首页> 外文会议>2012 18th IEEE-NPSS Real Time Conference. >High performance FPGA-based DMA interface for PCIe
【24h】

High performance FPGA-based DMA interface for PCIe

机译:用于PCIe的基于FPGA的高性能DMA接口

获取原文
获取原文并翻译 | 示例

摘要

We present a data communication suite developed for use in the Track Engine Trigger for the IceCube Neutrino Observatory (South Pole). It is a PCIe-based system implemented in Xilinx FPGAs with a bus master DMA on a 4-lane, generation 2 link. The suite contains DMA controller hardware IPs, test benches, Linux driver and user application for DMA and PIO transfers into memory modules and FIFOs. The Linux driver uses streaming mapping, vector write functionality, race condition controllers, page-wise memory allocation, wait queues and Message Signaled Interrupt (MSI) to facilitate high performance and throughput. The DMA which is based on the Xilinx' bus master DMA, produces measured transfer speeds up to 748MB/s (read) and 784MB/s (write) using the Xilinx VC707 Virtex-7 board. The hardware has been verified on different platforms with different FPGAs. Besides the original IceCube application, the suite has also been used for the development of readout electronics for particle physics experiments. Other applications are also considered.
机译:我们提供了一个数据通信套件,该套件是为IceCube Neutrino天文台(南极)的跟踪引擎触发器开发的。它是在Xilinx FPGA中实现的基于PCIe的系统,在4通道,第2代链路上具有总线主控DMA。该套件包含DMA控制器硬件IP,测试平台,Linux驱动程序以及用于将DMA和PIO传输到内存模块和FIFO中的用户应用程序。 Linux驱动程序使用流映射,向量写入功能,竞争条件控制器,逐页内存分配,等待队列和消息信号中断(MSI)来促进高性能和吞吐量。基于Xilinx总线主控DMA的DMA使用Xilinx VC707 Virtex-7板可产生高达748MB / s(读取)和784MB / s(写入)的测量传输速度。该硬件已在具有不同FPGA的不同平台上进行了验证。除了原始的IceCube应用程序外,该套件还用于开发用于粒子物理实验的读出电子设备。还考虑了其​​他应用。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号