Menu Close

Weekly Status Report #5

Team Status Report

DEMO PLAN

To demonstrate what our ultimate application will look like, Riki will demonstrate a software implementation of time carving. The demo will show pre-processed videos, so we will not be running the algorithm live. To show the acceleration side of the project, John and Maxwell will demonstrate the SoC interacting with the programmable logic. In the demo, the programmable fabric will compute the energy map of a single frame of video supplied by the SoC. The SoC will provide the user-interface and external communication.

RISKS

Since we still don’t know about the algorithm, our midpoint demo might be incomplete.

 

Maxwell Johnson

Personal work:

This week I have been working on connecting to the DRAM through our FPGA. I worked with Xilinx’s Memory Interface Generator (MIG) tool to create an example design that writes to and reads from the boards DDR3 memory. The tool generates a user interface to the memory in the form of an IP block. I can instantiate this as a module in our design to allow the programmable fabric to interact with the memory.

Progress:

Interfacing to the memory is a big step forward in my section of the project. I am a few days behind the schedule in our Gantt chart, but I am catching up to where I am scheduled to be. In the next 4 days, I need to be able to interface with a program running on the SoC through the memory. I anticipate that I’ll be able to create my half of that interface in that time. The SoC is behind schedule, so I may have to reorder my pipeline so as to not stall 😛

Deliverables:

This week, I plan to have my half of the interface to the SoC completed. If the SoC side is also ready this week, John and I will be able to debug that interface and establish communication. In the latter half of the week, I will work on preparing a demo for the following week. I already have the sketch of a design for the energy function, so showing that in action would make an ideal demo. For the demo, we will work with a single frame, so the energy function needs to be slightly modified to exclude the time dimension.

John Zhang

Personal Work

 

Progress

 

Deliverables

 

Riki Khorana

Personal Work

This week I was working on securing my understanding of the graph cut algorithm found here. I have installed all the software necessary for implementation.

Progress

I am a few days behind on the gantt chart, mostly because I couldn’t get as much reading done over the spring break as I wanted to. I plan to make significant progress on the software implementation over the weekend.

Deliverables

I am going to explain the graph cut algorithm in detail next Monday to my teammates. The biggest deadline approaching is the midpoint demo happening in two weeks. I aim to finish the software implementation of the seam-carving algorithm next week so I definitely have something presentable.

Leave a Reply

Your email address will not be published. Required fields are marked *