Distributed 8x8 low-latency high-bandwidth cross-point cache queue on-chip router
A cache queue and low-latency technology, applied in the field of distributed 8X8 low-latency high-bandwidth cross-point cache queue on-chip routers, can solve problems such as unavailable transmission, and achieve the effects of reducing complexity, low latency, and improving performance
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0019] Such as figure 1 As shown, a distributed 8X8 low-latency high-bandwidth cross-point cache queue on-chip router includes three components: a processor label filter (1), a cluster label filter (2), and an output module (3). Its features are: the input nodes corresponding to output ports 0, 2, 4, and 6 are mounted with processor tag filters (1); the output nodes corresponding to output ports 1, 3, 5, and 7 are mounted with cluster tags Filter (2), each output port is mounted with an output module (3). When valid data is input, the processor label filter (1) at each node detects the data sent by the input port at the same time, and ignores the data when the processor label (cluster number and core number) of the data does not match itself, if Matching will enter the accepting operation, and at the same time send the output request signal to the corresponding output module, if the permission signal is received, the data will be sent directly to the output module (3), within...
Embodiment 2
[0021] This embodiment is basically the same as Embodiment 1, and the special features are as follows:
[0022] , processor tag filter (1)
[0023] Such as figure 2 : Processor tag filter (1) is designed for 2~3 stage variable pipeline. When the data is valid, the data is stored in the first-level register, and the processor label and the value in the processor label register are taken out at the same time. If they match, a match signal is generated. When the match signal is valid and the data is valid or the FIFO is not empty, an arbitration request is generated. Signal. After the matching is successful, the data enters the second register and waits for the response signal. If the response signal has arrived and the FIFO is empty before the next clock arrives, the data is directly sent to the data bus through the multiplexer and the tri-state gate. ; If not, the data enters the FIFO queue and waits for the arbitration output. When it does not pass through the FIFO, it is...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 