Remote Direct Memory Access ('RDMA') In A Parallel Computer

- IBM

Remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, including: receiving an RDMA read operation request that includes a virtual address representing a memory region at which to receive data to be transferred from a second node to the first node; responsive to the RDMA read operation request: translating the virtual address to a physical address; creating a local RDMA object that includes a counter set to the size of the memory region; sending a message that includes an DMA write operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node; and receiving the data to be transferred from the second node.

Skip to: Description  ·  Claims  · Patent History  ·  Patent History

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The field of the invention is data processing, or, more specifically, methods, apparatus, and products for remote direct memory access (‘RDMA’) in a parallel computer.

2. Description of Related Art

The development of the EDVAC computer system of 1948 is often cited as the beginning of the computer era. Since that time, computer systems have evolved into extremely complicated devices. Today's computers are much more sophisticated than early systems such as the EDVAC. Computer systems typically include a combination of hardware and software components, application programs, operating systems, processors, buses, memory, input/output devices, and so on. As advances in semiconductor processing and computer architecture push the performance of the computer higher and higher, more sophisticated computer software has evolved to take advantage of the higher performance of the hardware, resulting in computer systems today that are much more powerful than just a few years ago.

In high-performance computing (HPC), high-speed communications adapters use remote data memory access (RDMA) operations to move data between the memory of a local computer and the memory of a remote computer. The high-speed adapters which perform these operations characteristically work under a software stack known as Open Fabrics Enterprise Distribution (OFED).

SUMMARY OF THE INVENTION

Remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, including: receiving, by a kernel of the first node through an application programming interface (‘API’), an RDMA read operation request that includes a virtual address representing a memory region at which to receive data to be transferred from a second node to the first node; responsive to the RDMA read operation request: translating, by the kernel of the first node, the virtual address to a physical address; creating, by the kernel of the first node, a local RDMA object that includes a counter set to the size of the memory region; sending, from a messaging unit in the first node to a messaging unit in a second node, a message that includes an DMA write operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node representing the data to be transferred from the second node to the first node; and receiving, by the first node, the data to be transferred from the second node.

The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular descriptions of exemplary embodiments of the invention as illustrated in the accompanying drawings wherein like reference numbers generally represent like parts of exemplary embodiments of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 sets forth example apparatus for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 2 sets forth a block diagram of RDMA in a parallel computer according to embodiments of the present invention.

FIG. 3A sets forth a block diagram of an example Point-To-Point Adapter useful in systems for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 3B sets forth a block diagram of an example Global Combining Network Adapter useful in systems for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 4 sets forth a line drawing illustrating an example data communications network optimized for point-to-point operations useful in systems capable of RDMA in a parallel computer according to embodiments of the present invention.

FIG. 5 sets forth a line drawing illustrating an example global combining network useful in systems capable of RDMA in a parallel computer according to embodiments of the present invention.

FIG. 6 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 7 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 8 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 9 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 10 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

FIG. 11 sets forth a flow chart illustrating an example method for RDMA in a parallel computer according to embodiments of the present invention.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

Exemplary methods, apparatus, and products for RDMA in a parallel computer in accordance with the present invention are described with reference to the accompanying drawings, beginning with FIG. 1. FIG. 1 sets forth example apparatus for RDMA in a parallel computer according to embodiments of the present invention. The apparatus of FIG. 1 includes a parallel computer (100), non-volatile memory for the computer in the form of a data storage device (118), an output device for the computer in the form of a printer (120), and an input/output device for the computer in the form of a computer terminal (122). The parallel computer (100) in the example of FIG. 1 includes a plurality of compute nodes (102). The compute nodes (102) are coupled for data communications by several independent data communications networks including a high speed Ethernet network (174), a Joint Test Action Group (‘JTAG’) network (104), a global combining network (106) which is optimized for collective operations using a binary tree network topology, and a point-to-point network (108), which is optimized for point-to-point operations using a torus network topology. The global combining network (106) is a data communications network that includes data communications links connected to the compute nodes (102) so as to organize the compute nodes (102) as a binary tree. Each data communications network is implemented with data communications links among the compute nodes (102). The data communications links provide data communications for parallel operations among the compute nodes (102) of the parallel computer (100).

The compute nodes (102) of the parallel computer (100) are organized into at least one operational group (132) of compute nodes for collective parallel operations on the parallel computer (100). Each operational group (132) of compute nodes is the set of compute nodes upon which a collective parallel operation executes. Each compute node in the operational group (132) is assigned a unique rank that identifies the particular compute node in the operational group (132). Collective operations are implemented with data communications among the compute nodes of an operational group. Collective operations are those functions that involve all the compute nodes of an operational group (132). A collective operation is an operation, a message-passing computer program instruction that is executed simultaneously, that is, at approximately the same time, by all the compute nodes in an operational group (132) of compute nodes. Such an operational group (132) may include all the compute nodes (102) in a parallel computer (100) or a subset all the compute nodes (102). Collective operations are often built around point-to-point operations. A collective operation requires that all processes on all compute nodes within an operational group (132) call the same collective operation with matching arguments. A ‘broadcast’ is an example of a collective operation for moving data among compute nodes of an operational group. A ‘reduce’ operation is an example of a collective operation that executes arithmetic or logical functions on data distributed among the compute nodes of an operational group (132). An operational group (132) may be implemented as, for example, an MPI ‘communicator.’

‘MPI’ refers to ‘Message Passing Interface,’ a prior art parallel communications library, a module of computer program instructions for data communications on parallel computers. Examples of prior-art parallel communications libraries that may be improved for RDMA in a parallel computer according to embodiments of the present invention include MPI and the ‘Parallel Virtual Machine’ (‘PVM’) library. PVM was developed by the University of Tennessee, The Oak Ridge National Laboratory and Emory University. MPI is promulgated by the MPI Forum, an open group with representatives from many organizations that define and maintain the MPI standard. MPI at the time of this writing is a de facto standard for communication among compute nodes running a parallel program on a distributed memory parallel computer. This specification sometimes uses MPI terminology for ease of explanation, although the use of MPI as such is not a requirement or limitation of the present invention.

Some collective operations have a single originating or receiving process running on a particular compute node in an operational group (132). For example, in a ‘broadcast’ collective operation, the process on the compute node that distributes the data to all the other compute nodes is an originating process. In a ‘gather’ operation, for example, the process on the compute node that received all the data from the other compute nodes is a receiving process. The compute node on which such an originating or receiving process runs is referred to as a logical root.

Most collective operations are variations or combinations of four basic operations: broadcast, gather, scatter, and reduce. The interfaces for these collective operations are defined in the MPI standards promulgated by the MPI Forum. Algorithms for executing collective operations, however, are not defined in the MPI standards. In a broadcast operation, all processes specify the same root process, whose buffer contents will be sent. Processes other than the root specify receive buffers. After the operation, all buffers contain the message from the root process.

