, Clarkdale, has two separate dies (one for GMCH, one for processor) within the processor package. This tighter integration reduces memory latency even more.
  • A 14- to 19-stage instruction pipeline, depending on the micro-operation cache hit or miss.[7]
Translation lookaside buffer sizes[8][9]
CachePage Size
NameLevel4 KB2 MB1 GB
DTLB 1st 64 32 4
ITLB 1st 128 8 / logical core none
STLB 2nd 512 none none
All translation lookaside buffers (TLBs) are 4-way associative.[10]

Models and steppings[edit]

All Sandy Bridge processors with one, two, or four cores report the same CPUID model 0206A7h[11] and are closely related. The stepping number can not be seen from the CPUID but only from the PCI configuration space. The later Sandy Bridge-E processors with up to eight cores and no graphics are using CPUIDs 0206D6h and 0206D7h.[12] Ivy Bridge CPUs all have CPUID 0306A9h to date, and are built in four different configurations differing in the number of cores, L3 cache and GPU execution units.

Die Code NameCPUIDSteppingDie sizeTransistorsCoresGPU EUsL3 CacheSockets
Sandy Bridge-HE-4 0206A7h D2 216 mm2 1.16 billion 4 12 8 MB LGA 1155Socket G2BGA-1224BGA-1023
Sandy Bridge-H-2 J1 149 mm2 624 million 2 4 MB LGA 1155Socket G2BGA-1023
Sandy Bridge-M-2 Q0 131 mm2 504 million 6 3 MB
Sandy Bridge-EP-8 0206D6h C1 435 mm2 2.27 billion 8 N/A 20 MB LGA 2011
0206D7h C2
Sandy Bridge-EP-4 0206D6h M0 294 mm2 1.27 billion 4 N/A 10 MB LGA 2011
0206D7h M1
Ivy Bridge-M-2 0306A9h P0 94 mm2[13]   2 6[14] 3 MB[15] LGA 1155,
Socket G2,
BGA-1224,
BGA-1023
Ivy Bridge-H-2 L1 118 mm2[13]   2 16 4 MB
Ivy Bridge-HE-4 E1 160 mm2[13] 1.4 billion[16] 4 16 8 MB
Ivy Bridge-HM-4 N0 133 mm2[13]   4 6 6 MB[15]

Performance[edit]

  • The average performance increase, according to IXBT Labs and Semi Accurate as well as many other benchmarking sites, at clock to clock is 11.3% compared to the Nehalem Generation, which includes Bloomfield, Clarkdale, and Lynnfield processors.[17]
  • Around twice the integrated graphics performance compared to Clarkdale's (12 EUs comparison).

List of Sandy Bridge processors[edit]

1Processors featuring Intel's HD 3000 graphics are set in bold. Other processors feature HD 2000 graphics or no graphics core (Graphics Clock rate indicated by N/A).

  • This list may not contain all the Sandy Bridge processors released by Intel. A more complete listing can be found on Intel's website.

Desktop platform[edit]

[18] [19] [20]

Target
segment
Processor
Branding & Model
Cores
(Threads)
CPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date (Y-M-D)
Price
(USD)
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory
Extreme /
High-End
Core i7
Extreme
3970X 6 (12) 3.5 GHz 4.0 GHz N/A 15 MB 150 W 2012-11-12 $999 LGA
2011
DMI 2.0
PCIe 2.0[21]
Up to quad
channel
DDR3-1600[22]
3960X 3.3 GHz 3.9 GHz 130 W 2011-11-14
Core i7 3930K 3.2 GHz 3.8 GHz 12 MB $583
3820 4 (8) 3.6 GHz 10 MB 2012-02-13[23] $294
Performance 2700K 3.5 GHz 3.9 GHz 850 MHz 1350 MHz 8 MB 95 W 2011-10-24 $332 LGA
1155
DMI 2.0
PCIe 2.0
Up to dual
channel
DDR3-1333
2600K 3.4 GHz 3.8 GHz 2011-01-09 $317
2600 $294
2600S 2.8 GHz 65 W $306
Core i5 2550K 4 (4) 3.4 GHz N/A 6 MB 95 W 2012-01-30 $225
2500K 3.3 GHz 3.7 GHz 850 MHz 1100 MHz 2011-01-09 $216
2500 $205
2500S 2.7 GHz 65 W $216
2500T 2.3 GHz 3.3 GHz 650 MHz 1250 MHz 45 W
2450P 3.2 GHz 3.5 GHz N/A 95 W 2012-01-30 $195
2400 3.1 GHz 3.4 GHz 850 MHz 1100 MHz 2011-01-09 $184
2405S 2.5 GHz 3.3 GHz 65 W 2011-05-22 $205
2400S 2011-01-09 $195
2380P 3.1 GHz 3.4 GHz N/A 95 W 2012-01-30 $177
2320 3.0 GHz 3.3 GHz 850 MHz 1100 MHz 2011-09-04
2310 2.9 GHz 3.2 GHz 2011-05-22
2300 2.8 GHz 3.1 GHz 2011-01-09
Mainstream 2390T 2 (4) 2.7 GHz 3.5 GHz 650 MHz 3 MB 35 W 2011-02-20 $195
Core i3 2120T 2.6 GHz N/A 2011-09-04 $127
2100T 2.5 GHz 2011-02-20
2115C 2.0 GHz N/A 25 W 2012-05 $241 BGA
1284
2130 3.4 GHz 850 MHz 1100 MHz 65 W 2011-09-04 $138 LGA
1155
2125 3.3 GHz $134
2120 2011-02-20 $138
2105 3.1 GHz 2011-05-22 $134
2102 Q2 2011 $127
2100 2011-02-20 $117
Pentium G870 2 (2) 2012-06-03 $86
G860 3.0 GHz 2011-09-04
G860T 2.6 GHz 650 MHz 35 W 2012-06-03 $75
G850 2.9 GHz 850 MHz 65 W 2011-05-24 $86
G840 2.8 GHz $75
G645 2.9 GHz 09-03-2012 $64 Up to dual
channel
DDR3-1066
G640 2.8 GHz 06-03-2012
G632 2.7 GHz Q3 2011  
G630 2011-09-04 $75
G622 2.6 GHz Q2 2011  
G620 2011-05-24 $64
G645T 2.5 GHz 650 MHz 35 W 09-03-2012
G640T 2.4 GHz 06-03-2012
G630T 2.3 GHz 2011-09-04 $70
G620T 2.2 GHz 2011-05-24
Celeron G555 2.7 GHz 850 MHz 1000 MHz 2 MB 65 W 2012-09-02 $52
G550 2.6 GHz 2012-06-03
G540 2.5 GHz 2011-09-04
G530 2.4 GHz $42
G550T 2.2 GHz 650 MHz 35 W 2012-09-02
G540T 2.1 GHz 2012-06-03
G530T 2.0 GHz 2011-09-04 $47
G470 1 (2) 1.5 MB 2013-06-09 $37 Up to dual
channel
DDR3-1333
G465 1.9 GHz 2012-09-02 Up to dual
channel
DDR3-1066
G460 1.8 GHz 2011-12-11
G440 1 (1) 1.6 GHz 1 MB 2011-09-04
Target
segment
Processor
Branding & Model
Cores
(Threads)
CPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date (Y-M-D)
Price
(USD)
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory

Suffixes to denote:

  • K – Unlocked (adjustable CPU ratio up to 57 bins)
  • P – Versions clocked slightly higher than similar models, but with onboard-graphics deactivated.
  • S – Performance-optimized lifestyle (low power with 65W TDP)
  • T – Power-optimized lifestyle (ultra low power with 35-45W TDP)
  • X – Extreme performance (adjustable CPU ratio with no ratio limit)

NOTE:3960X,3930K and 3820 are actually of Sandy Bridge-E edition.

Server platform[edit]

Target
Segment
SocketProcessor
Branding & Model
Cores
(Threads)
CPU Clock rateGraphics Clock rateL3
Cache
InterfaceSupported
Memory
TDPRelease
Date
Price
(USD)
StandardTurboNormalTurbo
4P Server LGA
2011
Xeon E5 4650 8 (16) 2.7 GHz 3.3 GHz N/A 20 MB 2× QPI
DMI 2.0
PCIe 3.0
4x DDR3-1600 130 W 2012-05-14 $3616
4650L 2.6 GHz 3.1 GHz 115 W
4640 2.4 GHz 2.8 GHz 95 W $2725
4620 2.2 GHz 2.6 GHz 16 MB 4x DDR3-1333 $1611
4617 6 (6) 2.9 GHz 3.4 GHz 15 MB 4x DDR3-1600 130 W
4610 6 (12) 2.4 GHz 2.9 GHz 4x DDR3-1333 95 W $1219
4607 2.2 GHz N/A 12 MB 4x DDR3-1066 $885
4603 4 (8) 2.0 GHz 10 MB $551
2P Server 2687W 8 (16) 3.1 GHz 3.8 GHz 20 MB 4x DDR3-1600 150 W 2012-03-06 $1885
2690 2.9 GHz 3.8 GHz 135 W $2057
2680 2.7 GHz 3.5 GHz 130 W $1723
2670 2.6 GHz 3.3 GHz 115 W $1552
2665 2.4 GHz 3.1 GHz $1440
2660 2.2 GHz 3.0 GHz 95 W $1329
2658 2.1 GHz 2.4 GHz $1186
2650 2.0 GHz 2.8 GHz $1107
2650L 1.8 GHz 2.3 GHz 70 W
2648L 1.8 GHz 2.1 GHz $1186
2667 6 (12) 2.9 GHz 3.5 GHz 15 MB 130 W $1552
2640 2.5 GHz 3.0 GHz 4x DDR3-1333 95 W $884
2630 2.3 GHz 2.8 GHz $612
2620 2.0 GHz 2.5 GHz $406
2630L 2.0 GHz 2.5 GHz 60 W $662
2643 4 (8) 3.3 GHz 3.5 GHz 10 MB 4x DDR3-1600 130 W $884
2609 4 (4) 2.4 GHz N/A 4x DDR3-1066 80 W $246
2603 1.8 GHz $202
2637 2 (4) 3.0 GHz 3.5 GHz 5 MB 4x DDR3-1600 $884
LGA
1356
2470 8 (16) 2.3 GHz 3.1 GHz 20 MB 1× QPI
DMI 2.0
PCIe 3.0
3x DDR3-1600 95 W 2012-05-14 $1440
2450 2.1 GHz 2.9 GHz $1106
2450L 1.8 GHz 2.3 GHz 70 W
2440 6 (12) 2.4 GHz 2.9 GHz 15 MB 3x DDR3-1333 95 W $834
2430 2.2 GHz 2.7 GHz $551
2420 1.9 GHz 2.4 GHz $388
2430L 2.0 GHz 2.5 GHz 60 W $662
2407 4 (4) 2.2 GHz N/A 10 MB 3x DDR3-1066 80 W $250
2403 1.8 GHz $192
1P Server LGA
2011
1660 6 (12) 3.3 GHz 3.9 GHz 15 MB 2× QPI
DMI 2.0
PCIe 3.0
Up to quad
channel
DDR3-1600
130 W 2012-03-06 $1080
1650 3.2 GHz 3.8 GHz 12 MB $583
1620 4 (8) 3.6 GHz 3.8 GHz 10 MB $294
1607 4 (4) 3.0 GHz N/A Up to quad
channel
DDR3-1066
$244
1603 2.8 GHz $198
LGA
1356
1428L 6 (12) 1.8 GHz N/A 15 MB 1× QPI
DMI 2.0
PCIe 3.0
3x DDR3-1333 60 W Q2 2012 $395
1410 4 (8) 2.8 GHz 3.2 GHz 10 MB 80 W 2012-05-14  
Pentium 1407 2 (2) N/A 5 MB 3x DDR3-1066  
1405 1.2 GHz 1.8 GHz 40 W August 2012 $143
1403 2.6 GHz N/A 80 W 2012-05-14  
LGA
1155
Xeon E3 1290 4 (8) 3.6 GHz 4.0 GHz 8 MB DMI 2.0
PCIe 2.0
Up to dual
channel
DDR3-1333
95 W 2011-05-29 $885
1280 3.5 GHz 3.9 GHz 2011-04-03 $612
1275 3.4 GHz 3.8 GHz 850 MHz 1350 MHz $339
1270 N/A 80 W $328
1260L 2.4 GHz 3.3 GHz 650 MHz 1250 MHz 45 W $294
1245 3.3 GHz 3.7 GHz 850 MHz 1350 MHz 95 W $262
1240 N/A 80 W $250
1235 3.2 GHz 3.6 GHz 850 MHz 1350 MHz 95 W $240
1230 N/A 80 W $215
1225 4 (4) 3.1 GHz 3.4 GHz 850 MHz 1350 MHz 6 MB 95 W $194
1220 N/A 8 MB 80 W $189
1220L 2 (4) 2.2 GHz 3 MB 20 W
BGA
1284
1125C 4 (8) 2.0 GHz N/A 8 MB Up to dual
channel
DDR3-1600
40 W May 2012 $444
1105C 1.0 GHz 6 MB 25 W $333
LGA
1155
Pentium 350 2 (4) 1.2 GHz 3 MB Up to dual
channel
DDR3-1333
15 W November 2011 $159
BGA
1284
Celeron 725C 1 (2) 1.3 GHz 1.5 MB 10 W May 2012 $74

