CN105338433B - A kind of method of parallel computation broadcast communication - Google Patents
A kind of method of parallel computation broadcast communication Download PDFInfo
- Publication number
- CN105338433B CN105338433B CN201510671311.5A CN201510671311A CN105338433B CN 105338433 B CN105338433 B CN 105338433B CN 201510671311 A CN201510671311 A CN 201510671311A CN 105338433 B CN105338433 B CN 105338433B
- Authority
- CN
- China
- Prior art keywords
- cabinet
- broadcast
- optical fiber
- free light
- interchanger
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0003—Details
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/11—Arrangements specific to free-space transmission, i.e. transmission through air or vacuum
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Signal Processing (AREA)
- Optical Communication System (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of methods of parallel computation broadcast communication, the node in need for sending broadcast data transmit data to column head interchanger, convert free space optical for the signal in optical fiber by placing free light-emitting device on the cabinet top of column head interchanger；It places that free light is divided and receiving device is divided and is received on the cabinet for receiving data, the broadcast communication between cabinet is divided by free light and receiving device by free space optical sends all receiving ends in physical layer；Broadcast communication between interior of equipment cabinet server is then transmitted by wire link.The present invention has many advantages, such as to can be effectively reduced broadcast delay, improves Parallel Computing Performance.
Description
Technical field
Present invention relates generally to the field of broadcast communication in massive parallel processing, and it is wide to refer in particular to a kind of parallel computation
The method for broadcasting communication.
Background technique
Massive parallel processing needs height in computing capabilitys such as scientific algorithm, weather forecast, medical novel drugs researchs
Field, country economic development and strategic capability promoted on have vital status and effect.High-performance computer
Computing capability is constantly promoted by technological innovation, typical framework is to cooperate with to carry out parallel computation by computing cluster.
The parallel programming models such as MPI (Message Passing Interface) realize section by divide task
Cooperated computing between point.Wherein common broadcast, fence the one-to-all and all-to-one needs in operation such as synchronize and will disappear
Breath is distributed to all nodes of the whole network and converges.High-performance calculation is extremely sensitive to the delay of message transmission, reduces the complete of broadcast
Calculating speed can be effectively promoted at the time.The BlueGene series super computers of typical IBM uses independent tree type network
Network is for accelerating broadcast communication.
Summary of the invention
The technical problem to be solved in the present invention is that, for technical problem of the existing technology, the present invention provides one
The method that kind can be effectively reduced broadcast delay, the parallel computation broadcast communication for improving Parallel Computing Performance
In order to solve the above technical problems, the invention adopts the following technical scheme:
A kind of method of parallel computation broadcast communication, it is in need send broadcast data node transmit data to column head
Interchanger converts free sky for the signal in optical fiber by placing free light-emitting device on the cabinet top of column head interchanger
Between light；Free light light splitting is placed on the cabinet for receiving data and receiving device is divided and is received, so that between cabinet
Broadcast communication can be divided by free light and receiving device by free space optical sends all receiving ends in physical layer；Cabinet
Broadcast communication between internal server is then transmitted by wire link.
As a further improvement of the present invention: the free light-emitting device is optical fiber collimator.
As a further improvement of the present invention: the optical fiber collimator is one-way optical fiber collimator, for sending out through optical module
It penetrates end and signal that optical fiber transmits is converted into the free space optical in a direction.
As a further improvement of the present invention: the optical fiber collimator is bidirectional optical fiber collimator, for sending out through optical module
Penetrate the free space optical that the signal that end, optical fiber and optical fiber splitter transmit is converted into both direction.
