The annual Intel IDF information technology summit is coming. In annual April, Intel will bring a lot of exciting new technologies to us, will insult and move the series to apply to more fields while when will propose in 2009 intelligence calculates the idea. In 2010, the large-scale application of the embedded system will become the light spot of this conference.
IDF 2010 is by ” lead the machine intelligently and create tomorrow ” It is a theme, aim at catching intelligence computation and individualized Internet era opportunity, announce industry’s development pulses, the cooperative partner leads innovation, seize and recover the first chance increasing hand in hand. Intel Corporation global Vice President and concurrently the intersection of Chinese and Yang, president of district, chat, understand this IDF latest tendency for us March 25, carry on on relevant problem and answer thoroughly, let as follows us know in the form of interview. ...

Soon, AMD should release one’s own first batch of tabletop six key processor Phenom II X6 series 1000T too, not only will be well below 999 dollars of Intel at the price but also support the intellectual speed technology of the trends too, AMD is called ” Turbo CORE ” .

Similar to Intel Turbo Boost Rui frequently Technology, AMD Turbo CORE technology can adjust different central frequency intelligently too, it is suitable to be insensitive to much thread, but expect much the applied environment of the frequency, but Because AMD does not have power door that is accused of power gating Technology, so realize the principle different from some of way, mainly utilize P-State power to manage the state to switch over.

In Phenom II X6 processor, if three or more cores are in idle state, Turbo CORE will be started, reduces three pieces of idle central frequency among them to 800MHz, improve the whole voltage at the same time, improve other three pieces of central frequency 400MHz or 500MHz, depend on different types specifically. Otherwise, six cores will all run according to the normal frequency.

As for the new acceleration situation of four key Phenom II X4 series 900T, AMD was not explained specifically, but should reduce two pieces of idle central frequency, and improve by other two central speeds.

Type Primitive mainly frequently Accelerate at most It is buffered that tertiary TDP Release time Phenom II X6 1090T 3.2GHz3.6GHz6MB125WQ2 Phenom II X6 1075T 3.0GHz3.5GHz6MB125WQ3 Phenom II X6 1055T 2.8GHz3.3GHz6MB125W/95WQ2 Phenom II X6 1035T 2.6GHz3.1GHz6MB95WQ2 Phenom II X4 960T 3.0GHz3.4GHz6MB95WQ2

Obviously, AMD can not be done as splendidly as Intel. First of all, The idle core can’t be totally closed, can only switch over to the low-speed state ,Still there is energy consumption; Secondly, In order to improve other central frequency, must pressurize for the whole processor ,This will influence the consumption and limit and accelerate the range; Have it just, Accelerating or lowering the speed can only be directed against three or two The core, but can’t regulate every core alone . Even so, this kind of speed technology is still very useful in a situation that present much thread environments are deficient, there are six cores and cores of high frequency three at the same time in other words.

Turbo CORE totally because of hardware too in technology, Acquiescence turns on and is controlled by processor automatically ,Do not need to install any special software, drive or tool. Obviously, the speed that this kind of frequency switches over is very quick, it is enough to let Windows system shift the thread among different cores.

When Turbo CORE opens, even accelerate to the supreme frequency, the energy consumption of the whole processor will not exceed the biggest heat and design the consumption either, it will also be much lower when being idler.

In addition In Turbo CORE open complete course, Cool ‘ n ‘ Quiet cool and quiet power-saving technology influenced ,They will cooperate each other, check whether can reduce half the central frequency at any time, accelerate other cores at the same time.

Finally,Socket AM2 on it is existing /support 125W heat not to be lasted consumption only AM3 socket main board,and BIOSs new, can support Phenom II X6/X4 processor and Turbo CORE speed technology.

...

Weisheng released and combined the group VIA VX900 of the chip to the new style form chip of own VIA Nano 3000 serial processors a few days ago, join ChromotionHD 2.0, attack the high clear video mainly to accelerate.

ChromotionHD 2.0 video engine supports the blue only high clear hardware to accelerate, and offered and accelerated ability to FullHD high clear hardware of the latest edition H.264, WMV-HD/VC-1, MPEG-2, MPEG-4/DivX form. Support one pair of videos of way to export, the interface supports DisplayPort, HDMI, DVP, VGA and LVDS/TMDS.

VX900 chip group encapsulates the area 31x31mm, support Nano, Nano 3000, C7-M and Eden processor of Weisheng, a pair of passway DDR3 1066MHz/DDR2 667MHz memory. Built-in DirectX Chrome9 HCM 3D figure core key frequency 250MHz of 9.0 grades ,128bit 2D engine. Offer a PCI-E 2.0 x8 and three PCI-E 2.0 x1s to expand the trough, one pair of PCI slots, supports VIA Vinyl HD 8 sound channel audio frequency. Support pairs of SATA II expand, the intersection of SD/MMS/MMC and media reader, 8x USB 2.0 interface.

Weisheng says, chip group VX900 has already offered the project customer to use now, the price has not been announced.

...

After Nehalem-EP Xeon 5500 series four key processor raises great revolution the most over the past 15 years of the server field, Intel looked at eight more advanced key types in advance today, the code ” Nehalem-EX ” ,Face No. four and No. eight Server field.

Nehalem-EX will replace 6 series key Dunnington Xeon 7400 existingly, main technological the intersection of characteristic and brief introduction is as follows:

-Because of Nehalem framework, LGA1567 new interface

-45nm HKMG craft is made, the consumption 130W that heat is designed

-Eight kernels at most of single processor

-Support ultra thread technology, 16 thread at most

-QuickPath interconnected framework, offers four pieces of high bandwidth QPI to connect

-Through QPI bus and third party’s nodal controller, support No. eight to connect in parallel at most, 128 pieces of thread at most

-24MB shares the tertiary buffer memory

-Integrated memory controller

-Every processor supports 16 pieces of memory slot at most

-Support Turbo Boost dynamic speed technology

-It is interconnected to expand memory buffering and memory

-Compare the Xeon 7400 memory bandwidth of highest 9 times to promote

-Support the virtual machine real-timely to move dynamically, including Enhanced VMotion of VT FlexMigration Assist and VMware of Intel

-Support ” RAS ” from Itanium family stability /usability /but service ,Including ” MCA Recovery ” automatic check-up framework resumes

Intel expresses, there is RAS technology facing advanced enterprise’s customer, Nehalem-EX can bring lower disposition cost, higher systematic function, less electric rate bills, and realize the standardization of the elastic IT environment and help to accelerate the transformation from RISC platform to Intel Xeon/Itanium platform. Itanium faces system of the above-mentioned 8 Luis to dispose, can offer dependability and expansibility of the large-scale host computer rank for working environment of the key task mainly, especially suitable for highly much thread to support.

Nehalem-EX processor scheduled to release in the second half of 2009 ,Eight OEM manufacturers including IBM have finished at least 15 sets of relevant server systems design now, among them there are 128 thread systems.

pay attention to new Itanium LOGO

Will fight in different parts on many ways MP soon in Nehalem framework Server field

Nehalem-EX Xeon overview

No. four is employed: 64 thread, 64 pieces of memory slot

Compare Xeon 7400: 2 times of memory, 2.7 times of thread, 1.5 times of buffer memory, and RAS

Compare Xeon 7400: Memory bandwidth of 9 time, 2.5 times of database performance, 1.7 times of integer performance, 2.2 times of floating dot performance

12 makes ...

AMD released Opteron 6100 serial server processor products of new generation completely formally today, this global first batch of 8 core and 12 key the intersection of x86 and processor too, face pairs of way and four No. the intersection of server and market mainly.

First, new encapsulation form

Opteron 6100 serial processors develop codes ” Ma Nicaragua storehouse ” Magny-Cours ,Basically can be regarded as Two six cores ” Istanbul ” Istanbul Adopt many chip module MCMs Way encapsulate at silicon chip together coming, adopt new 1944 stitch Socket G34 encapsulate the interface ,But Non-core Uncore Very great change takes place partly: The memory controller upgrades and supports standard DDR3-1333 and low-voltage LV-DDR3-1333, and done the enhancement in many aspects in order to control buffering consistency.

Memory, the node of Opteron 6100 inside has two DDR3 passways, The whole processor supports to store in four passways ,Bandwidth 42.7GB/s of peak value of the theory ,But north the intersection of bridge and speed have only 1.8GHz, so two 64 north bridge it joins forces bandwidth to be only 28.8GB/s, this is for keeping the cost of the consumption too.

It is the structure sketch of 12 key Opteron 6100 processors to make the picture, among them red dotted line represents four memory passways, blue line represents internal HT buffers consistency to connect, the gray line represents outside buffers HT to connect, green line represents inconsistency I/O HT is connected, among them the thick line is that HT x16 bandwidth, fine rule are HT x8 bandwidth.

Opteron 6100 series been adopted and built up DAC 2.0 in succession the second generation of straight ,Two kernels Die within each processor Or the node Node Between if you can’t there is bus, link up directly,be be be is there bandwidth bottleneck problem at alleddedded ,Superior to Intel and pass the outside front bus FSB that year Connect two kernel methods.

The bottom is the interconnected situation in case of single No. and one pair of ways of a Opteron 6100 processor, no longer detail can consult here here as space is limited .

1 2 3 4 5 Make ...

To many server system managers, will meet the following question: User agree management, unify control, the intersection of data and store, the intersection of management and the intersection of person who dispose and question in unison in unison systematically, to these questions, the dawn company has put out Gridview 2.0 administrative system, this administrative system can realize the region characteristic of the wide-area network of the customer service and some characteristic of data resource, combine the detached island of information, unified views in many angles of the opportunity hierarchical structure offering users to employ resources.
Inherit dawn DCMM2, DCAS, dawn Gridview administrative system v1.1,etc. and control and manage main function and technological characteristic of the products in dawn Gridview2.0 administrative system, and increase multiple outstanding functional developments and the flying contingent become control the administrative system products. This system has offered the unified centralized type to control the platform Portal, possess expansibility, integrating, dependability and ease of use, have offered the integrated interface to various commercial, management tools researched and developed by oneself, thus meet at the same time different fields such as the environment, hardware, software to variant producer mainframes,etc. carry on the demand which control management.
Dawn Gridview is 2.0 of new generation server what administrative system adopt it is distributed, level, module suffixation of a nonsyllabic, can expanding, a customized design and framework, offer and synthesize the integrated cluster administrative system, the range of application is relatively extensive, suitable for the unit, cluster, grid foundation, it last several kinds of arrangement, dispositions, control, manage mainframe environment inside it thus not realize LAN and across wide-area network,etc. function. ...

For a long time, IBM has been to enter into an alliance with VMware all the time in the virtual field. But like other manufacturers, IBM realizes gradually that offers the importance chosen more to user, so this situation is changing slowly. Recently, IBM reveals that will choose Red Hat to offer the virtual technology of ground floor for the developer’s cloud technology for them. ...

The San Francisco autumn at last the end of September is on IDF 2009 forum, Intel showed 22nm craft to common people brilliant and round for the first time, and announce that would release the relevant products in the second half of 2011.

Certainly, the realization of the new semi-conductive craft is not that a family of Intel can accomplish, the back contributes the distinguished person of the semi-conductive apparatus quietly but not often publicises. The analyst Jagadish Iyer of Arete Research LLC Company points out in a report, Intel determines the supplier of 22nm photoengraving technological equipment finally soon, the selected finally one is Dutch ASML Holding NV and Japanese Nikon .

In fact in earlier 45nm generation, ASML and Nikon have become photoengraving apparatus suppliers of Intel in pairs, but Intel immersing photoengraving technology of application for the first time at 32nm node, only one of Nikon offers the relevant apparatus, ASML will come back again nowadays.

The photoengraving machine that ASML offers is this time ” NXT: 1950i ” ,Each average price 40 million euro 54,400,000 dollars ,Responsible for 40% Ago say FEOL Photoengraving layer; The photoengraving machine of Nikon is ” S620D ” ,Offer 30 million dollars ,Responsible for 60% And then dish BEOL Photoengraving layer. 22nm craft of Intel will 45 photoengraving layers, 55% among them need to carry on the immersing type photoengravingly.

On crafts in 22nm, Intel will continue using 193nm immersing type photoengravingly, and a street map revealed recently reveals it will keep lasting until 11nm craft node ,This just means being extremely purple outside EUV Photoengraving after pushing.

Intel can produce about 5000 brilliant and round 22nms per month when Jagadish Iyer has been estimated by the end of this year, four brilliant round factories go into operation with all strength among 2011-2012 years, each one can produce about 45,000 brilliant and round 22nms every month, it is about 180,000 pieces to total in one month.

ASML photoengraving machine

Photoengraving machine of Nikon

...

With storing the maturity of the network technique, the scheme that high capacity and complexity are higher is realized to become very common. The memory platform of different frameworks has increased the difficulty of management greatly, it is very difficult to get effective use so as to store resources at the same time. Users hope that can offer the basic framework of their appropriate memory through a kind of special technology, they reduce and manage the cost in effective help, and meet the harsh service requirement. The appearance of virtual technology, strengthen productivity for enterprises to a great extent, the utilization ratio of the assets of increasing, and effective management enterprise’s operation environment, and different virtual technology offers different support to users.

What is so virtual technology on earth? Come, say, shield physical the intersection of apparatus and complexity by way of shining upon or abstract on being in fact broad. It can simplify the management of the infrastructure effectively, increases utilization ratio and ability of IT resources, for example the server, network or memory. Virtual to simplify managing, helping users to be effective and centralized and shared resource, so as to utilize resources to the greatest extent.

Virtual technology used widely in trade, main application concentrate on the intersection of memory and virtual, the intersection of network and virtual, fictitious server and store virtually among them.

The memory is virtual

The memory is a main assembly to computer, it is determining that be able to operate scale and quantity of the application program on the host computer. With the progress of technology, memory technology is updating too, the cost is being reduced greatly too. And the fictitious memory makes the application program have logic memory independent of existing and physical memory resource especially, fictitious memory managing device VMM at the same time If you can’t develop make, can more run at memory at the same time a plurality of application program.

Fictitious memory managing device VMM

Then how did VMM work? In the fictitious memory, the memory address is divided into and fixes all kinds of block. A so-called paging of handlings keep at magnetic disc the intersection of memory and page that does not move about, read to the physical memory when needing, this utilizes existing physical memory effectively between different process. The magnetic disc space that VMM uses, in order to exchange the file, the similar physical memory of the function, transfer the memory data not used in the system to it, in order to more memory spaces offer the process of the activity, in addition make the memory more flexible.

The network is virtual

Mention the network is virtual, the fictitious LAN is its typical representative, it can offer a flexible and convenient network management environment, make the large-scale network easier to manage, can realize optimization the most of the network by concentrating on disposing the physical apparatus of different positions.

12 makes ...

Development and maturity of virtual technology with the data centre, a lot of enterprises have already carried on the central apparatus of data including server, storing, network virtually. A group of people or things of the king of general manager in ESG Chinese district introduce: “Several years ago, it needn’t be virtual that we talked about mostly using. It has already become an inevitable trend but virtual now, it is mainly because of the reason of two respects: First because the input of the data centre is too big, a lot of enterprises are difficult to pay; Second, because of the importance of the data, the cost of the delaying machine is larger and larger. This makes virtually to become an inexorable trend that the data centre develops, especially enterprise layer users, have not considered yet now there is few virtual technology. ”

ESGEnterprise Storage Group An investigation done toward 377 large-scale enterprises and 329 medium-sized enterprises reveals before this, 61% of the large-scale enterprise and 41% of the medium-sized enterprises has all disposed the server virtually. All in all, there are 52% that have already in all enterprises polled adopted the virtual technology of the server, the enterprises left have mostly expressed the interest in virtual technology too, and show that probably adopts this technology in the future.

It is only a forerunner of the virtual market that the server is virtual, under the leadership of virtual technology of the server, store the application of the virtual, virtual technology of network and run up too. But how manage and utilize the intersection of data and these fictitious and physical resource under in the center better, become customer’s problem badly in need of solution. The virtual era after just as IBM, a concept put forward in 2009, we have already entered now ” ” .

” back virtual era ” Management difficult problem

Virtual technology has really brought a lot of advantages to user, for example improved resource utilization ratios, having reduced server, quantity and cost,etc. stored, but the popularization of virtual application brought some questions at the same time.

NetApp company Greater China Chen Wen, general manager of district is beautiful to say: “It is virtual and very hot this one or two years, a lot of enterprises begin to adopt the virtual scheme. However, though some enterprises have disposed virtual application, is still limited to very small range, at a stage in initial stage of comparison, and we see some may not carry on complete test to employ inside the virtual environment. A lot of enterprises have formed a virtual ‘ detached island ‘ after disposing virtually, may have got some advantages in this virtual detached island, ones that have reached on benefit are saved. But enterprises want real, a large amount of virtual technology of adoption, also need some time. ”

A group of people or things of the king think too, say from management, because virtual technology has not realized standardization yet, so will bring some questions. For example, to most users, it has not been already totally virtual in the whole data centre, may have only adopted some one item of virtual technology, whether to or just employ partly or not to carry on virtually, other IT systems or employ hitting and has not adopted virtually, in this case, the complexity of management certainly will increase. ESG has been done on North American market one is counted, the question that about 40% of data centre users have a headache most, just carried on partly virtually in the data centre, namely should manage the fictitious environment, the environment not virtual of management, often let IT personnel feel headache deeply again.

Fortunately, manufacturers have seen already these of users’ are worried, and begin to try to offer the pointed solution. According to a group of people or things of the king, can already see increasingly much virtual management software on the North American market now, one of them is called ” Virtual Instruments ” Virtual for management company,that they develop out for products,for one management softwares virtual of station. Utilize by the software,need user only through one of simple interfaces,can server, storing at system, stand up network all management.

“This market is so, so long as there is market demand, so long as some some problems are not solved, manufacturers appear, specially solve this problem, thus open a new market. ” A group of people or things of the king appraise like this.

123 makes ...