Mobile platform[edit]

  • Core i5-2515E and Core i7-2715QE processors have support for ECC memory and PCI express port bifurcation.
  • All mobile processors, except Celeron and Pentium, use Intel's Graphics sub-system HD 3000 (12 EUs).
Target
Segment
Processor
Branding & Model
Cores /
Threads
CPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date
Price
(USD)
Motherboard
NormalTurbo
(1C/2C/4C)
NormalTurboInterfaceSocket
Extreme Core i7
Extreme
2960XM 4 (8) 2.7 GHz 3.7/3.6/3.4 GHz 650 MHz 1300 MHz 8 MB 55 W 2011-09-04 $1096 *DMI 2.0
*Memory: Up to
dual channel
DDR3-1600 MHz
*PCIe 2.0
Socket G2 /
BGA-1224 (in embedded products)[24]
2920XM 2.5 GHz 3.5/3.4/3.2 GHz 2011-01-05
Performance Core i7 2860QM 2.5 GHz 3.6/3.5/3.3 GHz 45 W 2011-09-04 $568
2820QM 2.3 GHz 3.4/3.3/3.1 GHz 2011-01-05
2760QM 2.4 GHz 3.5/3.4/3.2 GHz 6 MB 2011-09-04 $378
2720QM 2.2 GHz 3.3/3.2/3.0 GHz 2011-01-05
2715QE 2.1 GHz 3.0/2.9/2.7 GHz 1200 MHz
2710QE
2675QM 2.2 GHz 3.1/3.0/2.8 GHz 1200 MHz 2011-10-02 *DMI 2.0
*Memory: Up to
dual channel
DDR3-1333 MHz
*PCIe 2.0
2670QM 1100 MHz
2635QM 2.0 GHz 2.9/2.8/2.6 GHz 1200 MHz 2011-01-05
2630QM 1100 MHz
Mainstream 2640M 2 (4) 2.8 GHz 3.5/3.3 GHz 1300 MHz 4 MB 35 W 2011-09-04 $346 Socket G2 /
BGA-1023 (in embedded products)[24]
2620M 2.7 GHz 3.4/3.2 GHz 2011-02-20
2649M 2.3 GHz 3.2/2.9 GHz 500 MHz 1100 MHz 25 W
2629M 2.1 GHz 3.0/2.7 GHz $311
2655LE 2.2 GHz 2.9/2.7 GHz 650 MHz 1000 MHz $346
2677M 1.8 GHz 2.9/2.6 GHz 350 MHz 1200 MHz 17 W 2011-06-20 $317
2637M 1.7 GHz 2.8/2.5 GHz $289
2657M 1.6 GHz 2.7/2.4 GHz 1000 MHz 2011-02-20 $317
2617M 1.5 GHz 2.6/2.3 GHz 950 MHz $289
2610UE 2.4/2.1 GHz 850 MHz $317
Core i5 2557M 1.7 GHz 2.7/2.4 GHz 1200 MHz 3 MB 2011-06-20 $250
2537M 1.4 GHz 2.3/2.0 GHz 900 MHz 2011-02-20
2467M 1.6 GHz 2.3/2.0 GHz 1150 MHz 2011-06-19
2540M 2.6 GHz 3.3/3.1 GHz 650 MHz 1300 MHz 35 W 2011-06-20 $266
2520M 2.5 GHz 3.2/3.0 GHz $225
2515E 3.1/2.8 GHz 1100 MHz $266
2510E
2450M 1300 MHz 2012-01 $225
2435M 2.4 GHz 3.0/2.7 GHz 2011-10-02 OEM
2430M 1200 MHz $225
2410M 2.3 GHz 2.9/2.6 GHz 2011-06-20
Core i3 2370M 2.4 GHz N/A 1150 MHz 2012-01
2350M 2.3 GHz 2011-10-02
2348M 2013-01 OEM
2330E 2.2 GHz 1050 MHz 2011-06-19 $225
2330M 1100 MHz
2328M 2012-09
2312M 2.1 GHz Q2 2011 OEM
2310E 1050 MHz 2011-02-20
2310M 1100 MHz
2377M 1.5 GHz 350 MHz 1000 MHz 17 W Q3 2012 $225
2375M 2012-03
2367M 1.4 GHz 2011-10-02 $250
2365M 2012-09 $225
2357M 1.3 GHz 950 MHz 2011-06-19 OEM
2340UE 800 MHz $250
Pentium B915C 1.5 GHz N/A 15 W 2012-05 $138
997 2 (2) 1.6 GHz 350 MHz 1000 MHz 2 MB 17 W 2012-09-30 $134
987 1.5 GHz Q3 2012
977 1.4 GHz 2012-01
967 1.3 GHz 2011-10-02 OEM
957 1.2 GHz 800 MHz 2011-06-19 $134
B980 2.4 GHz 650 MHz 1150 MHz 35 W 2012-09 OEM
B970 2.3 GHz 2012-01 $125
B960 2.2 GHz 1100 MHz 2011-10-02 $134
B950 2.1 GHz 2011-06-19
B940 2.0 GHz
Celeron B840 1.9 GHz 1000 MHz 2011-09-04 $86
B830 1.8 GHz 1050 MHz 2012-09-30
B820[25] 1.7 GHz 2012-07-29
B815[26] 1.6 GHz 2012-01
B810E 1000 MHz 2011-06-19
B810 950 MHz 2011-03-13
B800 1.5 GHz 1000 MHz 2011-06-19 $80
887 350 MHz 17 W 09-30-2012 $86
877 1.4 GHz 2012-07-29
867 1.3 GHz January 2012 $134
857 1.2 GHz 2011-07-03
847 1.1 GHz 800 MHz 2011-06-19
847E
807 1 (2) 1.5 GHz 950 MHz 1.5 MB 2012-07-29 $70
725C 1.3 GHz N/A 10 W 2012-05 $74
827E 1 (1) 1.4 GHz 350 MHz 800 MHz 17 W 2011-07-03 $107
797 950 MHz 2012-01
787 1.3 GHz 2011-07-03
B730 1.8 GHz 650 MHz 1000 MHz 35 W 2012-07-29 $70
B720[27] 1.7 GHz 2012-01
B710 1.6 GHz 2011-06-19
807UE 1.0 GHz 350 MHz 800 MHz 1 MB 10 W 2011-11 $117

Suffixes to denote:

  • M – Mobile processors
    • XM – Unlocked
    • QM – Quad-core
  • E – Embedded mobile processors
    • QE – Quad-core
    • LE – Performance-optimized
    • UE – Power-optimized

Cougar Point chipset flaw[edit]

On January 31, 2011, Intel issued a recall on all 67-series motherboards due to a flaw in the Cougar Point Chipset.[28] A hardware problem, in which the chipset's SATA-II ports may fail over time, cause failure of connection to SATA-II devices, though data is not at risk.[29] Intel claims that this problem will affect only 5% of users over 3 years, however, heavier I/O workloads can exacerbate the problem.