As a further improvement of the present invention: when the method completes the whole network broadcast by four interchangers, it is assumed that network
Scale is that m arranges cabinet, and every row has n cabinet, the specific steps are as follows:
S100:(0,0) free light unidirectional emission equipment is both horizontally and vertically placed respectively on cabinet top, be used for
Broadcast data is sent to the cabinet of same row and same row；Horizontal direction places free light light splitting on (0, j) 1≤j≤n cabinet top
And receiving device is for receiving data, and free light unidirectional emission equipment is placed in vertical direction, is used for broadcast data same
Column are forwarded；Free light light splitting and receiving device are placed on (i, j) 1≤i≤m, 1≤j≤n cabinet top, it is vertical for receiving
The broadcast data that direction is transmitted across；
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400: the cabinet top interchanger of cabinet (i, 0) is handed over by the cabinet top that wired mode transmits data to (0,0) cabinet
It changes planes；
S500:(0,0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet,
Each cabinet is divided by free light and receiving device, receives the broadcast data of (0,0) cabinet transmission；
S600:(0, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
As a further improvement of the present invention: the method is completed the whole network by interchanger three times and is broadcasted, it is assumed that network rule
Mould is that m arranges cabinet, and every row has n cabinet, the specific steps are as follows:
S100:(i, 0) the two-way transmitting equipment of free light is placed below in a position (i, i) of cabinet top in 1≤i≤m cabinet,
Free light unidirectional emission equipment, (k, k) 0≤k≤m of cabinet top is placed above, k ≠ i position is placed free light light splitting and connect
Receiving unit；The two-way transmitting equipment of free light is placed below in a position (i, i) in (i, j) 0≤i≤m, 1≤j≤n cabinet cabinet top,
Be placed above the light splitting of free light and receiving device, (k, k) 0≤k≤m of cabinet top, k ≠ i position place free light be divided and
Receiving device；
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400:(i, 0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet,
Each cabinet is divided by free light and receiving device, receives the broadcast data of (i, 0) cabinet transmission；
S500:(i, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
Compared with the prior art, the advantages of the present invention are as follows:
1, the method for parallel computation broadcast communication of the present invention, is broadcasted using free light, and free light is compared with optical fiber, certainly
There is natural broadcast characteristic by mpsi layer, broadcast communication application can be carried.Transmission has the characteristics that low delay, using certainly
Broadcast delay can be effectively reduced by the method that mpsi layer carries out data broadcasting, improve the performance of parallel computation.
2, the method for parallel computation broadcast communication of the present invention, using free light broadcasting network and other unicast traffics every
From reduction network traffic load.
3, the method for parallel computation broadcast communication of the present invention can be effectively reduced the hardware capabilities such as switch network arbitration
Complexity reduces the capacity requirement of forwarding table.
4, the method for parallel computation broadcast communication of the present invention carries out broadcast communication using free mpsi layer, i.e., with wireless
Link replaces wire link, wiring complexity can be reduced, so that system maintenance is simple.
Detailed description of the invention
Fig. 1 is the schematic diagram that the present invention constructs broadcast system in specific application example.
Fig. 2 is the schematic diagram that the present invention uses free light unidirectional emission equipment in specific application example.
Fig. 3 is the schematic diagram that the present invention uses the light splitting of free light and receiving device in specific application example.
Fig. 4 is the schematic diagram that the present invention uses the two-way transmitting equipment of free light in specific application example.
Fig. 5 is the schematic illustration that the present invention realizes the whole network broadcasting method by four interchangers in embodiment 1.
Fig. 6 is the schematic illustration that the present invention realizes the whole network broadcasting method by interchanger three times in example 2.
Specific embodiment
The present invention is described in further details below with reference to Figure of description and specific embodiment.
The method of parallel computation broadcast communication of the present invention, it is logical to optimize parallel computation broadcast based on free space mpsi layer
The method of letter；In the method for the invention, it is in need send broadcast data node transmit data to column head interchanger, lead to
It crosses on the cabinet top in column head interchanger and places optical fiber collimator 1 and by the signal in optical fiber convert free space optical A；It is receiving
Wave beam splitter 2 is placed on the cabinet of data to be divided, and enables the broadcast communication between cabinet will be certainly by wave beam splitter 2
All receiving ends are sent in physical layer by spatial light A；Broadcast communication between interior of equipment cabinet server then passes through wire link
Transmission.
Specific steps are as follows:
S1: building broadcast system；Referring to Fig. 1, free light unidirectional emission equipment is placed on the cabinet top of column head interchanger
(such as: optical fiber collimator 1) places free light light splitting and receiving device (such as: wave beam splitter 2) on the cabinet for receiving data；Its
In, the signal that optical fiber collimator 1 can be used to transmit through optical module transmitting terminal 3 and optical fiber is converted into the free space in a direction
Light A (as shown in Figure 2), or two are converted into for the signal transmitted through optical module transmitting terminal 3, optical fiber and optical fiber splitter 4
The free space optical A (as shown in Figure 3) in direction.Free space optical A can be divided (as shown in Figure 4) by wave beam splitter 2, be made
The broadcast communication obtained between cabinet can send all receiving ends, i.e. light in physical layer for free space optical by wave beam splitter
Module receiving end 5.
S2: the broadcast system constructed by step S1 carries out broadcast communication using free space optical A.
Embodiment 1: being that the whole network broadcast (vacation is completed by four interchangers referring to Fig. 5 in a specific application example
If network size is that m arranges cabinet, every row has n cabinet), detailed step is as follows:
S100:(0,0) free light unidirectional emission equipment is both horizontally and vertically placed respectively on cabinet top, be used for
Broadcast data is sent to the cabinet of same row and same row.Horizontal direction places free light light splitting on (0, j) 1≤j≤n cabinet top
And receiving device is for receiving data, and free light unidirectional emission equipment is placed in vertical direction, is used for broadcast data same
Column are forwarded.Free light light splitting and receiving device are placed on (i, j) 1≤i≤m, 1≤j≤n cabinet top, it is vertical for receiving
The broadcast data that direction is transmitted across.
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400: the cabinet top interchanger of cabinet (i, 0) is handed over by the cabinet top that wired mode transmits data to (0,0) cabinet
It changes planes；
S500:(0,0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet,
Each cabinet is divided by free light and receiving device, receives the broadcast data of (0,0) cabinet transmission.
S600:(0, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
Embodiment 2: being that the whole network broadcast (vacation is completed by interchanger three times referring to Fig. 6 in a specific application example
If network size is that m arranges cabinet, every row has n cabinet), detailed step is as follows:
S100:(i, 0) the two-way transmitting equipment of free light is placed below in a position (i, i) of cabinet top in 1≤i≤m cabinet,
Free light unidirectional emission equipment, (k, k) 0≤k≤m of cabinet top is placed above, k ≠ i position is placed free light light splitting and connect
Receiving unit；The two-way transmitting equipment of free light is placed below in a position (i, i) in (i, j) 0≤i≤m, 1≤j≤n cabinet cabinet top,
Be placed above the light splitting of free light and receiving device, (k, k) 0≤k≤m of cabinet top, k ≠ i position place free light be divided and
Receiving device.
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400:(i, 0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet,
Each cabinet is divided by free light and receiving device, receives the broadcast data of (i, 0) cabinet transmission.
S500:(i, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
The above is only the preferred embodiment of the present invention, protection scope of the present invention is not limited merely to above-described embodiment,
All technical solutions belonged under thinking of the present invention all belong to the scope of protection of the present invention.It should be pointed out that for the art
For those of ordinary skill, several improvements and modifications without departing from the principles of the present invention should be regarded as protection of the invention
Range.
Claims (8)
1. a kind of method of parallel computation broadcast communication, which is characterized in that the node in need for sending broadcast data by data
It is sent to column head interchanger, is turned the signal in optical fiber by placing free light-emitting device on the cabinet top of column head interchanger
Turn to free space optical；Free light light splitting is placed on the cabinet for receiving data and receiving device is divided and is received, so that
Broadcast communication between cabinet can be divided by free light and receiving device send free space optical in physical layer it is all
Receiving end；Broadcast communication between interior of equipment cabinet server is then transmitted by wire link；
When the method completes the whole network broadcast by four interchangers, it is assumed that network size is that m arranges cabinet, and every row has n cabinet,
Specific step is as follows:
S100:(0,0) free light unidirectional emission equipment is both horizontally and vertically placed respectively on cabinet top, be used for same
The cabinet of one row and same row sends broadcast data；Horizontal direction is placed free light light splitting and is connect on (0, j) 1≤j≤n cabinet top
Receiving unit for receiving data, places free light unidirectional emission equipment in vertical direction, be used for by broadcast data same row into
Row forwarding；Free light light splitting and receiving device are placed on (i, j) 1≤i≤m, 1≤j≤n cabinet top, for receiving vertical direction
The broadcast data being transmitted across；
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400: the cabinet top interchanger of cabinet (i, 0) is exchanged by the cabinet top that wired mode transmits data to (0,0) cabinet
Machine；
S500:(0,0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet, each
Cabinet is divided by free light and receiving device, receives the broadcast data of (0,0) cabinet transmission；
S600:(0, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
2. the method for parallel computation broadcast communication according to claim 1, which is characterized in that the free light-emitting device
For optical fiber collimator.
3. the method for parallel computation broadcast communication according to claim 2, which is characterized in that the optical fiber collimator is single
To optical fiber collimator, the signal for transmitting through optical module transmitting terminal and optical fiber is converted into the free space optical in a direction.
4. the method for parallel computation broadcast communication according to claim 2, which is characterized in that the optical fiber collimator is double
To optical fiber collimator, the signal for transmitting through optical module transmitting terminal, optical fiber and optical fiber splitter is converted into oneself of both direction
By spatial light.
5. a kind of method of parallel computation broadcast communication, which is characterized in that the node in need for sending broadcast data by data
It is sent to column head interchanger, is turned the signal in optical fiber by placing free light-emitting device on the cabinet top of column head interchanger
Turn to free space optical；Free light light splitting is placed on the cabinet for receiving data and receiving device is divided and is received, so that
Broadcast communication between cabinet can be divided by free light and receiving device send free space optical in physical layer it is all
Receiving end；Broadcast communication between interior of equipment cabinet server is then transmitted by wire link；
The method completes the whole network broadcast by interchanger three times, it is assumed that network size is that m arranges cabinet, and every row has n cabinet, has
Steps are as follows for body:
S100:(i, 0) the two-way transmitting equipment of free light is placed below in a position (i, i) of cabinet top in 1≤i≤m cabinet, above
Free light unidirectional emission equipment, (k, k) 0≤k≤m of cabinet top are placed, k ≠ i position is placed free light light splitting and received and set
It is standby；The two-way transmitting equipment of free light is placed below in a position (i, i) in (i, j) 0≤i≤m, 1≤j≤n cabinet cabinet top, above
Free light light splitting and receiving device, (k, k) 0≤k≤m of cabinet top are placed, k ≠ i position is placed free light light splitting and received
Equipment；
S200: broadcast source node transmits data to cabinet top interchanger (i, j) by wired mode；
S300: cabinet top interchanger transmits data to the interchanger (i, 0) of same arrangement head cabinet by wired mode；
S400:(i, 0) broadcast data is sent to all cabinets of same row using free light by the cabinet top interchanger of cabinet, each
Cabinet is divided by free light and receiving device, receives the broadcast data of (i, 0) cabinet transmission；
S500:(i, j) 0≤j≤n cabinet to all cabinets of same row forwarding broadcast data, broadcast by completion the whole network.
6. the method for parallel computation broadcast communication according to claim 5, which is characterized in that the free light-emitting device
For optical fiber collimator.
7. the method for parallel computation broadcast communication according to claim 6, which is characterized in that the optical fiber collimator is single
To optical fiber collimator, the signal for transmitting through optical module transmitting terminal and optical fiber is converted into the free space optical in a direction.
8. the method for parallel computation broadcast communication according to claim 6, which is characterized in that the optical fiber collimator is double
To optical fiber collimator, the signal for transmitting through optical module transmitting terminal, optical fiber and optical fiber splitter is converted into oneself of both direction
By spatial light.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510671311.5A CN105338433B (en) | 2015-10-16 | 2015-10-16 | A kind of method of parallel computation broadcast communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510671311.5A CN105338433B (en) | 2015-10-16 | 2015-10-16 | A kind of method of parallel computation broadcast communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105338433A CN105338433A (en) | 2016-02-17 |
| CN105338433B true CN105338433B (en) | 2019-01-08 |
Family
ID=55288653
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510671311.5A Active CN105338433B (en) | 2015-10-16 | 2015-10-16 | A kind of method of parallel computation broadcast communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105338433B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1298113A (en) * | 1999-11-30 | 2001-06-06 | 王宝根 | Infrared multispectral scanning thermal imager |
| CN101487915A (en) * | 2008-01-17 | 2009-07-22 | 3M创新有限公司 | Optical apparatus holder |
| CN103441942A (en) * | 2013-08-26 | 2013-12-11 | 重庆大学 | Data center network system and data communication method based on software definition |
| CN204575907U (en) * | 2015-05-04 | 2015-08-19 | 深圳市鹏大光电技术有限公司 | A kind of forward direction optoelectric hybrid device of Erbium-Doped Fiber Amplifier (EDFA) |
| CN104904143A (en) * | 2012-11-26 | 2015-09-09 | 维斯柯科技有限公司 | Methods and systems for passive optical switching |
| CN104949770A (en) * | 2015-07-13 | 2015-09-30 | 天津津航技术物理研究所 | TDLAS (tunable diode laser absorption spectroscopy) gas temperature measurement and detection device |
-
2015
- 2015-10-16 CN CN201510671311.5A patent/CN105338433B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1298113A (en) * | 1999-11-30 | 2001-06-06 | 王宝根 | Infrared multispectral scanning thermal imager |
| CN101487915A (en) * | 2008-01-17 | 2009-07-22 | 3M创新有限公司 | Optical apparatus holder |
| CN104904143A (en) * | 2012-11-26 | 2015-09-09 | 维斯柯科技有限公司 | Methods and systems for passive optical switching |
| CN103441942A (en) * | 2013-08-26 | 2013-12-11 | 重庆大学 | Data center network system and data communication method based on software definition |
| CN204575907U (en) * | 2015-05-04 | 2015-08-19 | 深圳市鹏大光电技术有限公司 | A kind of forward direction optoelectric hybrid device of Erbium-Doped Fiber Amplifier (EDFA) |
| CN104949770A (en) * | 2015-07-13 | 2015-09-30 | 天津津航技术物理研究所 | TDLAS (tunable diode laser absorption spectroscopy) gas temperature measurement and detection device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105338433A (en) | 2016-02-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200342297A1 (en) | Tree Topology Based Computing System and Method | |
| Ding et al. | Security information transmission algorithms for IoT based on cloud computing | |
| CN103179385B (en) | A kind of easily extensible NVR system based on polycaryon processor | |
| US9331868B2 (en) | Method of bus virtualization in computing machine intercommunications | |
| CN104185999A (en) | Packet-flow interconnect fabric | |
| CN104780333A (en) | High-bandwidth video source interface adaptation device based on FPGA (Field Programmable Gate Array) | |
| CN112100265A (en) | Multi-source data processing method and device for big data architecture and block chain | |
| US11601489B2 (en) | Method for transmitting stream, streaming server and storage medium | |
| CN102387084A (en) | System structure based on Rapid IO (Input Output) protocol packet exchange | |
| CN104348913A (en) | Tight-coupling extensible big data interaction method | |
| CN108710644A (en) | One kind is about government affairs big data processing method | |
| US9575918B2 (en) | Collaboration server | |
| CN105681131B (en) | Main preparation system and its parallel output method | |
| CN103475870B (en) | A kind of distributed extendable intelligent video surveillance system | |
| CN105338433B (en) | A kind of method of parallel computation broadcast communication | |
| CN110635931B (en) | Link group configuration method and device | |
| CN201387555Y (en) | Comprehensive remote control monitoring system | |
| CN206237455U (en) | A kind of middle and primary schools' Students ' Comprehensive information management cloud platform based on big data analysis | |
| CN103227840A (en) | IOT (Internet of things)-oriented high-concurrency high-availability data acquisition system | |
| Janet et al. | Optimizing data movement within cloud environment using efficient compression techniques | |
| US10708127B1 (en) | Low-latency network switching device with latency identification and diagnostics | |
| CN203377910U (en) | Mass high-speed remote-sensing data real-time baseband processing system | |
| CN103678521A (en) | Distributed file monitoring system based on Hadoop frame | |
| CN113254723A (en) | Big data storage method and device under cloud network architecture | |
| JP5869134B2 (en) | System and method for asynchronous handshake based interconnection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |