Abstract:
A network-on-chip flit transmission method and a network-on-chip flit transmission apparatus. The method includes: receiving a current flit including an operation ID Op ID, an operation type Op Type, and a payload; when determining that acceleration processing needs to be performed on the current flit, determining, according to the Op Type, whether the payload of the current flit affects an execution result of a destination node of the current flit, where the execution result is a result obtained by executing, by the destination node, a task corresponding to the Op ID; and forwarding the current flit to a next node if a determining result is yes; and discarding the current flit if a determining result is no. Therefore, not only a quantity of flits that need to be transmitted on a network-on-chip can be greatly reduced, but also a calculation amount of the destination node can be reduced.
Abstract:
An optical network-on-chip and a method and an apparatus for dynamically adjusting optical link bandwidth is presented, wherein each fixedly interconnected optical transceiver in a cluster in the optical network-on-chip is configured to establish a link between the cluster and one cluster in other n-x clusters to exchange an optical signal; and a main controller is configured to allocate x adaptively interconnected transceivers to k fixed links with the heaviest communication traffic according to a set rule and communication traffic of fixed links established by n-x fixedly interconnected optical transceivers in the cluster; and for an adaptively interconnected optical transceiver in the x adaptively interconnected optical transceivers, control the adaptively interconnected optical transceiver to establish a link, except the fixed link, between two clusters connected by the fixed link.
Abstract:
A computer device and a method for reading or writing data by a computer device are provided. In the computer device, a central processing unit (CPU) is connected to a cloud controller using a double data rate (DDR) interface. Because the DDR interface has a high data transmission rate, interruption of CPU can be avoided. In addition, the CPU converts a read or write operation request into a control command and writes the control command into a transmission queue in the cloud controller. Because the cloud controller performs a read operation or a write operation on a network device according to operation information in the control command, after writing the control command into the transmission queue, the CPU does not need to wait for an operation performed by the cloud controller and can continue to perform other processes.
Abstract:
A memory access device allocate a memory resource to a node in a system to reduce maintenance overheads of the system and implement flexible scheduling of memory resources. The device includes a cloud control device on the side of a requesting node and a cloud control device on the side of a contributing node. The cloud control device on the side of the requesting node generates a request packet for accessing to-be-accessed data stored in the contributing node which provides memory resource, sends the request packet to the contributing node t. The cloud control device on the side of the contributing node receives the request packet, provides a request message to the contributing node, generates the response packet for the contributing node, and sends the response packet to the requesting node.
Abstract:
A computer device and a method for reading or writing data by a computer device are provided. In the computer device, a central processing unit (CPU) is connected to a cloud controller using a double data rate (DDR) interface. Because the DDR interface has a high data transmission rate, interruption of CPU can be avoided. In addition, the CPU converts a read or write operation request into a control command and writes the control command into a transmission queue in the cloud controller. Because the cloud controller performs a read operation or a write operation on a network device according to operation information in the control command, after writing the control command into the transmission queue, the CPU does not need to wait for an operation performed by the cloud controller and can continue to perform other processes.
Abstract:
An optical network-on-chip and a method and an apparatus for dynamically adjusting optical link bandwidth is presented, wherein each fixedly interconnected optical transceiver in a cluster in the optical network-on-chip is configured to establish a link between the cluster and one cluster in other n-x clusters to exchange an optical signal; and a main controller is configured to allocate x adaptively interconnected transceivers to k fixed links with the heaviest communication traffic according to a set rule and communication traffic of fixed links established by n-x fixedly interconnected optical transceivers in the cluster; and for an adaptively interconnected optical transceiver in the x adaptively interconnected optical transceivers, control the adaptively interconnected optical transceiver to establish a link, except the fixed link, between two clusters connected by the fixed link.
Abstract:
A network-on-chip flit transmission method and a network-on-chip flit transmission apparatus. The method includes: receiving a current flit including an operation ID Op ID, an operation type Op Type, and a payload; when determining that acceleration processing needs to be performed on the current flit, determining, according to the Op Type, whether the payload of the current flit affects an execution result of a destination node of the current flit, where the execution result is a result obtained by executing, by the destination node, a task corresponding to the Op ID; and forwarding the current flit to a next node if a determining result is yes; and discarding the current flit if a determining result is no. Therefore, not only a quantity of flits that need to be transmitted on a network-on-chip can be greatly reduced, but also a calculation amount of the destination node can be reduced.
Abstract translation:一种片上片传输方法和片上片上传输装置。 该方法包括:接收包括操作ID Op ID,操作类型Op类型和有效载荷的当前飞行; 当确定需要对当前飞行进行加速处理时,根据Op类型确定当前飞行的有效载荷是否影响当前飞行的目的地节点的执行结果,其中执行结果是获得的结果 通过目的地节点执行与该操作ID相对应的任务; 并且如果确定结果为是,则将所述当前飞行转发到下一个节点; 并且如果确定结果为否则丢弃当前的飞行。 因此,不仅可以大大减少在片上网络上需要发送的数量的闪烁,还可以减少目的地节点的计算量。