Intel stopped production of flawed B2 stepping chipsets and began producing B3 stepping chipsets with the silicon fix. Shipping of these new chipsets started on 14 February 2011 and Intel estimated full recovery volume in April 2011.[30] Motherboard manufacturers (such as ASUS and Gigabyte Technology) and computer manufacturers (such as Dell and Hewlett-Packard) stopped selling products that involved the flawed chipset and offered support for affected customers. Options ranged from swapping for B3 motherboards to product refunds.[31][32]

Sandy Bridge processor sales were temporarily on hold, as one cannot use the CPU without a motherboard. However, processor release dates were not affected.[33] After two weeks, Intel continued shipping some chipsets, but manufacturers had to agree to a set of terms that will prevent customers from encountering the bug.[34]

Limitations[edit]

Overclocking[edit]

With Sandy Bridge, Intel has tied the speed of every bus (USB, SATA, PCI, PCI-E, CPU cores, Uncore, memory etc.) to a single internal clock generator issuing the basic 100 MHz Base Clock (BClk).[35]With CPUs being multiplier locked, the only way to overclock is to increase the BClk, which can be raised by only 5–7% without other hardware components failing. As a work around, Intel made available K/X-series processors, which feature unlocked multipliers; with a multiplier cap of 57 for Sandy Bridge.[36] For the Sandy Bridge E platform, there is alternative method known as the BClk ratio overclock.[37]

During IDF (Intel Developer Forum) 2010, Intel demonstrated an unknown Sandy Bridge CPU running stably overclocked at 4.9 GHz on air cooling.[38][39]

Chipset[edit]

Non-K edition CPUs can overclock up to four bins from its turbo multiplier. Refer here for chipset support.

vPro remote-control (Intel Insider)[edit]

Sandy and Ivy Bridge processors with vPro capability have security features that can remotely disable a PC or erase information from hard drives. This can be useful in the case of a lost or stolen PC. The commands can be received through 3G signals, Ethernet, or Internet connections. AES encryption acceleration will be available, which can be useful for video conferencing and VoIP applications.[40][41]

Software development kit[edit]

With the introduction of the Sandy Bridge microarchitecture, Intel also introduced the Intel Data Plane Development Kit (Intel DPDK) to help developers of communications applications take advantage of the platform in packet processing applications, and network processors.[42]

Roadmap[edit]

Intel demonstrated the Haswell architecture in September 2011, released in 2013 as the successor to Sandy Bridge and Ivy Bridge.[43]

See also[edit]

 

 

http://en.wikipedia.org/wiki/Ivy_Bridge_(microarchitecture)

Ivy Bridge (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Ivy Bridge
Ivy Bridge Codename Logo.jpg
CPUID code 0306A9h
Product code 80637 (desktop)
L1 cache 64 KB per core
L2 cache 256 KB per core
L3 cache MB to 8 MB shared
GPU HD Graphics 2500
650 MHz to 1150 MHz
HD Graphics 4000
350 MHz to 1300 MHz
HD Graphics P4000
650 MHz to 1250 MHz
Predecessor Sandy Bridge
Successor Haswell
Socket(s)

Ivy Bridge is the codename for a line of processors based on the 22 nm manufacturing process developed by Intel. The name is also applied more broadly to the22 nm die shrink of the Sandy Bridge microarchitecture based on FinFET ("3D") tri-gate transistors, which is also used in the Xeon and Core i7 Ivy Bridge-EX(Ivytown), Ivy Bridge-EP and Ivy Bridge-E microprocessors released in 2013.

Ivy Bridge processors are backwards compatible with the Sandy Bridge platform, but such systems might require a firmware update (vendor specific).[1] In 2011, Intel released the 7-series Panther Point chipsets with integrated USB 3.0 to complement Ivy Bridge.[2]

Volume production of Ivy Bridge chips began in the third quarter of 2011.[3] Quad-core and dual-core-mobile models launched on April 29, 2012 and May 31, 2012 respectively.[4] Core i3 desktop processors, as well as the first 22 nm Pentium, were announced and available the first week of September, 2012.[5]

 

 

Overview[edit]

The Ivy Bridge CPU microarchitecture is a shrink from Sandy Bridge and remains largely unchanged.

Notable improvements include:[6][7]

  • 22 nm Tri-gate transistor ("3-D") technology (up to 50% less power consumption at the same performance level as 2-D planar transistors).[8]
  • A new random number generator and the RdRand instruction,[9] codenamed Bull Mountain.[10]

Ivy Bridge features and performance[edit]

The mobile and desktop Ivy Bridge chips also include significant changes over Sandy Bridge:

Translation lookaside buffersizes[19][20]
CachePage Size
NameLevel4 KB2 MB1 GB
DTLB 1st 64 32 4
ITLB 1st 128 8 / logical core none
STLB 2nd 512 none none

Benchmark comparisons[edit]

Compared to Sandy Bridge:

  • 3% to 5% increase in CPU performance when compared clock for clock[21][22]
  • 25% to 68% increase in integrated GPU performance.[23]

Thermal performance and heat issues[edit]

Ivy Bridge's temperatures are reportedly 10°C higher compared to Sandy Bridge when overclocked, even at default voltage setting.[24] Impress PC Watch, a Japanese website, performed experiments that confirmed earlier speculations that this is because Intel used a poor quality (and perhaps lower cost) thermal interface material (thermal paste, or "TIM") between the chip and the heat spreader, instead of thefluxless solder of previous generations.[25][26][27] The mobile Ivy Bridge processors are not affected by this issue because they do not use a heat spreader between the chip and cooling system.

Enthusiast reports describe the TIM used by Intel as low-quality,[27] and not up to par for a "premium" CPU, with some speculation that this is by design to encourage sales of prior processors.[25] Further analyses caution that the processor can be damaged or void its warranty if home users attempt to remedy the matter.[25][28] The TIM has much lower thermal conductivity, causing heat to trap on the die.[24]Experiments replacing this with a higher quality TIM or other heat removal methods showed a substantial temperature drop, and improvements to the voltages and clocking sustainable by Ivy Bridge chips.[25][29]

Intel claims that the smaller die of Ivy Bridge and the related increase in thermal density is expected to result in higher temperatures when the CPU is overclocked; Intel also stated that this is as expected and will likely not improve in future revisions.[30]

Models and steppings[edit]

All Ivy Bridge processors with one, two, or four cores report the same CPUID model 0x000306A9, and are built in four different configurations differing in the number of cores, L3 cache and GPU execution units.

Die Code NameCPUIDSteppingDie SizeDie DimensionsTransistorsCoresGPU EUsL3 CacheSockets
Ivy Bridge-M-2 0x000306A9 P0 94 mm2[31] 7.656 x 12.223 mm   2 6[32] 3 MB[33] LGA 1155,
Socket G2,
BGA-1224,
BGA-1023
Ivy Bridge-H-2 L1 118 mm2[31] 8.141 x 14.505 mm   2 16 4 MB
Ivy Bridge-HE-4 E1 160 mm2[31] 8.141 x 19.361 mm 1.4 billion[34] 4 16 8 MB
Ivy Bridge-HM-4 N0 133 mm2[31] 7.656 x 17.349 mm   4 6 6 MB[33]

Ivy Bridge-E features[edit]

Ivy Bridge-E
CPUID code 0306Fxh
Product code 80633
L1 cache 32 KB per core
L2 cache 256 KB per core
L3 cache 15 MB shared
Predecessor Sandy Bridge-E
Successor Haswell-E
Socket(s)
Ivy Bridge-EN
CPUID code 0306Fxh
Product code 80634
L1 cache 32 KB per core
L2 cache 256 KB per core
L3 cache 10 MB to 25 MB shared
Predecessor Sandy Bridge-EN
Successor Haswell-EN
Socket(s)
Ivy Bridge-EP
CPUID code 0306Fxh
Product code 80635
L1 cache 32 KB per core
L2 cache 256 KB per core
L3 cache 10 MB to 30 MB shared
Predecessor Sandy Bridge-EP
Successor Haswell-EP
Socket(s)
Ivy Bridge-EX
CPUID code 0306Fxh
Product code 80636
L1 cache 32 KB per core
L2 cache 256 KB per core
L3 cache 12 MB to 37.5 MB shared
Predecessor Westmere-EX
Successor ?
Socket(s)

Ivy Bridge-E is the follow-up to Sandy Bridge-E, using the same CPU core as the Ivy Bridge processor, but in an LGA 2011 or LGA 1356 package for workstations and servers.

  • New RAS features for Ivybridge-EX
  • Dual-Memory Controller for Ivybridge-EP
  • No integrated GPU
  • Up to 15 CPU cores
  • Up to 37.5 MB L3 cache.[35]
  • Thermal design power between 60 W and 155 W
  • Support for up to 8 DIMMS of DDR3-1866 memory per socket

Models and Steppings[edit]

The Ivy Bridge-E family is made in three different versions, by number of cores, and for three market segments: the basic Ivy Bridge-E is a single-socket processor sold as Core i7-49xx and is only available in the six-core S1 stepping, with some versions limited to four active cores.

Ivy Bridge-EN (Xeon E5-14xx v2 and Xeon E5-24xx v2) is the model for single- and dual-socket servers using LGA 1356 with up to 10 cores, while Ivy Bridge-EP (nd Xeon E5-16xx v2, Xeon E5-26xx v2 and Xeon E5-46xx v2) scales up to four LGA 2011 sockets and up to 12 cores per chip and Ivybridge-EX will have up to 15 cores and scale to 8 sockets.

Die Code NameCPUIDSteppingDie sizeTransistorsCoresL3 CacheSocket
Ivy Bridge-E-6 0x0306Fx S1 256.5 mm² 1.86 billion 6 15 MB LGA 2011
Ivy Bridge-EN-6 LGA 1356
Ivy Bridge-EP-6 LGA 2011
Ivy Bridge-EN-10 M1 346.5 mm² 2.86 billion 10 25 MB LGA 1356
Ivy Bridge-EP-10 LGA 2011
Ivy Bridge-EX-15 C1 541 mm² 4.3 billion 15 37.5 MB LGA 2011
 

List of Ivy Bridge and Ivy Bridge-E processors[edit]

Processors featuring Intel's HD 4000 graphics (or HD P4000 for Xeon) are set in bold. Other processors feature HD 2500 graphics unless indicated by N/A.

Desktop processors[edit]

List of announced desktop processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
CPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date
Release
price
(USD)
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory
Extreme /
High-End
6 (12) Core i7
Extreme
4960X 3.6 GHz 4.0 GHz N/A 15 MB 130 W 2013-09-10 $999[36] LGA
2011
DMI 2.0
PCIe 3.0*
Up to quad
channel
DDR3-1866
Core i7 4930K 3.4 GHz 3.9 GHz 12 MB $583[36]
4 (8) 4820K 3.7 GHz 10 MB $323[36]
Performance 3770K 3.5 GHz 3.9 GHz 650 MHz 1150 MHz 8 MB 77 W 2012-04-23 $332 LGA
1155
DMI 2.0
PCIe 3.0
Up to dual
channel
DDR3-1600[37]
3770 3.4 GHz $294
3770S 3.1 GHz 65 W
3770T 2.5 GHz 3.7 GHz 45 W
Mainstream 4 (4) Core i5 3570K 3.4 GHz 3.8 GHz 6 MB 77 W $225
3570 2012-05-31[38] $205
3570S 3.1 GHz 65 W
3570T 2.3 GHz 3.3 GHz 45 W
3550 3.3 GHz 3.7 GHz 77 W 2012-04-23
3550S 3.0 GHz 65 W
3475S 2.9 GHz 3.6 GHz 1100 MHz 2012-05-31[38] $201
3470 3.2 GHz 77 W $184
3470S 2.9 GHz 65 W
2 (4) 3470T 3 MB 35 W
4 (4) 3450 3.1 GHz 3.5 GHz 6 MB 77 W 2012-04-23
3450S 2.8 GHz 65 W
3350P 3.1 GHz 3.3 GHz N/A 69 W 2012-09-03 $177
3340 650 MHz 1050 MHz 77 W 2013-09-01 $182
3340S 2.8 GHz 65 W
3335S 2.7 GHz 3.2 GHz 2012-09-03 $194
3330S $177
3330 3.0 GHz 77 W $182
2 (4) Core i3 3250 3.5 GHz N/A 3 MB 55 W 2013-06-09 $138 DMI 2.0
PCIe 2.0
3245 3.4 GHz $134
3240 2012-09-03 $138
3225 3.3 GHz $134
3220 $117
3210 3.2 GHz 2013-01-20
3250T 3.0 GHz 35 W 2013-06-09 $138
3240T 2.9 GHz 2012-09-03
3220T 2.8 GHz $117
2 (2) Pentium G2140 3.3 GHz 55 W 2013-06-09 $86
G2130 3.2 GHz 2013-01-20
G2120 3.1 GHz 2012-09-03
G2120T 2.7 GHz 35 W 2013-06-09 $75
G2100T 2.6 GHz 2012-09-03
G2030 3.0 GHz 55 W 2013-06-09 $64 Dual channel DDR3-1333
G2020 2.9 GHz 2013-01-20
G2010 2.8 GHz
G2030T 2.6 GHz 35 W 2013-06-09
G2020T 2.5 GHz 2013-01-20
2 (2) Celeron G1630 2.8 GHz 2 MB 55 W 2013-09-01 $52
G1620 2.7 GHz 2013-01-20
G1610 2.6 GHz $42
G1620T 2.4 GHz 35 W 2013-09-01
G1610T 2.3 GHz 2013-01-20

