CN107491408A - A kind of calculation type server node - Google Patents
A kind of calculation type server node Download PDFInfo
- Publication number
- CN107491408A CN107491408A CN201710635683.1A CN201710635683A CN107491408A CN 107491408 A CN107491408 A CN 107491408A CN 201710635683 A CN201710635683 A CN 201710635683A CN 107491408 A CN107491408 A CN 107491408A
- Authority
- CN
- China
- Prior art keywords
- computing unit
- unit
- pcie
- groups
- signals
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/38—Information transfer, e.g. on bus
- G06F13/42—Bus transfer protocol, e.g. handshake; Synchronisation
- G06F13/4204—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus
- G06F13/4221—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being an input/output bus, e.g. ISA bus, EISA bus, PCI bus, SCSI bus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2213/00—Indexing scheme relating to interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F2213/0026—PCI express
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Transfer Systems (AREA)
- Multi Processors (AREA)
Abstract
The present invention provides a kind of calculation type server node, it is characterised in that comprising centralized processing unit, administrative unit, computing unit 1 and computing unit 2, computing unit 1 carries out data communication with computing unit 2 by UPI buses;Computing unit 1 and computing unit 2 are respectively provided with three PCIE bus control units PE1, PE2, PE3.The present invention is combined by using the OCUlink connectors based on PCIE expansion techniques and Cable interconnection techniques, is broken through traditional calculations server and is being applied upper above-mentioned limitation.The present invention has higher data storage capacities, more flexible PCIE extended capabilities, more kinds of disk storage applications.
Description
Technical field
The present invention relates to computer realm, and in particular to server node architecture design.
Background technology
The calculation type server of current traditional OCP standards can only support 3.5 cun of disk positions of 1 piece of standard, the biography of SATA disks
Defeated speed only has 6 ~ 12Gbps, can not meet the application of current big data and high speed data transfer.What simultaneity factor was supported
PCIE device is supported by CPU0, and the resource free time on CPU1 does not use, and can produce " isolated island " effect, and system resource makes
With inequality, can not meet the needs of current Distributed Application.
The content of the invention
The present invention is solution above-mentioned technical problem.Therefore, the present invention provides a kind of calculation type server node, it has branch
Hold multiple storage dish extensions, message transmission rate is high, solves the advantages of " isolated island " effect.
To achieve these goals, the present invention adopts the following technical scheme that.
A kind of calculation type server node, comprising centralized processing unit, administrative unit, computing unit 1 and computing unit 2,
Computing unit 1 carries out data communication with computing unit 2 by UPI buses.Computing unit 1 and computing unit 2 are respectively provided with three
PCIE bus control units PE1, PE2, PE3.
X16 signals in the PE1 of computing unit 1 are divided into 2 independent x8 signals and for connecting Mezz card connectors
ConnectionA and Connect B.X16 signals in the PE2 of computing unit 1 are switched over by resistance can connect concentration
Processing unit supports to be integrated in the QAT data enciphering/deciphering technologies that centralized processing unit carries, or is connected to x24 physics rule
The PCIe card for standard of being arranged in pairs or groups on the PCIE Slot of lattice;X16 signals in the PE3 of computing unit 1 are divided into the x8 signals of 2 groups of independence,
Wherein 1 group of x8 is connected on the PCIE Slot of x24 physical specifications, and 1 group of x8 is connected to the PCIE Slot of x8 physical specifications in addition
On.
The x16 signals of PE1 in computing unit 2 are divided into the x4 signals of 4 groups of independence, are drawn and carried out by 4 groups of OCUlink
PCIE expanded applications.
Preferably, centralized processing unit mainly realizes the control to low speed, function simple device in system, total by SATA
Line connects 2 M.2 equipment, for depositing the operating system needed for computing system, such as Windows, Linux or Vmware.
Preferably, centralized processing unit supports 8 SATA devices by high density OCUlink connectors, supports that establishment is soft
RAID, for depositing the data needed for computing system.
Preferably, centralized processing unit is signally attached to meet on the connector Connect C of OCP specifications by 4 groups of KR,
PHY cards can be arranged in pairs or groups to support to be integrated into the network controller in centralized processing unit, realized and supported according to the different PHY that arranges in pairs or groups
4*1Gb or 4*10Gb networks power port and optical port network.
Preferably, management control unit is mainly used in realizing to key electronic device in computing system and whole system operation
The monitoring of state.
Preferably, administrative unit is designed based on the chips of Aspeed 2500, has special management port to be used for supporting visitor
Management of the family remotely based on standard IMPI instructions, while USB and USB interface, the demand for being easy to local management are also provided.
Beneficial effects of the present invention:
1st, higher data storage capacities
The present invention supports 16 memory banks, 12 of more traditional OCP nodes, density lifting 33%, can support that 1024GB's is quick
Data storage capacities.