A scatter operation, like the broadcast operation, is also a one-to-many collective operation. In a scatter operation, the logical root divides data on the root into segments and distributes a different segment to each compute node in the operational group (132). In scatter operation, all processes typically specify the same receive count. The send arguments are only significant to the root process, whose buffer actually contains sendcount*N elements of a given datatype, where N is the number of processes in the given group of compute nodes. The send buffer is divided and dispersed to all processes (including the process on the logical root). Each compute node is assigned a sequential identifier termed a ‘rank.’ After the operation, the root has sent sendcount data elements to each process in increasing rank order. Rank 0 receives the first sendcount data elements from the send buffer. Rank 1 receives the second sendcount data elements from the send buffer, and so on.

A gather operation is a many-to-one collective operation that is a complete reverse of the description of the scatter operation. That is, a gather is a many-to-one collective operation in which elements of a datatype are gathered from the ranked compute nodes into a receive buffer in a root node.

A reduction operation is also a many-to-one collective operation that includes an arithmetic or logical function performed on two data elements. All processes specify the same ‘count’ and the same arithmetic or logical function. After the reduction, all processes have sent count data elements from compute node send buffers to the root process. In a reduction operation, data elements from corresponding send buffer locations are combined pair-wise by arithmetic or logical operations to yield a single corresponding element in the root process' receive buffer. Application specific reduction operations can be defined at runtime. Parallel communications libraries may support predefined operations. MPI, for example, provides the following pre-defined reduction operations:

MPI_MAX maximum MPI_MIN minimum MPI_SUM sum MPI_PROD product MPI_LAND logical and MPI_BAND bitwise and MPI_LOR logical or MPI_BOR bitwise or MPI_LXOR logical exclusive or MPI_BXOR bitwise exclusive or

In addition to compute nodes, the parallel computer (100) includes input/output (‘I/O’) nodes (110, 114) coupled to compute nodes (102) through the global combining network (106). The compute nodes (102) in the parallel computer (100) may be partitioned into processing sets such that each compute node in a processing set is connected for data communications to the same I/O node. Each processing set, therefore, is composed of one I/O node and a subset of compute nodes (102). The ratio between the number of compute nodes to the number of I/O nodes in the entire system typically depends on the hardware configuration for the parallel computer (102). For example, in some configurations, each processing set may be composed of eight compute nodes and one I/O node. In some other configurations, each processing set may be composed of sixty-four compute nodes and one I/O node. Such example are for explanation only, however, and not for limitation. Each I/O node provides I/O services between compute nodes (102) of its processing set and a set of I/O devices. In the example of FIG. 1, the I/O nodes (110, 114) are connected for data communications I/O devices (118, 120, 122) through local area network (‘LAN’) (130) implemented using high-speed Ethernet.

The parallel computer (100) of FIG. 1 also includes a service node (116) coupled to the compute nodes through one of the networks (104). Service node (116) provides services common to pluralities of compute nodes, administering the configuration of compute nodes, loading programs into the compute nodes, starting program execution on the compute nodes, retrieving results of program operations on the compute nodes, and so on. Service node (116) runs a service application (124) and communicates with users (128) through a service application interface (126) that runs on computer terminal (122).

The parallel computer (100) of FIG. 1 operates generally for RDMA in a parallel computer. The parallel computer (100) of FIG. 1 includes a plurality of nodes, such as the first node (102a) and the second node (102b), that may be embodied as a compute node, an I/O node, or other module of automated computing machinery that forms a parallel computer (100). In the example of FIG. 1, each node (102a, 102b) includes a messaging unit (101a, 101b). The messaging units (101a, 101b) of FIG. 1 may be embodied, for example, as a communications adapter for point-to-point data communications with other nodes in the parallel computer (100), as a library of data communications operations for carrying out data communications with other nodes in the parallel computer (100), or any combination thereof. In the example of FIG. 1, the messaging units (101a, 101b) may carry out data communications with other nodes in the parallel computer (100), for example, over a torus network as described with reference to FIGS. 1-4.

The parallel computer (100) of FIG. 1 carries out RDMA in a parallel computer by receiving, by a kernel (103a) of the first node (102a) through an application programming interface (‘API’) (109a), an RDMA read operation request that includes a virtual address representing a memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a). The parallel computer (100) of FIG. 1 further carries out RDMA in a parallel computer by translating, by the kernel (103a) of the first node (102a), the virtual address to a physical address and by creating a local RDMA object (107a) that includes a counter set to the size of the memory (105a) region. The parallel computer (100) of FIG. 1 further carries out RDMA in a parallel computer by sending, from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA write operation request, the physical address of the memory (105a) region on the first node (102a), the physical address of the local RDMA object (107a) on the first node (102a), and a remote virtual address on the second node (102b) representing the data to be transferred from the second node (102b) to the first node (102a). The parallel computer (100) of FIG. 1 further carries out RDMA in a parallel computer by receiving, by the first node (102a), the data to be transferred from the second node (102b).

The parallel computer (100) of FIG. 1 can also carry out RDMA in a parallel computer by receiving, by a kernel (103a) of a first node (102a) through an API (109a), an RDMA write operation request that includes a virtual address representing a memory (105a) region in the first node (102a) from which to send data to a second node (102b). The parallel computer (100) of FIG. 1 can further carry out RDMA in a parallel computer by translating, by the kernel (103a) of the first node (102a), the virtual address to a physical address and creating a local RDMA object (107b) that includes a counter set to the size of the memory (105a) region in the first node (102a) from which to send data to a second node (102b). The parallel computer (100) of FIG. 1 can further carry out RDMA in a parallel computer by sending, from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA read operation request, the physical address of the memory region on the first node (102a), the physical address of the local RDMA object on the first node (102a), and a remote virtual address on the second node (102b) representing a memory region in the second node (102b) for receiving data from the first node (102a). The parallel computer (100) of FIG. 1 can further carry out RDMA in a parallel computer by sending, from the first node (102a), the data to be received by the second node (102b).

The arrangement of nodes, networks, and I/O devices making up the example apparatus illustrated in FIG. 1 are for explanation only, not for limitation of the present invention. Apparatus capable of RDMA in a parallel computer according to embodiments of the present invention may include additional nodes, networks, devices, and architectures, not shown in FIG. 1, as will occur to those of skill in the art. The parallel computer (100) in the example of FIG. 1 includes sixteen compute nodes (102); parallel computers capable of RDMA in a parallel computer according to embodiments of the present invention sometimes include thousands of compute nodes. In addition to Ethernet (174) and JTAG (104), networks in such data processing systems may support many data communications protocols including for example TCP (Transmission Control Protocol), IP (Internet Protocol), and others as will occur to those of skill in the art. Various embodiments of the present invention may be implemented on a variety of hardware platforms in addition to those illustrated in FIG. 1.

RDMA in a parallel computer according to embodiments of the present invention is generally implemented on a parallel computer that includes a plurality of compute nodes organized for collective operations through at least one data communications network. In fact, such computers may include thousands of such compute nodes. Each compute node is in turn itself a kind of computer composed of one or more computer processing cores, its own computer memory, and its own input/output adapters. For further explanation, therefore, FIG. 2 sets forth a block diagram of an example compute node (102) useful in a parallel computer capable of RDMA in a parallel computer according to embodiments of the present invention. The compute node (102) of FIG. 2 includes a plurality of processing cores (165) as well as RAM (156). The processing cores (165) of FIG. 2 may be configured on one or more integrated circuit dies. Processing cores (165) are connected to RAM (156) through a high-speed memory bus (155) and through a bus adapter (194) and an extension bus (168) to other components of the compute node. Stored in RAM (156) is an application program (159), a module of computer program instructions that carries out parallel, user-level data processing using parallel algorithms.

Also stored RAM (156) is a parallel communications library (161), a library of computer program instructions that carry out parallel communications among compute nodes, including point-to-point operations as well as collective operations. A library of parallel communications routines may be developed from scratch for use in systems according to embodiments of the present invention, using a traditional programming language such as the C programming language, and using traditional programming methods to write parallel communications routines that send and receive data among nodes on two independent data communications networks. Alternatively, existing prior art libraries may be improved to operate according to embodiments of the present invention. Examples of prior-art parallel communications libraries include the ‘Message Passing Interface’ (‘MPI’) library and the ‘Parallel Virtual Machine’ (‘PVM’) library.

Also stored in RAM (156) is an operating system (162), a module of computer program instructions and routines for an application program's access to other resources of the compute node. It is typical for an application program and parallel communications library in a compute node of a parallel computer to run a single thread of execution with no user login and no security issues because the thread is entitled to complete access to all resources of the node. The quantity and complexity of tasks to be performed by an operating system on a compute node in a parallel computer therefore are smaller and less complex than those of an operating system on a serial computer with many threads running simultaneously. In addition, there is no video I/O on the compute node (102) of FIG. 2, another factor that decreases the demands on the operating system. The operating system (162) may therefore be quite lightweight by comparison with operating systems of general purpose computers, a pared down version as it were, or an operating system developed specifically for operations on a particular parallel computer. Operating systems that may usefully be improved, simplified, for use in a compute node include UNIX™, Linux™, Windows XP™, AIX™, IBM's i5/OS™, and others as will occur to those of skill in the art.

The example compute node (102) of FIG. 2 includes several communications adapters (172, 176, 180, 188) for implementing data communications with other nodes of a parallel computer. Such data communications may be carried out serially through RS-232 connections, through external buses such as USB, through data communications networks such as IP networks, and in other ways as will occur to those of skill in the art. Communications adapters implement the hardware level of data communications through which one computer sends data communications to another computer, directly or through a network. Examples of communications adapters useful in apparatus useful for RDMA in a parallel computer include modems for wired communications, Ethernet (IEEE 802.3) adapters for wired network communications, and 802.11b adapters for wireless network communications.

The data communications adapters in the example of FIG. 2 include a Gigabit Ethernet adapter (172) that couples example compute node (102) for data communications to a Gigabit Ethernet (174). Gigabit Ethernet is a network transmission standard, defined in the IEEE 802.3 standard, that provides a data rate of 1 billion bits per second (one gigabit). Gigabit Ethernet is a variant of Ethernet that operates over multimode fiber optic cable, single mode fiber optic cable, or unshielded twisted pair.

The data communications adapters in the example of FIG. 2 include a JTAG Slave circuit (176) that couples example compute node (102) for data communications to a JTAG Master circuit (178). JTAG is the usual name used for the IEEE 1149.1 standard entitled Standard Test Access Port and Boundary-Scan Architecture for test access ports used for testing printed circuit boards using boundary scan. JTAG is so widely adapted that, at this time, boundary scan is more or less synonymous with JTAG. JTAG is used not only for printed circuit boards, but also for conducting boundary scans of integrated circuits, and is also useful as a mechanism for debugging embedded systems, providing a convenient “back door” into the system. The example compute node of FIG. 2 may be all three of these: It typically includes one or more integrated circuits installed on a printed circuit board and may be implemented as an embedded system having its own processing core, its own memory, and its own I/O capability. JTAG boundary scans through JTAG Slave (176) may efficiently configure processing core registers and memory in compute node (102) for use in dynamically reassigning a connected node to a block of compute nodes useful in systems for RDMA in a parallel computer according to embodiments of the present invention.

The data communications adapters in the example of FIG. 2 include a Point-To-Point Network Adapter (180) that couples example compute node (102) for data communications to a network (108) that is optimal for point-to-point message passing operations such as, for example, a network configured as a three-dimensional torus or mesh. The Point-To-Point Adapter (180) provides data communications in six directions on three communications axes, x, y, and z, through six bidirectional links: +x (181), −x (182), +y (183), −y (184), +z (185), and −z (186).

The data communications adapters in the example of FIG. 2 include a Global Combining Network Adapter (188) that couples example compute node (102) for data communications to a global combining network (106) that is optimal for collective message passing operations such as, for example, a network configured as a binary tree. The Global Combining Network Adapter (188) provides data communications through three bidirectional links for each global combining network (106) that the Global Combining Network Adapter (188) supports. In the example of FIG. 2, the Global Combining Network Adapter (188) provides data communications through three bidirectional links for global combining network (106): two to children nodes (190) and one to a parent node (192).

The example compute node (102) includes multiple arithmetic logic units (‘ALUs’). Each processing core (165) includes an ALU (166), and a separate ALU (170) is dedicated to the exclusive use of the Global Combining Network Adapter (188) for use in performing the arithmetic and logical functions of reduction operations, including an allreduce operation. Computer program instructions of a reduction routine in a parallel communications library (161) may latch an instruction for an arithmetic or logical function into an instruction register (169). When the arithmetic or logical function of a reduction operation is a ‘sum’ or a ‘logical OR,’ for example, the collective operations adapter (188) may execute the arithmetic or logical operation by use of the ALU (166) in the processing core (165) or, typically much faster, by use of the dedicated ALU (170) using data provided by the nodes (190, 192) on the global combining network (106) and data provided by processing cores (165) on the compute node (102).

Often when performing arithmetic operations in the global combining network adapter (188), however, the global combining network adapter (188) only serves to combine data received from the children nodes (190) and pass the result up the network (106) to the parent node (192). Similarly, the global combining network adapter (188) may only serve to transmit data received from the parent node (192) and pass the data down the network (106) to the children nodes (190). That is, none of the processing cores (165) on the compute node (102) contribute data that alters the output of ALU (170), which is then passed up or down the global combining network (106). Because the ALU (170) typically does not output any data onto the network (106) until the ALU (170) receives input from one of the processing cores (165), a processing core (165) may inject the identity element into the dedicated ALU (170) for the particular arithmetic operation being perform in the ALU (170) in order to prevent alteration of the output of the ALU (170). Injecting the identity element into the ALU, however, often consumes numerous processing cycles. To further enhance performance in such cases, the example compute node (102) includes dedicated hardware (171) for injecting identity elements into the ALU (170) to reduce the amount of processing core resources required to prevent alteration of the ALU output. The dedicated hardware (171) injects an identity element that corresponds to the particular arithmetic operation performed by the ALU. For example, when the global combining network adapter (188) performs a bitwise OR on the data received from the children nodes (190), dedicated hardware (171) may inject zeros into the ALU (170) to improve performance throughout the global combining network (106).

In the example of FIG. 2, the compute node (102) may utilize message unit (‘MU’) hardware for I/O data transport across I/O links and, for flexible I/O configurations, across an I/O torus. An I/O software architecture may specify a network layer on which I/O services are built. The network layer components may be modeled after the Open Fabrics Remote Direct Memory Access (‘RDMA’) framework or OpenFabrics Enterprise Distribution (‘OFED’) framework, an organization of companies and individuals providing open source software in the high-performance-computing (‘HPC’) arena. As such, internal network interfaces may be modeled after the OFED interfaces and processes running in the I/O node environment may communicate over I/O links using standard OFED RDMA verbs.

For further explanation, FIG. 3A sets forth a block diagram of an example Point-To-Point Adapter (180) useful in systems for RDMA in a parallel computer according to embodiments of the present invention. The Point-To-Point Adapter (180) is designed for use in a data communications network optimized for point-to-point operations, a network that organizes compute nodes in a three-dimensional torus or mesh. The Point-To-Point Adapter (180) in the example of FIG. 3A provides data communication along an x-axis through four unidirectional data communications links, to and from the next node in the −x direction (182) and to and from the next node in the +x direction (181). The Point-To-Point Adapter (180) of FIG. 3A also provides data communication along a y-axis through four unidirectional data communications links, to and from the next node in the −y direction (184) and to and from the next node in the +y direction (183). The Point-To-Point Adapter (180) of FIG. 3A also provides data communication along a z-axis through four unidirectional data communications links, to and from the next node in the −z direction (186) and to and from the next node in the +z direction (185).

For further explanation, FIG. 3B sets forth a block diagram of an example Global Combining Network Adapter (188) useful in systems for RDMA in a parallel computer according to embodiments of the present invention. The Global Combining Network Adapter (188) is designed for use in a network optimized for collective operations, a network that organizes compute nodes of a parallel computer in a binary tree. The Global Combining Network Adapter (188) in the example of FIG. 3B provides data communication to and from children nodes of a global combining network through four unidirectional data communications links (190), and also provides data communication to and from a parent node of the global combining network through two unidirectional data communications links (192).

For further explanation, FIG. 4 sets forth a line drawing illustrating an example data communications network (108) optimized for point-to-point operations useful in systems capable of RDMA in a parallel computer according to embodiments of the present invention. In the example of FIG. 4, dots represent compute nodes (102) of a parallel computer, and the dotted lines between the dots represent data communications links (103) between compute nodes. The data communications links are implemented with point-to-point data communications adapters similar to the one illustrated for example in FIG. 3A, with data communications links on three axis, x, y, and z, and to and fro in six directions +x (181), −x (182), +y (183), −y (184), +z (185), and −z (186). The links and compute nodes are organized by this data communications network optimized for point-to-point operations into a three dimensional mesh (105). The mesh (105) has wrap-around links on each axis that connect the outermost compute nodes in the mesh (105) on opposite sides of the mesh (105). These wrap-around links form a torus (107). Each compute node in the torus has a location in the torus that is uniquely specified by a set of x, y, z coordinates. Readers will note that the wrap-around links in the y and z directions have been omitted for clarity, but are configured in a similar manner to the wrap-around link illustrated in the x direction. For clarity of explanation, the data communications network of FIG. 4 is illustrated with only 27 compute nodes, but readers will recognize that a data communications network optimized for point-to-point operations for use in RDMA in a parallel computer in accordance with embodiments of the present invention may contain only a few compute nodes or may contain thousands of compute nodes. For ease of explanation, the data communications network of FIG. 4 is illustrated with only three dimensions, but readers will recognize that a data communications network optimized for point-to-point operations for use in RDMA in a parallel computer in accordance with embodiments of the present invention may in facet be implemented in two dimensions, four dimensions, five dimensions, and so on. Several supercomputers now use five dimensional mesh or torus networks, including, for example, IBM's Blue Gene Q™.

For further explanation, FIG. 5 sets forth a line drawing illustrating an example global combining network (106) useful in systems capable of RDMA in a parallel computer according to embodiments of the present invention. The example data communications network of FIG. 5 includes data communications links (103) connected to the compute nodes so as to organize the compute nodes as a tree. In the example of FIG. 5, dots represent compute nodes (102) of a parallel computer, and the dotted lines (103) between the dots represent data communications links between compute nodes. The data communications links are implemented with global combining network adapters similar to the one illustrated for example in FIG. 3B, with each node typically providing data communications to and from two children nodes and data communications to and from a parent node, with some exceptions. Nodes in the global combining network (106) may be characterized as a physical root node (202), branch nodes (204), and leaf nodes (206). The physical root (202) has two children but no parent and is so called because the physical root node (202) is the node physically configured at the top of the binary tree. The leaf nodes (206) each has a parent, but leaf nodes have no children. The branch nodes (204) each has both a parent and two children. The links and compute nodes are thereby organized by this data communications network optimized for collective operations into a binary tree (106). For clarity of explanation, the data communications network of FIG. 5 is illustrated with only 31 compute nodes, but readers will recognize that a global combining network (106) optimized for collective operations for use in RDMA in a parallel computer in accordance with embodiments of the present invention may contain only a few compute nodes or may contain thousands of compute nodes.

In the example of FIG. 5, each node in the tree is assigned a unit identifier referred to as a ‘rank’ (250). The rank actually identifies a task or process that is executing a parallel operation according to embodiments of the present invention. Using the rank to identify a node assumes that only one such task is executing on each node. To the extent that more than one participating task executes on a single node, the rank identifies the task as such rather than the node. A rank uniquely identifies a task's location in the tree network for use in both point-to-point and collective operations in the tree network. The ranks in this example are assigned as integers beginning with 0 assigned to the root tasks or root node (202), 1 assigned to the first node in the second layer of the tree, 2 assigned to the second node in the second layer of the tree, 3 assigned to the first node in the third layer of the tree, 4 assigned to the second node in the third layer of the tree, and so on. For ease of illustration, only the ranks of the first three layers of the tree are shown here, but all compute nodes in the tree network are assigned a unique rank.

For further explanation, FIG. 6 sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The parallel computer (100) of FIG. 6 includes a plurality of nodes (102a, 102b). In the example of FIG. 6, each node (102a, 102b) may be embodied, for example, as a compute node, an I/O node, or other module of automated computing machinery that forms a parallel computer (100).

In the example of FIG. 6, each node (102a, 102b) includes a messaging unit (101a, 101b). The messaging units (101a, 101b) of FIG. 6 may be embodied, for example, as a communications adapter for point-to-point data communications with other nodes in the parallel computer (100), as a library of data communications operations for carrying out data communications with other nodes in the parallel computer (100), or any combination thereof. In the example of FIG. 6, the messaging units (101a, 101b) may carry out data communications with other nodes in the parallel computer (100), for example, over a torus network as described above with reference to FIGS. 1-4.

The example method of FIG. 6 includes receiving (602), by a kernel (103a) of the first node (102a) through an application programming interface (‘API’) (109a), an RDMA read operation request (111) that includes a virtual address (113) representing a memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a). In the example of FIG. 6, the API (109a) may include one or more software routines that an application-level program can invoke to carry out RDMA operations, including an RDMA read operation designed to allow a node to read data from a specified memory location on another node. In the example of FIG. 6, the read operation request (111) includes a virtual address (113) representing a memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a). The virtual address (113) of FIG. 6 points to a location in virtual memory that is mapped to actual physical memory of the first node (102a), for example, by a page table that is accessible by the kernel (103a).

The example method of FIG. 6 also includes translating (604), by the kernel (103a) of the first node (102a), the virtual address (113) to a physical address. In the example of FIG. 6, translating (604) the virtual address (113) to a physical address in physical memory on the first node (102a) may be carried out, for example, by looking up the virtual address (113) in a page table stored on the first node (102a), by looking up the virtual address (113) in a translation lookaside buffer (‘TLB’) stored on the first node (102a), by using an address translation algorithm, and so on.

The example method of FIG. 6 also includes creating (606), by the kernel (103a) of the first node (102a), a local RDMA object (107a) that includes a counter set to the size of the memory (105a) region. In the example of FIG. 6, the local RDMA object (107a) may be embodied, for example, as a data structure useful in controlling the execution of an RDMA operation. The local RDMA object (107a) can include a counter set to the size of the memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a). In the example of FIG. 6, the size of the memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a) may be expressed in any unit of measure, for example, such as byte, kilobyte, megabyte, and so on. The counter may be useful for determining the amount of free memory remaining in the memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a), as the counter can be decremented as data is received as part of the transfer of data from the second node (102b).

The example method of FIG. 6 also includes sending (608), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA write operation request (612), the physical address (614) of the memory (105a) region on the first node (102a), the physical address (617) of the local RDMA object (107a) on the first node (102a), and a remote virtual address (616) on the second node (102b) representing the data to be transferred from the second node (102b) to the first node (102a). In the example of FIG. 6, a DMA write operation request (612) is sent from the first node (102a) to the second node (102b), to prompt the second node (102b) to write data stored in memory on the second node (102b) to memory on the first node (102a). In such an example, the DMA write operation request (612) is accompanied by the physical address (614) of the memory (105a) region on the first node (102a) that the data is to be written to. The DMA write operation request (612) is also accompanied by the physical address (617) of the local RDMA object (107a) on the first node (102a) and a remote virtual address (616) on the second node (102b). The remote virtual address (616) on the second node (102b) identifies the location, in the memory of the second node (102b), of the data to be transferred from the second node (102b) to the first node (102a).

The example method of FIG. 6 also includes receiving (610), by the first node (102a), the data (618) to be transferred from the second node (102b). In the example of FIG. 6, the first node (102a) may receive (610) the data (618) to be transferred from the second node (102b) in one or more data transfers. The data (618) to be transferred from the second node (102b) may be received by the first node (102a) over a point-to-point network such as, for example, the torus network described above with reference to FIGS. 1-4.

The example method of FIG. 6 may be utilized to implement standard data communications protocols in a parallel computer. One example of a data communications protocol that may be implemented utilizing the method of FIG. 6, for example, is the compute node verbs protocol, which is similar to the IBverbs protocol that is standardized by OFED. Compute node verbs enable communication over torus links between a compute node kernel (‘CNK’) running on a compute node and user space processes running on I/O nodes. Compute node verbs communicate with the vRNIC verbs provider in the OFED framework on Linux. The compute node verbs module allows the CNK to drive the torus links using the wire protocol defined by the vRNIC.

Compute node verbs interfaces are closely modeled on the OFED RDMA user space verbs interface. On Linux, user space processes use the standard OFED RDMA verbs interface. The compute node verbs module is light-weight and there may be differences from standard OFED RDMA verbs to maintain the low-noise requirements of CNK. The compute node verbs module retains freedom of action in implementation so long as high level requirements are met.

For further explanation, FIG. 7 sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The example method of FIG. 7 is similar to the example method of FIG. 6 as it also includes:

    • receiving (602), by a kernel (103a) of the first node (102a) through an application programming interface (‘API’) (109a), an RDMA read operation request (111) that includes a virtual address (113) representing a memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a);
    • translating (604), by the kernel (103a) of the first node (102a), the virtual address (113) to a physical address;
    • creating (606), by the kernel (103a) of the first node (102a), a local RDMA object (107a) that includes a counter set to the size of the memory (105a) region;
    • sending (608), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA write operation request (612), the physical address (614) of the memory (105a) region on the first node (102a), the physical address (617) of the local RDMA object (107a) on the first node (102a), and a remote virtual address (616) on the second node (102b) representing the data to be transferred from the second node (102b) to the first node (102a); and
    • receiving (610), by the first node (102a), the data (618) to be transferred from the second node (102b)

The example method of FIG. 7 also includes receiving (702), by the messaging unit (101b) in the second node (102b), the message that includes an DMA write operation request (612), the physical address (614) of the memory (105a) region on the first node (102a), the physical address (617) of the local RDMA object (107a) on the first node (102a), and a remote virtual address (616) on the second node (102b) representing the data to be transferred from the second node (102b) to the first node (102a). In the example of FIG. 7, the second node (102b) may receive (702) the message over a point-to-point network such as, for example, the torus network described above with reference to FIGS. 1-4.

The example method of FIG. 7 also includes translating (704), by a kernel (103b) on the second node (102b), the remote virtual address (616) on the second node (102b) to a physical address on the second node (102b). In the example of FIG. 7, translating (704) the remote virtual address (616) on the second node (102b) to a physical address in physical memory on the second node (102b) may be carried out, for example, by looking up the remote virtual address (616) in a page table stored on the second node (102b), by looking up the remote virtual address (616) in a TLB stored on the second node (102b), by using an address translation algorithm, and so on.

The example method of FIG. 7 also includes transferring (706), by one or more DMA operations (708) executing on the messaging unit (101b) in the second node (102b), the data (618) to be transferred from the second node (102b) to the first node (102a). In the example of FIG. 7, the one or more DMA operations (708) executing on the messaging unit (101b) in the second node (102b) may be embodied, for example, as computer program instructions executing on computer hardware, such as a processor, that transfer data between messaging units in nodes. In the example of FIG. 7, the DMA operations (708) executing on the messaging unit (101b) may include, for example, computer program instructions for retrieving data to be transferred from computer memory, encapsulating the data to be transferred, and transmitting the data over a computer network. In the example of FIG. 7, the DMA operations (708) transfer data without using the central processing unit(s) of the nodes, such that the data transfers occur without creating heavy workload requirements for the central processing unit(s) of the nodes.

For further explanation, FIG. 8 sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The example method of FIG. 8 is similar to the example method of FIG. 6 as it also includes:

    • receiving (602), by a kernel (103a) of the first node (102a) through an application programming interface (‘API’) (109a), an RDMA read operation request (111) that includes a virtual address (113) representing a memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a);
    • translating (604), by the kernel (103a) of the first node (102a), the virtual address (113) to a physical address;
    • creating (606), by the kernel (103a) of the first node (102a), a local RDMA object (107a) that includes a counter set to the size of the memory (105a) region;
    • sending (608), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA write operation request (612), the physical address (614) of the memory (105a) region on the first node (102a), the physical address (617) of the local RDMA object (107a) on the first node (102a), and a remote virtual address (616) on the second node (102b) representing the data to be transferred from the second node (102b) to the first node (102a); and
    • receiving (610), by the first node (102a), the data (618) to be transferred from the second node (102b)

The example method of FIG. 8 also includes decrementing (802), by the kernel (103a) on the first node (102a), the counter in the local RDMA object (107a) by an amount equal to the size of the data (618) received from the second node (102b). As described above with reference to FIG. 6, the local RDMA object (107a) can include a counter set to the size of the memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a). The size of the memory (105a) region at which to receive data to be transferred from a second node (102b) to the first node (102a) may be expressed in any unit of measure, for example, such as byte, kilobyte, megabyte, and so on. For example, if the size of the memory (105a) region at which to receive data is 12 megabytes, the counter may be set to 12. In the example of FIG. 8, the kernel (103a) on the first node (102a) can decrement (802) the counter in the local RDMA object (107a) by an amount equal to the size of the data (618) received from the second node (102b). For example, if 2 megabytes of data is received from the second node (102b), the counter may be decremented (802) by 2, thereby causing the value of the counter to be set to 10.

The example method of FIG. 8 also includes notifying (804), by the kernel (103a) on the first node (102a), a user-level application (812) that the data (618) to be transferred from the second node (102b) to the first node (102a) has been received when the counter in the local RDMA object (107a) is equal to zero. In the example of FIG. 8, when the counter in the local RDMA object (107a) is equal to zero, the data (618) that has been transferred from the second node (102b) to the first node (102a) has taken up the entire memory (105a) region at which to receive data. As such, no additional data can be stored in the memory (105a) region at which to receive data without overwriting data received in response to the RDMA read operation request (111) and subsequent DMA write operation request (612).

In the example of FIG. 8, notifying (804) the user-level application (812) that the data (618) to be transferred from the second node (102b) to the first node (102a) has been received can include raising (806) an interrupt in the user-level application. In the example of FIG. 8, raising (806) an interrupt in the user-level application (812) may be carried out by issuing an interrupt request (IRQ) indicating the need for attention by an interrupt handler. Raising (806) an interrupt in the user-level application (812) can cause a context switch to the interrupt handler that can signal, to the user-level application (812), that the RDMA read operation request (111) and subsequent DMA write operation request (612) have completed successfully.

In the example of FIG. 8, notifying (804) the user-level application (812) that the data (618) to be transferred from the second node (102b) to the first node (102a) has been received can alternatively include being (808) polled by the user-level application (812). In the example of FIG. 8, the user-level application (812) may include computer program instructions that enable the user-level application (812) to actively sample to the status of the RDMA read operation request (111) and subsequent DMA write operation request (612), for example, by periodically checking the value of the counter in the local RDMA object (107a).

The example of FIG. 8 also includes sending (810) an error message indicating that the RDMA read operation request (111) cannot be serviced. In the example of FIG. 8, the RDMA read operation request (111) may not be serviced, for example, due to a communications failure between the first node (102a) and the second node (102b), due to an error reading from memory in the second node (102b), due to an error writing to memory in the first node (102a), and so on. In response to determining that an error has occurred, an error message can be sent (810) to the first node (102a) so that the first node (102a) can retry the RDMA read operation request (111) or simply proceed without executing the RDMA read operation request (111).

For further explanation, FIG. 9 sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The example method of FIG. 9 includes receiving (908), by a kernel (103a) of a first node (102a) through an API (906), an RDMA write operation request (902) that includes a virtual address (904) representing a memory (105a) region in the first node (102a) from which to send data to a second node (102b). In the example of FIG. 9, the API (906) may include one or more software routines that an application-level program can invoke to carry out RDMA operations, including an RDMA write operation designed to allow a node to write data from a specified location in memory on the node to a specified location in memory on another node. In the example of FIG. 9, the write operation request (902) includes a virtual address (904) representing a memory (105a) region from which to retrieve data that is to be transferred to a second node (102b). The virtual address (904) of FIG. 9 points to a location in virtual memory that is mapped to actual physical memory of the first node (102a), for example, by a page table that is accessible by the kernel (103a).

The example method of FIG. 9 also includes translating (910), by the kernel (103a) of the first node (102a), the virtual address (904) to a physical address. In the example of FIG. 9, translating (910) the virtual address (904) to a physical address in physical memory on the first node (102a) may be carried out, for example, by looking up the virtual address (904) in a page table stored on the first node (102a), by looking up the virtual address (904) in a TLB stored on the first node (102a), by using an address translation algorithm, and so on.

The example method of FIG. 9 also includes creating (914), by the kernel (103a) of the first node (102a), a local RDMA object (912) that includes a counter set to the size of the memory (105a) region in the first node (102a) from which to send data to a second node (102b). In the example of FIG. 9, the local RDMA object (912) may be embodied, for example, as a data structure useful in controlling the execution of an RDMA operation. The local RDMA object (912) can include a counter set to the size of the memory (105a) region from which to send data to a second node (102b). In the example of FIG. 9, the size of the memory (105a) region from which to send data to a second node (102b) may be expressed in any unit of measure, for example, such as byte, kilobyte, megabyte, and so on. The counter may be useful for determining the amount data that is to be transferred to the second node (102b), as the counter can be decremented as data is sent to the second node (102b).

The example method of FIG. 9 also includes sending (916), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA read operation request (920), the physical address (922) of the memory region on the first node (102a), the physical address (905) of the local RDMA object on the first node (102a), and a remote virtual address (924) on the second node (102b) representing a memory region in the second node (102b) for receiving data from the first node (102a). In the example of FIG. 9, a DMA read operation request (920) is sent from the first node (102a) to the second node (102b), to prompt the second node (102b) to read data stored in memory on the first node (102a) to be written to memory on the second node (102b). In such an example, the DMA read operation request (920) is accompanied by the physical address (922) of the memory (105a) region on the first node (102a) that the data is to be read from. The DMA read operation request (920) is also accompanied by the physical address (925) of the local RDMA object (912) on the first node (102a) and a remote virtual address (924) on the second node (102b). The remote virtual address (924) on the second node (102b) identifies the location, in the memory of the second node (102b), where data from the first node (102a) is to be written into memory of the second node (102b).

The example method of FIG. 9 also includes sending (926), from the first node (102a), the data (928) to be received by the second node (102b). In the example of FIG. 9, the first node (102a) may send (926) the data (928) to the second node (102b) in one or more data transfers. The data (928) to be transferred to the second node (102b) may be received by the second node (102b) over a point-to-point network such as, for example, the torus network described above with reference to FIGS. 1-4.

For further explanation, FIG. 10 sets forth sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The example method of FIG. 10 is similar to the example method of FIG. 9, as it also includes:

    • receiving (908), by a kernel (103a) of a first node (102a) through an API (906), an RDMA write operation request (902) that includes a virtual address (904) representing a memory (105a) region in the first node (102a) from which to send data to a second node (102b);
    • translating (910), by the kernel (103a) of the first node (102a), the virtual address (904) to a physical address;
    • creating (914), by the kernel (103a) of the first node (102a), a local RDMA object (912) that includes a counter set to the size of the memory (105a) region in the first node (102a) from which to send data to a second node (102b);
    • sending (916), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA read operation request (920), the physical address (922) of the memory region on the first node (102a), the physical address (905) of the local RDMA object on the first node (102a), and a remote virtual address (924) on the second node (102b) representing a memory region in the second node (102b) for receiving data from the first node (102a); and
    • sending (926), from the first node (102a), the data (928) to be received by the second node (102b).

The example of FIG. 10 also includes receiving (1002), by the messaging unit (101b) in the second node (102b), the message that includes the DMA read operation request (920), the physical address (922) of the memory region on the first node (102a), the physical address (905) of the local RDMA object on the first node (102a), and a remote virtual address (924) on the second node (102b) representing a memory region in the second node (102b) for receiving data from the first node (102a). In the example of FIG. 10, the second node (102b) may receive (1002) the message over a point-to-point network such as, for example, the torus network described above with reference to FIGS. 1-4.

The example of FIG. 10 also includes translating (1004), by a kernel (103b) on the second node (102b), the remote virtual address (924) on the second node (102b) to a physical address on the second node (102b). In the example method of FIG. 10, translating (1004) the remote virtual address (924) on the second node (102b) to a physical address on the second node (102b) may be carried out, for example, by looking up the remote virtual address (924) in a page table stored on the second node (102b), by looking up the remote virtual address (924) in a TLB stored on the second node (102b), by using an address translation algorithm, and so on

The example of FIG. 10 also includes receiving (1006), by one or more DMA operations executing on the messaging unit (101b) in the second node, the data (928) to be transferred the first node (102a) to the second node (102b). In the example of FIG. 10, the one or more DMA operations (708) executing on the messaging unit (101b) in the second node (102b) may be embodied, for example, as computer program instructions executing on computer hardware, such as a processor, that transfer data between messaging units in nodes. In the example of FIG. 10, the DMA operations (708) executing on the messaging unit (101b) may include, for example, computer program instructions for receiving data to be to computer memory in the second node (102b). In the example of FIG. 10, the DMA operations (708) receive data without using the central processing unit(s) of the second node (102b), such that the data transfers occur without creating heavy workload requirements for the central processing unit(s) of the second node (102b).

For further explanation, FIG. 11 sets forth sets forth a flow chart illustrating an example method for RDMA in a parallel computer (100) according to embodiments of the present invention. The example method of FIG. 11 is similar to the example method of FIG. 9, as it also includes:

    • receiving (908), by a kernel (103a) of a first node (102a) through an API (906), an RDMA write operation request (902) that includes a virtual address (904) representing a memory (105a) region in the first node (102a) from which to send data to a second node (102b);
    • translating (910), by the kernel (103a) of the first node (102a), the virtual address (904) to a physical address;
    • creating (914), by the kernel (103a) of the first node (102a), a local RDMA object (912) that includes a counter set to the size of the memory (105a) region in the first node (102a) from which to send data to a second node (102b);
    • sending (916), from a messaging unit (101a) in the first node (102a) to a messaging unit (101b) in a second node (102b), a message that includes an DMA read operation request (920), the physical address (922) of the memory region on the first node (102a), the physical address (905) of the local RDMA object on the first node (102a), and a remote virtual address (924) on the second node (102b) representing a memory region in the second node (102b) for receiving data from the first node (102a); and
    • sending (926), from the first node (102a), the data (928) to be received by the second node (102b).

The example method of FIG. 11 also includes decrementing (1102), by the kernel (103a) on the first node (102a), the counter by an amount equal to the size of data (928) sent to the second node (102b). As described above with reference to FIG. 9, the local RDMA object (912) can include a counter set to the size of data (928) to be sent to the second node (102b). The size of the data (928) to be sent to the second node (102b) may be expressed in any unit of measure, for example, such as byte, kilobyte, megabyte, and so on. For example, if the size of the data (928) to be sent to the second node (102b) is 12 megabytes, the counter may be set to 12. In the example method of FIG. 11, the kernel (103a) on the first node (102a) can decrement (1102) the counter in the local RDMA object (912) by an amount equal to the size of the data (928) received by the second node (102b). For example, if 2 megabytes of data is received by the second node (102b), the counter may be decremented (1102) by 2, thereby causing the value of the counter to be set to 10.

The example method of FIG. 11 also includes notifying (1104), by the kernel (103a) on the first node (102a), a user-level application (812) that the data (928) to be transferred to the second node (102b) has been transferred when the counter is equal to zero. In the example method of FIG. 11, when the counter in the local RDMA object (912) is equal to zero, all of the data (928) that is to be transferred to the second node (102b) has been transferred. As such, no additional data needs to be transferred in response to the RDMA write operation request (902) and subsequent DMA read operation request (920).

In the example method of FIG. 11, notifying (1104) a user-level application (812) that the data (928) to be transferred to the second node (102b) has been transferred can include raising (1106) an interrupt in the user-level application (812). In the example method of FIG. 11, raising (1106) an interrupt in the user-level application (812) may be carried out by issuing an IRQ indicating the need for attention by an interrupt handler. Raising (1106) an interrupt in the user-level application (812) can cause a context switch to the interrupt handler that can signal, to the user-level application (812), that the RDMA write operation request (902) and subsequent DMA read operation request (920) have completed successfully.

In the example method of FIG. 11, notifying (1104) a user-level application (812) that the data (928) to be transferred to the second node (102b) has been transferred can alternatively include being (1108) polled by the user-level application (812). In the example of FIG. 11, the user-level application (812) may include computer program instructions that enable the user-level application (812) to actively sample to the status of the RDMA write operation request (902) and subsequent DMA read operation request (920), for example, by periodically checking the value of the counter in the local RDMA object (912).

The example method of FIG. 11 also includes sending (1110) an error message to the user-level application (812) indicating that the RDMA write operation request cannot be serviced. In the example method of FIG. 11, the RDMA write operation request (902) may not be serviced, for example, due to a communications failure between the first node (102a) and the second node (102b), due to an error reading from memory in the first node (102a), due to an error writing to memory in the second node (102b), and so on. In response to determining that an error has occurred, an error message can be sent (1110) to the first node (102a) so that the first node (102a) can retry the RDMA write operation request (912) or simply proceed without executing the RDMA write operation request (912).

As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or computer program product. Accordingly, aspects of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a “circuit,” “module” or “system.” Furthermore, aspects of the present invention may take the form of a computer program product embodied in one or more computer readable medium(s) having computer readable program code embodied thereon.

Any combination of one or more computer readable medium(s) may be utilized. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the “C” programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider).

Aspects of the present invention are described above with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other devices to function in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

It will be understood from the foregoing description that modifications and changes may be made in various embodiments of the present invention without departing from its true spirit. The descriptions in this specification are for purposes of illustration only and are not to be construed in a limiting sense. The scope of the present invention is limited only by the language of the following claims.

Claims

1. A method of remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, the method comprising:

receiving, by a kernel of the first node through an application programming interface (‘API’), an RDMA read operation request that includes a virtual address representing a memory region at which to receive data to be transferred from a second node to the first node;
responsive to the RDMA read operation request:
translating, by the kernel of the first node, the virtual address to a physical address;
creating, by the kernel of the first node, a local RDMA object that includes a counter set to the size of the memory region;
sending, from a messaging unit in the first node to a messaging unit in a second node, a message that includes an DMA write operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node representing the data to be transferred from the second node to the first node; and
receiving, by the first node responsive to the second node's execution of the DMA write operation request, the data to be transferred from the second node.

2. The method of claim 1 further comprising:

receiving, by the messaging unit in the second node, the message;
translating, by a kernel on the second node, the remote virtual address on the second node to a physical address on the second node; and
transferring, by one or more DMA operations executing on the messaging unit in the second node, the data to be transferred from the second node to the first node.

3. The method of claim 1 further comprising:

decrementing, by the kernel on the first node, the counter by an amount equal to the size of data received from the second node; and
notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received when the counter is equal to zero.

4. The method of claim 3 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received includes raising an interrupt in the user-level application.

5. The method of claim 3 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received includes being polled by the user-level application.

6. The method of claim 1 further comprising sending an error message indicating that the RDMA read operation request cannot be serviced.

7. A method of remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, the method comprising:

receiving, by a kernel of a first node through an application programming interface (‘API’), an RDMA write operation request that includes a virtual address representing a memory region in the first node from which to send data to a second node;
responsive to the RDMA write operation request:
translating, by the kernel of the first node, the virtual address to a physical address;
creating, by the kernel of the first node, a local RDMA object that includes a counter set to the size of the memory region;
sending, from a messaging unit in the first node to a messaging unit in a second node, a message that includes an DMA read operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node representing a memory region in the second node for receiving data from the first node; and
sending, from the first node responsive to the second node's execution of the DMA read operation request, the data to be received by the second node.

8. The method of claim 7 further comprising:

receiving, by the messaging unit in the second node, the message;
translating, by a kernel on the second node, the remote virtual address on the second node to a physical address on the second node; and
receiving, by one or more DMA operations executing on the messaging unit in the second node, the data to be transferred the first node to the second node.

9. The method of claim 7 further comprising:

decrementing, by the kernel on the first node, the counter by an amount equal to the size of data sent to the second node; and
notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred when the counter is equal to zero.

10. The method of claim 9 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred includes raising an interrupt in the user-level application.

11. The method of claim 9 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred includes being polled by the user-level application.

12. The method of claim 7 further comprising sending an error message to a user-level application indicating that the RDMA write operation request cannot be serviced.

13. An apparatus for remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, the apparatus comprising a computer processor, a computer memory operatively coupled to the computer processor, the computer memory having disposed within it computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

receiving, by a kernel of the first node through an application programming interface (‘API’), an RDMA read operation request that includes a virtual address representing a memory region at which to receive data to be transferred from a second node to the first node;
responsive to the RDMA read operation request:
translating, by the kernel of the first node, the virtual address to a physical address;
creating, by the kernel of the first node, a local RDMA object that includes a counter set to the size of the memory region;
sending, from a messaging unit in the first node to a messaging unit in a second node, a message that includes an DMA write operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node representing the data to be transferred from the second node to the first node; and
receiving, by the first node responsive to the second node's execution of the DMA write operation request, the data to be transferred from the second node.

14. The apparatus of claim 13 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

receiving, by the messaging unit in the second node, the message;
translating, by a kernel on the second node, the remote virtual address on the second node to a physical address on the second node; and
transferring, by one or more DMA operations executing on the messaging unit in the second node, the data to be transferred from the second node to the first node.

15. The apparatus of claim 13 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

decrementing, by the kernel on the first node, the counter by an amount equal to the size of data received from the second node; and
notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received when the counter is equal to zero.

16. The apparatus of claim 15 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received includes raising an interrupt in the user-level application.

17. The apparatus of claim 15 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred from the second node to the first node has been received includes being polled by the user-level application.

18. The apparatus of claim 13 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the step of sending an error message indicating that the RDMA read operation request cannot be serviced.

19. An apparatus for remote direct memory access (‘RDMA’) in a parallel computer, the parallel computer including a plurality of nodes, each node including a messaging unit, the apparatus comprising a computer processor, a computer memory operatively coupled to the computer processor, the computer memory having disposed within it computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

receiving, by a kernel of a first node through an application programming interface (‘API’), an RDMA write operation request that includes a virtual address representing a memory region in the first node from which to send data to a second node;
responsive to the RDMA write operation request:
translating, by the kernel of the first node, the virtual address to a physical address;
creating, by the kernel of the first node, a local RDMA object that includes a counter set to the size of the memory region;
sending, from a messaging unit in the first node to a messaging unit in a second node, a message that includes an DMA read operation request, the physical address of the memory region on the first node, the physical address of the local RDMA object on the first node, and a remote virtual address on the second node representing a memory region in the second node for receiving data from the first node; and
sending, from the first node responsive to the second node's execution of the DMA read operation request, the data to be received by the second node.

20. The apparatus of claim 19 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

receiving, by the messaging unit in the second node, the message;
translating, by a kernel on the second node, the remote virtual address on the second node to a physical address on the second node; and
receiving, by one or more DMA operations executing on the messaging unit in the second node, the data to be transferred the first node to the second node.

21. The apparatus of claim 19 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the steps of:

decrementing, by the kernel on the first node, the counter by an amount equal to the size of data sent to the second node; and
notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred when the counter is equal to zero.

22. The apparatus of claim 21 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred includes raising an interrupt in the user-level application.

23. The apparatus of claim 21 wherein notifying, by the kernel on the first node, a user-level application that the data to be transferred to the second node has been transferred includes being polled by the user-level application.

24. The apparatus of claim 19 further comprising computer program instructions that, when executed by the processor, cause the apparatus to carry out the step of sending an error message indicating that the RDMA write operation request cannot be serviced.

Patent History

Publication number: 20120331243
Type: Application
Filed: Jun 24, 2011
Publication Date: Dec 27, 2012
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION (Armonk, NY)
Inventors: Michael E. Aho (Rochester, MN), Thomas M. Gooding (Rochester, MN), Michael B. Mundy (Rochester, MN), Andrew T. Tauferner (Rochester, MN)
Application Number: 13/167,950

Classifications