 Requires a compatible Motherboard

Suffixes to denote:

  • K - Unlocked (adjustable CPU multiplier up to 63 bins)
  • S - Performance-optimized lifestyle (low power with 65 W TDP)
  • T - Power-optimized lifestyle (ultra low power with 35-45 W TDP)
  • P - No on-die video chipset
  • X – Extreme performance (adjustable CPU ratio with no ratio limit)

Server processors[edit]

Additional high-end server processors based on the Ivy Bridge architecture, code named Ivytown, were announced September 10, 2013 at the Intel Developer Forum, after the usual one year interval between consumer and server product releases.[39][40][41] The Ivy Bridge-EP processor line announced in September 2013 has up to 12 cores and 30 MB third level cache, with rumors of Ivy Bridge-EX up to 15 cores and an increased third level cache of up to 37.5 MB,[42][43] although an early leaked lineup of Ivy Bridge-E included processors with a maximum of 6 cores.[44] Both Core-i7 and Xeon versions are produced: the Xeon versions marketed as Xeon E5-2600 V2 act as drop-in replacements for the existing Sandy Bridge-EN and Sandy Bridge-EP based Xeon E5, and Core-i7 versions designated i7-4820K, i7-4930K, i7-4960X were released on September 10, 2013 remained compatible with X79 and LGA2011 hardware.[43][45]

A new Ivy Bridge-EX line marketed as Xeon E7 V2 had no corresponding predecessor using the Sandy Bridge microarchitecture but instead followed the older Westmere-EX processors.

Target
Segment
Cores
(Threads)
Processor
Branding & Model
CPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date
Price
(USD)
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory
2P Server 12 (24) Xeon E5 2697v2 2.7 GHz 3.5 GHz N/A 30 MB 130 W 2013-09-10 $2614 LGA
2011
2× QPI
DMI 2.0
PCIe 3.0
Up to quad
channel
DDR3-1866
2695v2 2.4 GHz 3.2 GHz 115 W $2336
2692v2 2.2 GHz 3.0 GHz 2013-06 OEM (Tianhe-2)
2651v2 1.8 GHz ? ? 2013-09-10  
10 (20) 2690v2 3.0 GHz 3.6 GHz 25 MB 130 W $2057
2680v2 2.8 GHz 3.6 GHz 115 W $1723
2670v2 2.5 GHz 3.3 GHz $1552
2660v2 2.2 GHz 3.0 GHz 95 W $1389
2658v2 2.4 GHz $1440
2650Lv2 1.7 GHz 2.1 GHz 70 W $1219 Up to quad
channel
DDR3-1600
2648Lv2 1.9 GHz 2.5 GHz $1218 Up to quad
channel
DDR3-1866
8 (16) 2687Wv2 3.4 GHz 4.0 GHz 150 W $2108
2667v2 3.3 GHz 130 W $2057
2650v2 2.6 GHz 3.4 GHz 20 MB 95 W $1166
2640v2 2.0 GHz 2.5 GHz $885 Up to quad
channel
DDR3-1600
2628Lv2 1.9 GHz 2.4 GHz 70 W $1000
6 (12) 2643v2 3.5 GHz 3.8 GHz 25 MB 130 W $1552 Up to quad
channel
DDR3-1866
2630v2 2.6 GHz 3.1 GHz 15 MB 80 W $612 Up to quad
channel
DDR3-1600
2630Lv2 2.4 GHz 2.8 GHz 60 W
2620v2 2.1 GHz 2.6 GHz 80 W $406
2618Lv2 2.0 GHz N/A 50 W $520 Up to quad
channel
DDR3-1333
4 (8) 2637v2 3.5 GHz 3.8 GHz 130 W $996 Up to quad
channel
DDR3-1866
4 (4) 2609v2 2.5 GHz N/A 10 MB 80 W $294 Up to quad
channel
DDR3-1333
2603v2 1.8 GHz $202
1P Server 8 (16) 1680v2 3.0 GHz 3.9 GHz 25 MB 130 W $1723 0× QPI
DMI 2.0
PCIe 3.0
Up to quad
channel
DDR3-1866
6 (12) 1660v2 3.7 GHz 4.0 GHz 15 MB $1080
1650v2 3.5 GHz 3.9 GHz 12 MB $583
4 (8) 1620v2 3.7 GHz 10 MB $294
4 (4) 1607v2 3.0 GHz N/A $244 Up to quad
channel
DDR3-1600
4 (8) Xeon E3 1290v2 3.7 GHz 4.1 GHz 8 MB 87 W 2012-05-14 $885 LGA
1155
DMI 2.0
PCIe 3.0?
Up to dual
channel
DDR3-1600
1280v2 3.6 GHz 4.0 GHz 69 W $623
1275v2 3.5 GHz 3.9 GHz 650 MHz 1.25 GHz 77 W $350
1270v2 N/A 69 W $339
1265Lv2 2.5 GHz 3.5 GHz 650 MHz 1.15 GHz 45 W $305
1245v2 3.4 GHz 3.8 GHz 650 MHz 1.25 GHz 77 W $273
1240v2 N/A 69 W $261
1230v2 3.3 GHz 3.7 GHz $230
4 (4) 1225v2 3.2 GHz 3.6 GHz 650 MHz 1.25 GHz 77 W $224
1220v2 3.1 GHz 3.5 GHz N/A 69 W $203
2 (4) 1220Lv2 2.3 GHz 3 MB 17 W $189

? Requires a compatible Motherboard

Mobile processors[edit]

Target
segment
Cores
(Threads)
Processor
Branding & Model
Programmable TDPCPU TurboGraphics Clock rateL3
Cache
Release
Date
Price
(USD)
SDP[46]cTDP downNominal TDPcTDP up1-coreNormalTurbo
Performance 4 (8) Core i7 3940XM N/A 45 W / ? GHz 55 W / 3.0 GHz 65 W / ? GHz 3.9 GHz 650 MHz 1350 MHz 8 MB 2012-09-30 $1096
3920XM 45 W / ? GHz 55 W / 2.9 GHz 65 W / ? GHz 3.8 GHz 1300 MHz 2012-04-23
3840QM N/A 45 W / 2.8 GHz N/A 2012-09-30 $568
3820QM 45 W / 2.7 GHz 3.7 GHz 1250 MHz 2012-04-23
3740QM 1300 MHz 6 MB 2012-09-30 $378
3720QM 45 W / 2.6 GHz 3.6 GHz 1250 MHz 2012-04-23
3635QM 45 W / 2.4 GHz 3.4 GHz 1200 MHz 2012-09-30 N/A
3632QM 35 W / 2.2 GHz 3.2 GHz 1150 MHz $378
3630QM 45 W / 2.4 GHz 3.4 GHz
3615QM 45 W / 2.3 GHz 3.3 GHz 1200 MHz 2012-04-23
3612QM 35 W / 2.1 GHz 3.1 GHz 1100 MHz
3610QM 45 W / 2.3 GHz 3.3 GHz
Mainstream 2 (4) 3689Y 7 W / ? GHz 10 W / 13 W / 1.5 GHz 2.6 GHz 350 MHz 850 MHz 4 MB 2013-01-07 $362
3687U N/A 14 W / ? GHz 17 W / 2.1 GHz 25 W / 3.1 GHz 3.3 GHz 1200 MHz 2013-01-20 $346
3667U 14 W / ? GHz 17 W / 2.0 GHz 25 W / 3.0 GHz 3.2 GHz 1150 MHz 2012-06-03
3537U 14 W / ? GHz 25 W / 2.9 GHz 3.1 GHz 1200 MHz 2013-01-20
3555LE N/A 25 W / 2.5 GHz N/A 3.2 GHz 550 MHz 1000 MHz 2012-06-03 $360
3540M 35 W / 3.0 GHz 3.7 GHz 650 MHz 1300 MHz 2013-01-20 $346
3525M 35 W / 2.9 GHz 3.6 GHz 1350 MHz Q3 2012  
3520M 1250 MHz 2012-06-03 $346
3517U 14 W / ? GHz 17 W / 1.9 GHz 25 W / 2.8 GHz 3.0 GHz 350 MHz 1150 MHz
3517UE 14 W / ? GHz 17 W / 1.7 GHz 25 W / 2.6 GHz 2.8 GHz 1000 MHz $330
Core i5 3610ME N/A 35 W / 2.7 GHz N/A 3.3 GHz 650 MHz 950 MHz 3 MB $276
3439Y 7 W / ? GHz 10 W / ? GHz 13 W / 1.5 GHz 2.3 GHz 350 MHz 850 MHz 2013-01-07 $250
3437U' N/A 14 W / ? GHz 17 W / 1.9 GHz 25 W / 2.4 GHz 2.9 GHz 650 MHz 1200 MHz 2013-01-20 $225
3427U 14 W / ? GHz 17 W / 1.8 GHz 25 W / 2.3 GHz 2.8 GHz 350 MHz 1150 MHz 2012-06-03
3380M N/A 35 W / 2.9 GHz N/A 3.6 GHz 650 MHz 1250 MHz 2013-01-20 $266
3365M 35 W / 2.8 GHz 3.5 GHz 1350 MHz Q3 2012  
3360M 1200 MHz 2012-06-03 $266
3340M 35 W / 2.7 GHz 3.4 GHz 1250 MHz 2013-01-20 $225
3339Y 7 W / ? GHz 10 W / ? GHz 13 W / 1.5 GHz 2.0 GHz 350 MHz 850 MHz 2013-01-07 $250
3337U N/A 14 W / ? GHz 17 W / 1.8 GHz 2.7 GHz 350 MHz 1100 MHz 2013-01-20 $225
3320M N/A 35 W / 2.6 GHz 3.3 GHz 650 MHz 1200 MHz 2012-06-03
3317U 14 W / ? GHz 17 W / 1.7 GHz 2.6 GHz 350 MHz 1050 MHz
3230M N/A 35 W / 2.6 GHz 3.2 GHz 650 MHz 1100 MHz 2013-01-20
3210M 35 W / 2.5 GHz 3.1 GHz 2012-06-03
Core i3 3229Y 7 W / ? GHz 10 W / ? GHz 13 W / 1.4 GHz N/A 350 MHz 850 MHz 2013-01-07 $250
3227U N/A 14 W / ? GHz 17 W / 1.9 GHz 1100 MHz 2013-01-20 $225
3217U 14 W / ? GHz 17 W / 1.8 GHz 1050 MHz 2012-06-24
3217UE 14 W / ? GHz 17 W / 1.6 GHz 900 MHz July 2013 $261
3130M N/A 35 W / 2.6 GHz 650 MHz 1100 MHz 2013-01-20 $225
3120M 35 W / 2.5 GHz 2012-09-30
3120ME 35 W / 2.4 GHz 900 MHz July 2013
3110M 1000 MHz 2012-06-24
2 (2) Pentium 2030M 35 W / 2.5 GHz 1100 MHz 2 MB 2013-01-20 $134
2020M 35 W / 2.4 GHz 2012-09-30
2127U 17 W / 1.9 GHz 350 MHz 2013-06-09
2117U 17 W / 1.8 GHz 1000 MHz 2012-09-30
2129Y 7 W 10 W / 1.1 GHz 850 MHz 2013-01-07 $150
Celeron 1019Y 7 W 10 W / 1.0 GHz 800 MHz 2013-04 $153
1020E N/A 35 W / 2.2 GHz 650 MHz 1000 MHz 2013-01-20 $86
1020M 35 W / 2.1 GHz
1005M 35 W / 1.9 GHz 2013-06-09
1000M 35 W / 1.8 GHz 2013-01-20
1037U 17 W / 1.8 GHz 350 MHz
1017U 17 W / 1.6 GHz 2013-06-09
1007U 17 W / 1.5 GHz 2013-01-20
1047UE 17 W / 1.4 GHz 900 MHz $134
1 (1) 927UE 17 W / 1.5 GHz $107
  • M - Mobile processor
  • Q - Quad-core
  • U - Ultra-low power
  • X - 'Extreme'
  • Y - Extreme-ultra low power

Roadmap[edit]

Intel demonstrated the Haswell architecture in September 2011, which began release in 2013 as the successor to Sandy Bridge and Ivy Bridge.[47]

See also[edit]

 

 

http://en.wikipedia.org/wiki/Haswell_(microarchitecture)

Haswell (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Haswell
L1 cache 64 KB per core
L2 cache 256 KB per core
L3 cache MB to 8 MB shared
GPU HD Graphics 4200, 4400,
4600, 5000, Iris 5100
or Iris Pro 5200
200 MHz to 1.3 GHz
Predecessor Sandy Bridge (tock)
Ivy Bridge (tick)
Successor Broadwell (tick)
Skylake (tock)
Socket(s)

Haswell is the codename for a processor microarchitecture developed by Intel as the successor to the Ivy Bridge architecture.[1] It uses the 22 nm process.[2] Intel officially announced CPUs with this microarchitecture on June 4, 2013 at Computex Taipei 2013.[3] With Haswell, Intel introduced a low-power processor designed for convertible or 'hybrid' Ultrabooks, having the Y suffix. Intel demonstrated a working Haswell chip at the 2011 Intel Developer Forum.[4]

Haswell CPUs are used in conjunction with the Intel 8 Series chipsets.

 

 

Design[edit]

A Haswell wafer with a pin for scale.

The Haswell architecture is specifically designed[5] to optimize the power savings and performance benefits from the move to FinFET (non-planar, "3D") transistors on the improved 22 nm process node.[6]

Haswell has been launched in three major forms:[7]

  1. Desktop version (LGA1150 socket): Haswell-DT
  2. Mobile/Laptop version (PGA socket): Haswell-MB
  3. BGA version:
    • 47 W and 57 W TDP classes: Haswell-H (For "All-in-one" systems, Mini-ITX form factor motherboards, and other small footprint formats.)
    • 13.5 W and 15 W TDP classes (MCP): Haswell-ULT (For Intel's UltraBook platform.)
    • 10 W TDP class (SoC): Haswell-ULX (For tablets and certain UltraBook-class implementations.)

Notes[edit]

  • ULT = Ultra Low TDP; ULX = Ultra Low eXtreme TDP.
  • Only certain quad-core variants and BGA R-series SKUs will receive GT3 (Intel HD 5000, Intel Iris 5100), or GT3e (Intel Iris Pro 5200) integrated graphics. All other models will get GT2 (Intel HD 4X00) integrated graphics.[8] See also Intel HD Graphics for more detailed specifications.
  • Due to low power requirements of tablet and UltraBook platforms, Haswell-ULT and Haswell-ULX will only be available in dual-core. All other versions will be available in dual- or quad-core variants.

Performance[edit]

Compared to Ivy Bridge:

  • Approximately 8% better vector processing performance.[9]
  • Up to 6% faster single-threaded performance.
  • 6% faster multi-threaded performance.
  • Desktop variants of Haswell draw between 8% and 23% more power under load than Ivy Bridge.[9][10][11]
  • A 6% increase in sequential CPU performance (eight execution ports per core versus six).[9]
  • Up to 20% performance increase over the integrated HD4000 GPU (Haswell HD4600 vs Ivy Bridge's built-in Intel HD4000).[9]
  • Total performance improvement on average is about 3%.[9]
  • Around 15 °C hotter than Ivy Bridge and unable to break 4.2 GHz easily.[12][13][14][15][16][17]

Technology[edit]

Features carried over from Ivy Bridge[edit]

New features[edit]

  • Wider Core: fourth ALU, third AGU, second branch prediction unit, deeper buffers, higher cache bandwidth, improved front-end and memory controller
  • Haswell New Instructions[22] (HNI, includes Advanced Vector Extensions 2 (AVX2), gatherBMI1+BMI2, LZCNT and FMA3 support).[23]
  • The instruction decode queue, which holds instructions after they have been decoded, is no longer statically partitioned between the two threads that each core can service.[18]
  • New sockets – LGA 1150 for desktops and rPGA947 & BGA1364 for the mobile market.[24]
  • Z97 (performance) and H97 (mainstream) chipsets for the Haswell Refresh and Broadwell, in Q2 2014.[25]
  • New socket – LGA 2011-3 with X99 chipset for the Enthusiast-Class Desktop Platform Haswell-E.[26]
  • Intel Transactional Synchronization Extensions (TSX), on selected models.[27]
  • Graphics support in hardware for Direct3D 11.1 and OpenGL 4.0.[28]
  • DDR4 for the enterprise/server variant (Haswell-EX).[29]
  • DDR4 for the Enthusiast-Class Desktop Platform Haswell-E.[30]
  • Variable Base clock (BClk)[31] like LGA 2011.[32]
  • There are four versions of the integrated GPU: GT1, GT2, GT3 and GT3e, where GT3 version has 40 execution units (EUs). Haswell's predecessor, Ivy Bridge, has a maximum of 16 EUs. GT3e version with 40 EUs and on-package 128 MB of embedded DRAM (eDRAM), called Crystal Well, is available only in mobile H-SKUs and desktop (BGA-only) R-SKUs. Effectively, this eDRAM is a Level 4 cache — shared dynamically between the on-die GPU and CPU, and serving as a victim cache to the CPU's L3 cache.[33][34][35][36][37]
  • Support for Thunderbolt technology and Thunderbolt 2.0.[38]
  • Fully integrated voltage regulator (FIVR), thereby moving some of the components from motherboard onto the CPU.[39][40][41] FIVR is implemented as a separate 13x8 mm on-package die, manufactured in90 nm process.[42]
  • New advanced power-saving system.
  • 37, 47, 57 W thermal design power (TDP) mobile processors.[19]
  • 35, 45, 65, 84, 95 and 130–140 W (high-end, Haswell-E) TDP desktop processors.[19]
  • 15 W TDP processors for the Ultrabook platform (multi-chip package like Westmere)[43] leading to reduced heat which results in thinner as well as lighter Ultrabooks, but performance level will be lower than the 17 W version.[44]
  • Shrink of the Platform Controller Hub (PCH), from 65 nm to 32 nm.[45]
Translation lookaside buffersizes[46][47]
CachePage Size
NameLevel4 KB2 MB1 GB
DTLB 1st 64 32 4
ITLB 1st 128 8 / logical core none
STLB 2nd 1024 none

Expected Server features[edit]

  • Release not before end of 2014.
  • Haswell-EP having up to 14–15 cores, and Haswell-EX with up to 18–20 cores.
  • A new cache design.
  • Up to 35 MB total unified cache (Last Level Cache (LLC)) for Haswell-EP[48][49] and up to 40 MB for Haswell-EX.
  • It is possible that Socket R3 will replace LGA 2011 for server Haswells.[50][51]

List of Haswell processors[edit]

Desktop processors[edit]

  • Intel HD Graphics in following variants:
    • R-series desktop processors feature Intel Iris Pro 5200 graphics (GT3e).[53]
    • All other currently known i3, i5 and i7 desktop processors include Intel HD 4600 graphics (GT2).[54]
    • The exceptions are processors 4130 and 4130T, which include HD 4400 graphics (GT2).
    • Pentium processors contain Intel HD Graphics (GT1).
  • Haswell-based desktop Celerons are planned for the first quarter of 2014.[55]

List of announced desktop processors is as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
CPU Clock rateGraphics Clock rateL3
Cache
GPU
eDRAM
TDPRelease
Date
Release
price
(USD)
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory
Performance 4 (8) Core i7 4771 3.5 GHz 3.9 GHz 350 MHz[56] 1.2 GHz 8 MB N/A 84 W September 1, 2013 $320 LGA
1150
DMI 2.0
PCIe 3.0
Up to dual
channel
DDR3-1600[57]
4770K 1.25 GHz June 2, 2013[58] $339
4770 3.4 GHz 1.2 GHz $303
4770S 3.1 GHz 65 W
4770R 3.2 GHz 200 MHz 1.3 GHz 6 MB 128 MB $ BGA
4770T 2.5 GHz 3.7 GHz 350 MHz[56] 1.2 GHz 8 MB N/A 45 W $303 LGA
1150
4765T 2.0 GHz 3.0 GHz 35 W
Mainstream 4 (4) Core i5 4670K 3.4 GHz 3.8 GHz 6 MB 84 W $242
4670 $213
4670S 3.1 GHz 65 W
4670R 3.0 GHz 3.7 GHz 200 MHz 1.3 GHz 4 MB 128 MB $ BGA
4670T 2.3 GHz 3.3 GHz 350 MHz[56] 1.2 GHz 6 MB N/A 45 W $213 LGA
1150
4570 3.2 GHz 3.6 GHz 1.15 GHz 84 W $192
4570S 2.9 GHz 65 W
4570R 2.7 GHz 3.2 GHz 200 MHz 4 MB 128 MB $ BGA
2 (4) 4570T 2.9 GHz 3.6 GHz N/A 35 W $192 LGA
1150
4 (4) 4440 3.1 GHz 3.3 GHz 350 MHz[56] 1.1 GHz 6 MB 84 W September 1, 2013 $187
4440S 2.8 GHz 65 W
4430 3.0 GHz 3.2 GHz 84 W June 2, 2013[58] $182
4430S 2.7 GHz 65 W
2 (4) Core i3 4340 3.6 GHz N/A 1.15 GHz 4 MB 54 W September 1, 2013 $157
4330 3.5 GHz $147
4330T 3.0 GHz 200 MHz 35 W $138
4330TE 2.4 GHz 350 MHz 1 GHz $122
4130 3.4 GHz 1.15 GHz 3 MB 54 W $129
4130T 2.9 GHz 200 MHz 35 W $131
2 (2) Pentium G3430 3.3 GHz 350 MHz 1.1 GHz 54 W $93
G3420 3.2 GHz 1.15 GHz $82
G3420T 2.7 GHz 200 MHz 1.1 GHz 35 W $
G3320TE 2.3 GHz 350 MHz 1 GHz $70
G3220 3.0 GHz 1.1 GHz 54 W $64 Up to dual
channel
DDR3-1333
G3220T 2.6 GHz 200 MHz 35 W $64

 Requires a compatible motherboard

Suffixes to denote:

  • K - Unlocked (adjustable CPU multiplier up to 63x)
  • S - Performance-optimized lifestyle (low power with 65 W TDP)
  • T - Power-optimized lifestyle (ultra low power with 35–45 W TDP)
  • R - BGA packaging / High performance GPU (currently Iris Pro 5200 (GT3e))

Server processors[edit]

List of announced server processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelCPU Clock rateGraphics Clock rateL3
Cache
TDPRelease
Date
Release
price
(USD)
tray / box
Motherboard
NormalTurboNormalTurboSocketInterfaceMemory
 Server 4 (8) Xeon E3 1285v3 HD P4600 (GT2) 3.6 GHz 4.0 GHz 350 MHz 1.3 GHz 8 MB 84 W June 2, 2013 $662 / — LGA
1150
DMI 2.0
PCIe 3.0
up to dual
channel
DDR3-1600
w/ ECC
1285Lv3 3.1 GHz 3.9 GHz 1.25 GHz 65 W $774 / —
1280v3 N/A 3.6 GHz 4.0 GHz N/A 82 W $612 / —
1275v3 HD P4600 (GT2) 3.5 GHz 3.9 GHz 350 MHz 1.25 GHz 84 W $339 / $350
1270v3 N/A N/A 80 W $328 / —
1268Lv3 HD P4600 (GT2) 2.3 GHz 3.3 GHz 350 MHz 1 GHz 45 W $310 / —
1265Lv3 HD (GT1) 2.5 GHz 3.7 GHz 1.2 GHz $294 / —
1245v3 HD P4600 (GT2) 3.4 GHz 3.8 GHz 84 W $276 / $287
1240v3 N/A N/A 80 W $262 / $273
1230v3 3.3 GHz 3.7 GHz $240 / $250
1230Lv3 1.8 GHz 2.8 GHz 25 W $250 / —
4 (4) 1225v3 HD P4600 (GT2) 3.2 GHz 3.6 GHz 350 MHz 1.2 GHz 84 W $213 / $224
1220v3 N/A 3.1 GHz 3.5 GHz N/A 80 W $193 / —
2 (4) 1220Lv3 1.1 GHz 1.3 GHz 4 MB 13 W September 1, 2013

 Requires a compatible motherboard

Suffixes to denote:

  • L - Low power

Mobile processors[edit]

List of announced mobile processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelProgrammable TDP[60]CPU TurboGraphics Clock rateL3
Cache
GPU
eDRAM
Release
Date
Price
(USD)
SDP[61]cTDP downNominal TDPcTDP up1-coreNormalTurbo
Performance  4 (8)  Core i7 4930MX HD 4600 (GT2) N/A N/A 57 W / 3.0 GHz 65 W / 3.7 GHz  3.9 GHz 400 MHz 1350 MHz 8 MB N/A June 2, 2013[62] $1096
4960HQ Iris Pro 5200 (GT3e) N/A 47 W / 2.6 GHz 55 W / 3.6 GHz  3.8 GHz 200 MHz 1300 MHz 6 MB 128 MB[34] September 1, 2013[63] $657
4950HQ 47 W / 2.4 GHz 55 W / 3.4 GHz  3.6 GHz June 2, 2013[62]
4900MQ HD 4600 (GT2) 47 W / 2.8 GHz 55 W / 3.6 GHz  3.8 GHz 400 MHz 8 MB N/A $570
4860EQ Iris Pro 5200 (GT3e) 47 W / 1.8 GHz N/A  3.2 GHz 750 MHz 1000 MHz 6 MB 128 MB August 2012 $508
4850EQ 47 W / 1.6 GHz 650 MHz $466
4850HQ 47 W / 2.3 GHz 55 W / 3.3 GHz  3.5 GHz 200 MHz 1300 MHz June 2, 2013[62] $468
4800MQ HD 4600 (GT2) 47 W / 2.7 GHz 55 W / 3.5 GHz  3.7 GHz 400 MHz N/A $380
4750HQ Iris Pro 5200 (GT3e) 47 W / 2.0 GHz 55 W / 3.0 GHz  3.2 GHz 200 MHz 1200 MHz 128 MB $440
4702MQ HD 4600 (GT2) 37 W / 2.2 GHz 45 W / 2.9 GHz 400 MHz 1150 MHz N/A $383
4702HQ
4700MQ 47 W / 2.4 GHz 55 W / 3.2 GHz  3.4 GHz
4700HQ 1200 MHz
4700EQ 1000 MHz $378
Mainstream  2 (4) 4650U HD 5000 (GT3) 11.5 W / 800 MHz 15 W / 1.7 GHz N/A  3.3 GHz 200 MHz 1100 MHz 4 MB $454
4610Y HD 4200 (GT2) 6 W / 800 MHz 9.5 W / 800 MHz 11.5 W / 1.7 GHz  2.9 GHz 850 MHz September 1, 2013 N/A
4600M HD 4600 (GT2) N/A N/A 37 W / 2.9 GHz  3.6 GHz 400 MHz 1300 MHz $346
4600U HD 4400 (GT2) 11.5 W / 800 MHz 15 W / 2.1 GHz  3.3 GHz 200 MHz 1100 MHz $398
4558U Iris 5100 (GT3) 23 W / 800 MHz 28 W / 2.8 GHz 1200 MHz June 2, 2013[62] $454
4550U HD 5000 (GT3) 11.5 W / 800 MHz 15 W / 1.5 GHz  3.0 GHz 1100 MHz
4500U HD 4400 (GT2) 15 W / 1.8 GHz 25 W / ? $398
 Core i5 4402E HD 4600 (GT2) N/A 25 W / 1.6 GHz N/A  2.7 GHz 400 MHz 900 MHz 3 MB September 1, 2013 $266
4400E N/A 37 W / 2.7 GHz  3.3 GHz 1000 MHz
4350U HD 5000 (GT3) 11.5 W / 800 MHz 15 W / 1.4 GHz  2.9 GHz 200 MHz 1100 MHz June 2, 2013[62] $342
4330M HD 4600 (GT2) N/A 37 W / 2.8 GHz  3.5 GHz 400 MHz 1250 MHz September 1, 2013 $266
4302Y HD 4200 (GT2) 4.5 W / 800 MHz N/A 11.5 W / 1.6 GHz  2.3 GHz 200 MHz 850 MHz N/A
4300Y 6 W / 800 MHz 9.5 W / 800 MHz $304
4300M HD 4600 (GT2) N/A N/A 37 W / 2.6 GHz  3.3 GHz 400 MHz 1250 MHz $225
4300U HD 4400 (GT2) 11.5 W / 800 MHz 15 W / 1.9 GHz  2.9 GHz 200 MHz 1100 MHz $287
4288U Iris 5100 (GT3) 23 W / 800 MHz 28 W / 2.6 GHz  3.1 GHz 1200 MHz June 2, 2013[62] $342
4258U 28 W / 2.4 GHz  2.9 GHz 1100 MHz
4250U HD 5000 (GT3) 11.5 W / 800 MHz 15 W / 1.3 GHz  2.6 GHz 1000 MHz
4210Y HD 4200 (GT2) 6 W / 800 MHz 9.5 W / 800 MHz 11.5 W / 1.5 GHz  1.9 GHz 850 MHz September 1, 2013 $304
4202Y 4.5 W / 800 MHz N/A 11.5 W / 1.6 GHz  2.0 GHz N/A
4200Y 6 W / 800 MHz 9.5 W / 800 MHz 11.5 W / 1.4 GHz  1.9 GHz June 2, 2013[62] $304
4200U HD 4400 (GT2) N/A 11.5 W / 800 MHz 15 W / 1.6 GHz 25 W / ?  2.6 GHz 1000 MHz $287
4200H HD 4600 (GT2) N/A 47 W / 2.8 GHz N/A  3.4 GHz 400 MHz 1150 MHz September 1, 2013 $257
4200M N/A 37 W / 2.5 GHz  3.1 GHz $240
 Core i3 4158U Iris 5100 (GT3) 23 W / 800 MHz 28 W / 2.0 GHz N/A 200 MHz 1100 MHz June 2, 2013[62] $342
4102E HD 4600 (GT2) N/A 25 W / 1.6 GHz 400 MHz 900 MHz September 1, 2013 $225
4100E N/A 37 W / 2.4 GHz
4100M N/A 37 W / 2.5 GHz 1100 MHz N/A
4100U HD 4400 (GT2) 11.5 W / 800 MHz 15 W / 1.8 GHz 200 MHz 1000 MHz June 2, 2013[62] $287
4020Y HD 4200 (GT2) 6 W / 800 MHz 9.5 W / 800 MHz 11.5 W / 1.5 GHz 850 MHz September 1, 2013 $304
4012Y 4.5 W / 800 MHz N/A N/A
4010Y 6 W / 800 MHz 9.5 W / 800 MHz 11.5 W / 1.3 GHz June 2, 2013[62] $304
4010U HD 4400 (GT2) N/A 11.5 W / 800 MHz 15 W / 1.7 GHz 1000 MHz $287
4005U 950 MHz September 1, 2013 $281
4000M HD 4600 (GT2) N/A 37 W / 2.4 GHz 400 MHz 1100 MHz $240
 2 (2)  Pentium 3560Y HD Graphics 6 W / 800 MHz N/A 11.5 W / 1.2 GHz 200 MHz 850 MHz 2 MB N/A
3556U N/A 15 W / 1.7 GHz 1000 MHz
3550M 37 W / 2.3 GHz 400 MHz 1100 MHz
 Celeron 2980U 15 W / 1.6 GHz 200 MHz 1000 MHz $137
2955U 15 W / 1.4 GHz $132
2950M 37 W / 2.0 GHz 400 MHz 1100 MHz $86

Suffixes to denote:

  • M - Mobile processor
  • Q - Quad-core
  • U - Ultra-low power
  • X - 'Extreme'
  • Y - Extreme-low power
  • H - BGA1364 packaging

Roadmap[edit]

The Skylake microarchitecture will be the successor to the Haswell and Broadwell architectures.

See also[edit]

 

 

http://en.wikipedia.org/wiki/Broadwell_(microarchitecture)

Broadwell (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Broadwell
Predecessor Haswell
Successor Skylake

Broadwell is Intel's codename for the second processor in its Haswell microarchitecture. In keeping with Intel's tick-tock principle, Broadwell is the next step in semiconductor fabrication, with feature size reduced to 14 nanometers.[1][2]

Broadwell will adopt the Multi-Chip Package (MCP) design. The new layout might be also moving the integrated voltage regulator (IVR) off-die and back onto the motherboards, in an attempt to reduce CPU's heat production.[3]

Broadwell will be used in conjunction with Intel 9 Series chipsets.[4]

 

 

Expected variants[edit]

Broadwell is expected to launch in three major forms:[5]

  • Broadwell-D: desktop version (LGA 1150 socket)
  • BGA version:
    • Broadwell-H: 35 W and 55 W TDP classes, for "all-in-one" systems, Mini-ITX form factor motherboards, and other small footprint formats
    • Broadwell-U: less than 15 W TDP class (SoC), for Intel's ultrabook and NUC platforms
    • Broadwell-Y: less than 10 W TDP class (SoC), for tablets and certain ultrabook-class implementations
  • Broadwell-M: mobile/laptop version (PGA socket).

Instruction set extensions[edit]

Broadwell will introduce some instruction set architecture extensions:[6][7]

Roadmap[edit]

On September 10, 2013, Intel showcased the Broadwell 14 nm processor in a demonstration at IDF. Intel CEO Brian Krzanich claimed that the chip would allow systems to provide a 30 percent improvement in power use over the Haswell chips released in mid-2013.[10]

On October 21, 2013, a leaked Intel roadmap indicated a late 2014 or early 2015 release of the K-series Broadwell on the LGA 1150 platform, in parallel with the previously announced Haswell refresh. This will coincide with the release of Intel's 9-series chipset, which may be required for Broadwell processors due to a change in power specifications for its LGA 1150 socket.[11][12]

No new mobile roadmaps have yet leaked to clarify if mobile Broadwell will be available in 2014. A leaked slide shows Broadwell-E/EP/EX in 2015.[13]

 

 

 

http://en.wikipedia.org/wiki/Skylake_(microarchitecture)

Skylake (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Skylake
Predecessor Haswell (tock)
Broadwell (tick)
Successor Cannonlake (tick)

Skylake is the codename for a processor microarchitecture to be developed by Intel as the successor to the Haswell architecture.[1] Skylake will use a 14 nmprocess.[2]

There are no official details regarding this microarchitecture's development. The first Skylake processors are expected in 2015-2016.[3]

 

 

Architecture[edit]

 
 

Cannonlake

From Wikipedia, the free encyclopedia
 
 
Cannonlake
Predecessor Skylake

In keeping with Intel's tick-tock principle, the 10 nm shrink of Skylake is due out the year after the introduction of the microarchitecture and is rumored to be codenamed "Cannonlake" (sources from 2011 indicated Skymont was to be the codename); however no official announcement has been made. However, Cannonlake is referred to as being in development already, directly from Intel's job listing.[1] Further nodes are not clear either although latest Intel development (Q3 2012) indicates 7 nm node may reach production around 2017, with 5 nm in 2019.[2]

In 2009 Intel CEO Paul S. Otellini has been quoted as saying that silicon is in its last decade as the base material of the CPU.[3]

 
 
 
 
 
 

Bonnell (microarchitecture)

From Wikipedia, the free encyclopedia
 
 
Bonnell
Intel Atom 2009.svg
Intel Atom logo
Produced 2008–present
Common manufacturer(s)
  • Intel
Max. CPU clock rate 600 MHz to 2.13 GHz
FSB speeds 400 MHz to 667 MHz
Instruction set Intel Atom x86
Cores 1, 2
Successor Silvermont
Package(s)
Core name(s)
  • Silverthorne
  • Diamondville
  • Pineview
  • Tunnel Creek
  • Lincroft
  • Stellarton
  • Sodaville
  • Cedarview
List of Intel CPU microarchitectures
MicroarchitecturePipeline stages
P5 (Pentium) 5
P6 (Pentium Pro) 14
P6 (Pentium 3) 10
NetBurst (Willamette) 20
NetBurst (Northwood) 20
NetBurst (Prescott) 31
NetBurst (Cedar Mill) 31
Core 14
Bonnell 16

Bonnell is a CPU microarchitecture used by Intel Atom processors which can execute up to two instructions per cycle.[1][2] Like many other x86 microprocessors, it translates x86 instructions (CISC instructions) into simpler internal operations (sometimes referred to as micro-ops, effectively RISC style instructions) prior to execution. The majority of instructions produce one micro-op when translated, with around 4% of instructions used in typical programs producing multiple micro-ops. The number of instructions that produce more than one micro-op is significantly fewer than the P6 andNetBurst microarchitectures. In the Bonnell microarchitecture, internal micro-ops can contain both a memory load and a memory store in connection with an ALU operation, thus being more similar to the x86 level and more powerful than the micro-ops used in previous designs.[3] This enables relatively good performance with only two integer ALUs, and without any instruction reorderingspeculative execution or register renaming. The Bonnell microarchitecture therefore represents a partial revival of the principles used in earlier Intel designs such as P5 and the i486, with the sole purpose of enhancing the performance per watt ratio. However, Hyper-Threading is implemented in an easy (i.e. low-power) way to employ the whole pipeline efficiently by avoiding the typical single thread dependencies.[3]

 

 

First generation cores[edit]

Silverthorne microprocessor[edit]

On 2 March 2008, Intel announced a new single-core Atom Z5xx series processor (code-named Silverthorne), to be used in ultra-mobile PCs and mobile Internet devices (MIDs), which will supersede Stealey(A100 and A110). The processor has 47 million transistors on a 25 mm2 die, allowing for extremely economical production (~2500 chips on a single 300 mm diameter wafer).

An Atom Z500 processor's dual-thread performance is equivalent to its predecessor Stealey, but should outperform it on applications that can use simultaneous multithreading and SSE3.[4] They run from 0.8 to 2.0 GHz and have a TDP rating between 0.65 and 2.4 W that can dip down to 0.01 W when idle.[5] They feature 32 KB instruction L1 and 24 KB data L1 caches, 512 KB L2 cache and a 533 MT/s front-side bus. The processors are manufactured in 45 nm process.[6][7]

Diamondville microprocessor[edit]

 

The Intel Atom N270

On 2 March 2008, Intel announced lower-power variants of the Diamondville CPU named Atom N2xx. It was intended for use in nettops and the Classmate PC.[8][9][10] Like their predecessors, these are single-core CPUs with Hyper-Threading.

The N270 has a TDP rating of 2.5 W, runs at 1.6 GHz and has a 533 MHz FSB.[11] The N280 has a clock speed of 1.66 GHz and a 667 MHz FSB.[12]

On 22 September 2008, Intel announced a new 64-bit dual-core processor (unofficially code-named Dual Diamondville) branded Atom 330, to be used in desktop computers. It runs at 1.6 GHz and has a FSB speed of 533 MHz and a TDP rating of 8 W. Its dual core consists of two Diamondville dies on a single substrate.[13]

During 2009, Nvidia used the Atom 300 and their GeForce 9400M chipset on a mini-ITX form factor motherboard for their Ion platform.

First generation power requirements[edit]

The relatively low power Atom CPU was originally used with a cheaper, not so electricity-efficient chipset such as the Intel 945G

Although the Atom processor itself is relatively low-power for an x86 microprocessor, many chipsets commonly used with it dissipate significantly more power. For example, while the Atom N270 commonly used in netbooks through mid-2010 has a TDP rating of 2.5 W, an Intel Atom platform that uses the 945GSE Express chipset has a specified maximum TDP of 11.8 W, with the processor responsible for a relatively small portion of the total power dissipated. Individual figures are 2.5 W for the N270 processor, 6 W for the 945GSE chipset and 3.3 W for the 82801GBM I/O controller.[14][11][15][16] Intel also provides a US15W System Controller Hub-based chipset with a combined TDP of less than 5 W together with the Atom Z5xx (Silverthorne) series processors, to be used in ultra-mobile PCs and MIDs,[17] though some manufacturers have released ultra-thin systems running these processors (e.g. Sony VAIO X).

Initially, all Atom motherboards on the consumer market featured the Intel 945GC chipset, which uses 22 watts by itself. As of early 2009, only a few manufacturers are offering lower-power motherboards with a 945GSE or US15W chipset and an Atom N270, N280 or Z5xx series CPU.

Second generation cores[edit]

Pineview microprocessor[edit]

 

New Intel Atom N450 SLBMG 1.66GHz 512kB L2 BGA559

On 21 December 2009, Intel announced the N450, D510 and D410 CPUs with integrated graphics.[18] The new manufacturing process resulted in a 20% reduction in power consumption and a 60% smaller die size.[19][20] The Intel GMA 3150, a 45 nm shrink of the GMA 3100 with no HD capabilities, is included as the on-die GPU. Netbooks using this new processor were released on 11 January 2010.[19][21] The major new feature is longer battery life (10 or more hours for 6-cell systems).[22][23]

This generation of the Atom was codenamed Pineview, which is used in the Pine Trail platform. Intel's Pine Trail-M platform utilizes an Atom processor (codenamed Pineview-M) and Platform Controller Hub (codenamed Tiger Point). The graphics and memory controller have moved into the processor, which is paired with the Tiger Point PCH. This creates a more power-efficient 2-chip platform rather than the 3-chip one used with previous-generation Atom chipsets.[24]

On 1 March 2010 Intel introduced the N470 processor,[25] running at 1.83 GHz with a 667 MHz FSB and a TDP rating of 6.5 W.[26]

The new Atom N4xx chips became available on 11 January 2010.[27] It is used in netbook and nettop systems and includes an integrated single-channel DDR2memory controller and an integrated graphics core. It also features Hyper-Threading and is manufactured on a 45 nm process.[28] The new design uses half the power of the older Menlow platform. This reduced overall power consumption and size makes the platform more desirable for use in smartphones and other mobile internet devices.

The D4xx and D5xx series support the x86-64 bit instruction set and DDR2-800 memory. They are rated for embedded use. The series has an integrated graphics processor built directly into the CPU to help improve performance. The models are targeted at nettops and low-end desktops. They do not support SpeedStep.

The Atom D510 dual-core processor runs at 1.66 GHz, with 1 MB of L2 cache and a TDP rating of 13 W.[29] The single-core Atom D410 runs at 1.66 GHz, with 512 KB of L2 cache and a TDP rating of 10 W.[30]

Tunnel Creek microprocessor[edit]

Tunnel Creek is an embedded Atom processor used in the Queens Bay platform with the Topcliff PCH.

Lincroft microprocessor[edit]

The Lincroft (Z6xx) with the Whitney Point PCH is included in the Oak Trail tablet platform. Oak Trail is an Intel Atom platform based on Moorestown. Both platforms include a Lincroft microprocessor, but use two distinct input/output Platform Controller Hubs (I/O-PCH), codenamed Langwell and Whitney Point respectively. Oak Trail was presented on 11 April 2011 and was to be released in May 2011.[dated info][31]The Z670 processor, part of the Oak Trail platform, delivers improved video playback, faster Internet browsing and longer battery life, "without sacrificing performance" according to Intel. Oak Trail includes support for 1080p video decoding as well as HDMI. The platform also has improved power efficiency and allows applications to run on various operating systems, including Android, MeeGo and Windows.

Stellarton microprocessor[edit]

Stellarton is a Tunnel Creek CPU with an Altera Field Programmable Gate Array (FPGA).

Sodaville SoC[edit]

 

Sodaville is a consumer electronics Atom SoC.

Groveland SoC[edit]

 

Groveland is a consumer electronics Atom SoC.

Third generation cores[edit]

The 32 nm shrink of Bonnell is called Saltwell.

Cedarview microprocessor[edit]

Intel released their third-generation Cedar Trail platform (consisting of a range of Cedarview processors[32] and the NM10 southbridge chip) based on 32 nm process technology in the fourth quarter of 2011.[31]Intel stated that improvements in graphics capabilities, including support for 1080p video, additional display options including HDMI and DisplayPort, and enhancements in power consumption are to enable fanless designs with longer battery life.

The Cedar Trail platform includes two new CPUs, 32 nm-based N2800 (1.86 GHz) and N2600 (1.6 GHz), which replace the previous generation Pineview N4xx and N5xx processors. The CPUs also feature an integrated GPU that supports DirectX 9.

In addition to the netbook platform, two new Cedarview CPUs for nettops, D2700 and D3200, were released on 25 September 2011.[33]

In early March 2012 the N2800-based Intel DN2800MT motherboard[34] started to become available. Due to the use of a netbook processor, this Mini-ITX motherboard can reach idle power consumption as low as 7.1 W.[35]

Penwell SoC[edit]

Penwell is an Atom SoC that is part of the Medfield MID/Smartphone platform.

Berryville SoC[edit]

 

Berryville is a consumer electronics Atom SoC.

Cloverview SoC[edit]

Cloverview is an Atom SoC that is part of the Clover Trail tablet platform.

Centerton SoC[edit]

 

In December 2012 Intel launched the 64-bit Centerton family of Atom CPUs, designed specifically for use in Bordenville platform servers.[36] Based on the 32 nm Saltwell architecture, Centerton adds features previously unavailable in most Atom processors, such as Intel VT virtualization technology, and support for ECC memory.[37]

Briarwood SoC[edit]

 

Briarwood is an Atom SoC that is designed for a server platform.

Roadmap[edit]

 
 
 
 

Silvermont

From Wikipedia, the free encyclopedia
 
 
Silvermont
Predecessor Bonnell
Saltwell
Successor Goldmont

Silvermont is a new low power SoC processor microarchitecture from Intel. Silvermont will form the basis for two consumer SoC families; Merrifield intended forsmartphones and Bay Trail aimed at tabletshybrid devicesnetbooksnettops, and embedded/automotive systems. As well as Avoton SoCs for micro-servers and storage devices; and Rangeley SoCs targeting network and communication infrastructure.[1]

Silvermont was announced to the media on May 6, 2013 at Intel's headquarters at Santa ClaraCalifornia.[2][3] Intel has repeatedly said the first Bay Trail devices will be available during the Holiday 2013 timeframe, while leaked slides show the most recent release window for Bay Trail-T as August 28-September 13, 2013.[4]Both Avoton and Rangeley were announced as being available in the second half of 2013. The first Merrifield devices are expected in 1H14.[5]

 

 

Design[edit]

Silvermont will be the first Atom processor to feature an out-of-order architecture.[6]

Technology[edit]

List of Silvermont processors[edit]

Desktop processors (Bay Trail-D)[edit]

List of upcoming desktop processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelTDPCPU TurboGraphics Clock rateL3
Cache
Release
Date
Price
(USD)
1-coreNormalTurbo
Value  4 (4) Pentium J2850 Intel HD Graphics
(4 EU)
10W/ 2.4 GHz N/A 688 MHz 792 MHz 2 MB 3Q13 OEM
Celeron J1850 10W/ 2.0 GHz $82
 2 (2) J1750 10W/ 2.4 GHz 750 MHz 1 MB OEM
 4 (4) J1900[9] 10W/ 2.0 GHz ?? 2 MB OEM

Server Processors (Avoton)[edit]

List of upcoming server processors as follows: [10]

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelTDPCPU TurboGraphics Clock rateL3
Cache
Release
Date
Price
(USD)
1-coreNormalTurbo
Server  8 (8) Atom C2750 N/A 20W/ 2.4 GHz 2.6 GHz N/A N/A 4MB 3Q13 $171
C2730 12W/ 1.7 GHz 2.0 GHz $150
 4 (4) C2550 14W/ 2.4 GHz 2.6 GHz 2MB OEM
C2530 9W/ 1.7 GHz 2.0 GHz
 2 (2) C2350 6W/ 1.7 GHz 2.0 GHz 1MB

Communications Processors (Rangeley)[edit]

List of upcoming communications processors as follows: [11]

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelTDPCPU TurboGraphics Clock rateIntel
QuickAssist
L3
Cache
Release
Date
Price
(USD)
1-coreNormalTurbo
Communications  8 (8) Atom C2758 N/A 20W/ 2.4 GHz N/A N/A N/A Yes 4MB 3Q13 $171
C2738 20W/ 2.4 GHz No $171
C2718 18W/ 2.0 GHz Yes $150
 4 (4) C2558 15W/ 2.4 GHz Yes 2MB $86
C2538 15W/ 2.4 GHz No $86
C2518 15W/ 1.7 GHz Yes $75
 2 (2) C2358 7W/ 1.7 GHz 2.0 GHz Yes 1MB $49
C2358 7W/ 1.7 GHz 2.0 GHz No 1MB $49

Embedded/automotive processors (Bay Trail-I)[edit]

List of embedded processors as follows: [12]

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelTDPCPU TurboGraphics Clock rateL3
Cache
Release
Date
Price
(USD)
1-coreNormalTurbo
Embedded  4 (4) Atom E3845 Intel HD Graphics
(4 EU)
10W/ 1.91 GHz N/A 542 MHz 792 MHz 2MB 4Q13 $52
 2 (2) E3827 8W/ 1.75 GHz 1MB $41
E3826 7W/ 1.46 GHz 533 MHz 677 MHz $37
E3825 6W/ 1.33 GHz N/A $34
 1 (1) E3815 5W/ 1.46 GHz 400 MHz 512KB $31

Mobile processors (Bay Trail-M)[edit]

List of upcoming mobile processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelTDPCPU TurboGraphics Clock rateL3
Cache
Release
Date
Price
(USD)
1-coreNormalTurbo
Value  4 (4) Pentium N3510 Intel HD Graphics
(4 EU)
7.5W/ 2.0 GHz N/A 750 MHz N/A 2MB 3Q13 OEM
Celeron N2920[13] 7.5W/ 1.6 GHz ? OEM
N2910 7.5W/ 1.6 GHz 756 MHz $132
 2 (2) N2810 7.5W/ 2.0 GHz 1MB OEM
N2805 4.5W/ 1.46 GHz 667 MHz

Tablet processors (Bay Trail-T)[edit]

List of upcoming tablet and hybrid processors as follows:

Target
segment
Cores
(Threads)
Processor
Branding & Model
GPU ModelSDP[14]TDPMax CPU
Turbo
GraphicsClock rateL2
Cache
Memory
Standard
Max
Memory Bandwidth
Max
Memory Supported
Max Display ResolutionRelease
Date
Price
(USD)
NormalTurbo
Value  4 (4) Atom Z3770 Intel HD Graphics (4 EU) 2W/ 1.46 GHz ? 2.39 GHz 311 MHz 667 MHz 2 MB LPDDR3 1067 Dual Channel 17.1 GB/s 4GB 2560×1600 September 11, 2013 $37.00
Z3770D 2.2W/ 1.5 GHz 2.41 GHz 313 MHz 688 MHz DDR3L-RS 1333 Single Channel 10.6 GB/s 2GB 1920×1280
Z3740 2W/ 1.33 GHz 1.86 GHz 311 MHz 667 MHz LPDDR3 1067 Dual Channel 17.1 GB/s 4GB 2560×1600 $32.00
Z3740D 2.2W/ 1.33 GHz 313 MHz 688 MHz DDR3L-RS 1333 Single Channel 10.6 GB/s 2 GB 1920×1280
 2 (2) Z3680  ?/ 1.33 GHz 2.0 GHz 311 MHz 667 MHz 1 MB LPDDR3 1067 Single Channel 8.5 GB/s 1 GB 1280×800 ?
Z3680D  ?/ 1.33 GHz 311 MHz 688 MHz DDR3L-RS 1333 Single Channel 10.6 GB/s 2 GB 1920×1280

Smartphone processors (Merrifield)[edit]

Roadmap[edit]

Airmont is the 14 nm shrink of Silvermont.[6]

 

 

 

 

 

 

 

 

 

posted @ 2013-11-18 23:02  baihuahua  阅读(3500)  评论(0编辑  收藏  举报