2nd, more flexible PCIE extended capabilities
Pass through the resource allocation of optimization calculation unit PCIE controllers, in the integrated FPGA of collocation computing unit, more traditional OCP
Calculate node standard OCP Mezz cards still can use.Except meeting that standard OCP calculate nodes support 2 from computing unit 1
The application of PCIE expansion cards is outer, the mode extended by using OCUlink and Cable, can also support from computing unit 1 and calculate
Unit 2 supports the application of this more resource load stabilization of 2 PCIe cards respectively.
3rd, more kinds of disk storage application
The mode extended by using OCUlink and Cable, except meeting 1 3.5 cun of disk of traditional OCP calculate nodes support
Using, the application that can also support 2 2.5 cun of NVMe to store, a width of PCIE3.0 x4 of NVMe memory interface bands, it is possible to provide 32Gbps
Transmission rate, far above current 6 ~ 12Gbps of SATA transmission rate.
Brief description of the drawings
Fig. 1 is the calculation type server node architecture design figure of embodiment 1.
Fig. 2 is that the calculation type server node of embodiment 2 supports PCIE device topological diagram.
Embodiment
The invention will be further described with embodiment below in conjunction with the accompanying drawings.
Embodiment 1
As shown in figure 1, a kind of calculation type server node, comprising centralized processing unit, administrative unit, computing unit 1 and is calculated
Unit 2, computing unit 1 carry out data communication with computing unit 2 by UPI buses.Computing unit 1 and computing unit 2 are respectively provided with
Three PCIE bus control units PE1, PE2, PE3.
X16 signals in the PE1 of computing unit 1 are divided into 2 independent x8 signals and for connecting Mezz card connectors
ConnectionA and Connect B.X16 signals in the PE2 of computing unit 1 are switched over by resistance can connect concentration
Processing unit supports to be integrated in the QAT data enciphering/deciphering technologies that centralized processing unit carries, or is connected to x24 physics rule
The PCIe card for standard of being arranged in pairs or groups on the PCIE Slot of lattice;X16 signals in the PE3 of computing unit 1 are divided into the x8 signals of 2 groups of independence,
Wherein 1 group of x8 is connected on the PCIE Slot of x24 physical specifications, and 1 group of x8 is connected to the PCIE Slot of x8 physical specifications in addition
On.
The x16 signals of PE1 in computing unit 2 are divided into the x4 signals of 4 groups of independence, are drawn and carried out by 4 groups of OCUlink
PCIE expanded applications.
Centralized processing unit mainly realizes the control to low speed, function simple device in system, and 2 are connected by SATA buses
Individual M.2 equipment, for depositing the (SuSE) Linux OS needed for computing system.
Centralized processing unit supports 8 SATA devices by high density OCUlink connectors, supports to set up software PLC, uses
To deposit the data needed for computing system.
Centralized processing unit is signally attached to meet on the connector Connect C of OCP specifications by 4 groups of KR, can take
Support to be integrated into the network controller in centralized processing unit with PHY cards, realized by the PHY that arranges in pairs or groups and support 4*1Gb networks electricity
Mouthful.
Management control unit is mainly used in realizing to key electronic device in computing system and whole system running state
Monitoring.
Administrative unit is designed based on the chips of Aspeed 2500, has special management port to be used for supporting consumer remote
Based on the management of standard IMPI instructions, while also provide USB and USB interface, the demand for being easy to local management.
Embodiment 2
As shown in Fig. 2 calculation type server node supports the connection of PCIE device, server node bag computing unit 1(CPU0)
With computing unit 2(CPPU1).CPU0 PE3 PCIE controllers draw 16lane signal, and wherein 8lane is arrived by pinboard
On standard PCIE x8 slot, the FHHL cards of a standard can be supported;Other 8lane PCIE signal is connected similarly to transfer
Plate, signal is connected to by Cable by two preposition 2.5 cun of NVMe storage devices by two groups of x4 OCULink connectors.
Other 1 PCIE device in system is supported that CPU1 PE1 controllers draw 16Lane PCIE signal, are divided into 4 groups by CPU1
X4 signal, it is connected to by 4 x4 OCUlink connectors by Cable in Riser expansion board, there is 1 x16 on Riser plates
The Slot of specification, standard FHFL card can be supported.
Although above-mentioned the embodiment of the present invention is described with reference to accompanying drawing, model not is protected to the present invention
The limitation enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme those skilled in the art are not
Need to pay various modifications or deformation that creative work can make still within protection scope of the present invention.
Claims (6)
1. a kind of calculation type server node, it is characterised in that include centralized processing unit, administrative unit, computing unit 1 and meter
Unit 2 is calculated, computing unit 1 carries out data communication with computing unit 2 by UPI buses;Computing unit 1 and computing unit 2 are respectively set
There are three PCIE bus control units PE1, PE2, PE3;
X16 signals in the PE1 of computing unit 1 are divided into 2 independent x8 signals and for connecting Mezz card connectors
ConnectionA and Connect B;X16 signals in the PE2 of computing unit 1 are switched over by resistance can connect concentration
Processing unit supports to be integrated in the QAT data enciphering/deciphering technologies that centralized processing unit carries, or is connected to x24 physics rule
The PCIe card for standard of being arranged in pairs or groups on the PCIE Slot of lattice;X16 signals in the PE3 of computing unit 1 are divided into the x8 signals of 2 groups of independence,
Wherein 1 group of x8 is connected on the PCIE Slot of x24 physical specifications, and 1 group of x8 is connected to the PCIE Slot of x8 physical specifications in addition
On;
The x16 signals of PE1 in computing unit 2 are divided into the x4 signals of 4 groups of independence, are drawn by 4 groups of OCUlink and carry out PCIE expansions
Exhibition application.
2. calculation type server node according to claim 1, it is characterised in that the centralized processing unit is mainly realized
Control to low speed, function simple device in system, 2 M.2 equipment are connected by SATA buses, for depositing computing system institute
The operating system needed, such as Windows, Linux or Vmware.
3. calculation type server node according to claim 1, it is characterised in that the centralized processing unit passes through highly dense
Spend OCUlink connectors and support 8 SATA devices, support to set up software PLC, for depositing the data needed for computing system.
4. calculation type server node according to claim 1, it is characterised in that the centralized processing unit passes through 4 groups
KR is signally attached to meet on the connector Connect C of OCP specifications, and the PHY cards that can arrange in pairs or groups focus on to support to be integrated into
Network controller in unit, realized according to the different PHY that arranges in pairs or groups and support 4*1Gb or 4*10Gb networks power port and optical port network.
5. calculation type server node according to claim 1, it is characterised in that the management control unit is mainly used in
Realize the monitoring to key electronic device in computing system and whole system running state.
6. calculation type server node according to claim 1, it is characterised in that the administrative unit is based on Aspeed
2500 chips are designed, and have special management port to be used for the management for supporting consumer remote to be instructed based on standard IMPI, simultaneously
USB and USB interface, the demand for being easy to local management are also provided.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710635683.1A CN107491408B (en) | 2017-07-31 | 2017-07-31 | Computing server node |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710635683.1A CN107491408B (en) | 2017-07-31 | 2017-07-31 | Computing server node |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107491408A true CN107491408A (en) | 2017-12-19 |
CN107491408B CN107491408B (en) | 2023-09-15 |
Family
ID=60644908
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710635683.1A Active CN107491408B (en) | 2017-07-31 | 2017-07-31 | Computing server node |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107491408B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108287803A (en) * | 2018-01-22 | 2018-07-17 | 郑州云海信息技术有限公司 | A kind of PCIE expanding units of OCP interfaces |
CN111752346A (en) * | 2019-06-24 | 2020-10-09 | 北京京东尚科信息技术有限公司 | Server based on combined architecture |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140122833A1 (en) * | 2009-09-24 | 2014-05-01 | Mark Bradley Davis | Server on a chip and node cards comprising one or more of same |
US20150036681A1 (en) * | 2013-08-01 | 2015-02-05 | Advanced Micro Devices, Inc. | Pass-through routing at input/output nodes of a cluster server |
US20150120971A1 (en) * | 2013-10-25 | 2015-04-30 | Samsung Electronics Co., Ltd | Server system and storage system |
US20150186319A1 (en) * | 2013-12-26 | 2015-07-02 | Dirk F. Blevins | Computer architecture to provide flexibility and/or scalability |
CN204557308U (en) * | 2015-04-28 | 2015-08-12 | 浪潮电子信息产业股份有限公司 | A kind of high density Novel cutter flap-type server based on fusion architecture |
CN204650354U (en) * | 2015-04-23 | 2015-09-16 | 浪潮电子信息产业股份有限公司 | A kind of blade server framework supporting PCIe expansion based on fusion architecture |
CN204705946U (en) * | 2015-06-25 | 2015-10-14 | 浪潮电子信息产业股份有限公司 | A kind of novel 4 road server systems based on fusion architecture |
CN204833236U (en) * | 2015-08-21 | 2015-12-02 | 浪潮电子信息产业股份有限公司 | Support memory system of hybrid storage |
CN204945920U (en) * | 2015-09-28 | 2016-01-06 | 浪潮电子信息产业股份有限公司 | The IO expansion board of a kind of compatibility dissimilar computing node PCIE hot plug |
CN206162507U (en) * | 2016-11-07 | 2017-05-10 | 郑州云海信息技术有限公司 | Switching mechanism of PCIE bus |
CN207337393U (en) * | 2017-07-31 | 2018-05-08 | 郑州云海信息技术有限公司 | A kind of calculation type server node |
-
2017
- 2017-07-31 CN CN201710635683.1A patent/CN107491408B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140122833A1 (en) * | 2009-09-24 | 2014-05-01 | Mark Bradley Davis | Server on a chip and node cards comprising one or more of same |
US20150036681A1 (en) * | 2013-08-01 | 2015-02-05 | Advanced Micro Devices, Inc. | Pass-through routing at input/output nodes of a cluster server |
US20150120971A1 (en) * | 2013-10-25 | 2015-04-30 | Samsung Electronics Co., Ltd | Server system and storage system |
US20150186319A1 (en) * | 2013-12-26 | 2015-07-02 | Dirk F. Blevins | Computer architecture to provide flexibility and/or scalability |
CN204650354U (en) * | 2015-04-23 | 2015-09-16 | 浪潮电子信息产业股份有限公司 | A kind of blade server framework supporting PCIe expansion based on fusion architecture |
CN204557308U (en) * | 2015-04-28 | 2015-08-12 | 浪潮电子信息产业股份有限公司 | A kind of high density Novel cutter flap-type server based on fusion architecture |
CN204705946U (en) * | 2015-06-25 | 2015-10-14 | 浪潮电子信息产业股份有限公司 | A kind of novel 4 road server systems based on fusion architecture |
CN204833236U (en) * | 2015-08-21 | 2015-12-02 | 浪潮电子信息产业股份有限公司 | Support memory system of hybrid storage |
CN204945920U (en) * | 2015-09-28 | 2016-01-06 | 浪潮电子信息产业股份有限公司 | The IO expansion board of a kind of compatibility dissimilar computing node PCIE hot plug |
CN206162507U (en) * | 2016-11-07 | 2017-05-10 | 郑州云海信息技术有限公司 | Switching mechanism of PCIE bus |
CN207337393U (en) * | 2017-07-31 | 2018-05-08 | 郑州云海信息技术有限公司 | A kind of calculation type server node |
Non-Patent Citations (3)
Title |
---|
徐海琛,张献华: "高性能计算技术与大容量存储技术概述" * |
王巍;: "基于龙芯3A处理器的加固服务器设计与实现" * |
石峰;吴建飞;刘凯;徐欣;: "基于Xilinx FPGA的PCIE接口实现" * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108287803A (en) * | 2018-01-22 | 2018-07-17 | 郑州云海信息技术有限公司 | A kind of PCIE expanding units of OCP interfaces |
CN111752346A (en) * | 2019-06-24 | 2020-10-09 | 北京京东尚科信息技术有限公司 | Server based on combined architecture |
Also Published As
Publication number | Publication date |
---|---|
CN107491408B (en) | 2023-09-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN204833236U (en) | Support memory system of hybrid storage | |
CN104135514B (en) | Fusion type virtual storage system | |
CN102929363B (en) | A kind of method for designing of high-density blade server | |
CN108090014A (en) | The storage IO casees system and its design method of a kind of compatible NVMe | |
CN107220195A (en) | A kind of multichannel computer system architecture for supporting high density NVMe to store | |
CN107480094A (en) | A kind of pond server system architecture of fusion architecture | |
CN104408014A (en) | System and method for interconnecting processing units of calculation systems | |
CN107491408A (en) | A kind of calculation type server node | |
CN204705946U (en) | A kind of novel 4 road server systems based on fusion architecture | |
CN105224496B (en) | A kind of flow for dynamic reconfigurable system | |
CN104580527B (en) | A kind of more I/O high density multi-node server system design methods of cloud service-oriented device application | |
CN109033009A (en) | It is a kind of to support general and machine cabinet type server circuit board and system | |
CN202383569U (en) | Mainboard with multifunctional extensible peripheral component interconnect express (PCIE) interface device | |
CN209248436U (en) | A kind of expansion board clamping and server | |
CN207337393U (en) | A kind of calculation type server node | |
CN204557308U (en) | A kind of high density Novel cutter flap-type server based on fusion architecture | |
CN105159605A (en) | Storage server architecture | |
CN105511990B (en) | Device based on fusion architecture dual redundant degree storage control node framework | |
US11055252B1 (en) | Modular hardware acceleration device | |
CN102799708B (en) | Graphic processing unit (GPU) high-performance calculation platform device applied to electromagnetic simulation | |
CN104932837A (en) | Storage pool framework | |
CN107391428B (en) | Novel framework four-way server | |
CN106161169A (en) | A kind of multi-host network exchange system | |
CN105095143A (en) | Server node and complete machine cabinet server | |
CN209248518U (en) | A kind of solid state hard disk expansion board clamping and server |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |