UPGRADE YOUR BROWSER

We have detected your current browser version is not the latest one. Xilinx.com uses the latest web technologies to bring you the best online experience possible. Please upgrade to a Xilinx.com supported browser:Chrome, Firefox, Internet Explorer 11, Safari. Thank you!

cancel
Showing results for 
Search instead for 
Did you mean: 
Visitor rrfisher1
Visitor
151 Views
Registered: ‎03-06-2019

Understand the Correct Usage of the Floating-Point Accumulator in Vivado

Hi Vivado Community,

I am conducting research on potential speed-ups of various applications and would like to utilize a multiply and accumulate pipeline to quickly perform matrix multiplication; then time to see exactly how long it would take to perform the operation down to the clock cycles. The problem I seem to be having is that the data output from the floating-point accumulaor is a mirror of the data input to the IP. I've been though the various documentions and cannot determine the culprit of the phenomenon. I've included various pictures of what is happening in the ILA to help someone determine what could possibly be the problem.

3. ILA Capture Data Out Begin.PNG4. ILA Capture Multiply Data Out (MxM_data_out).PNG5. ILA Capture Accumulator Data Out (accum_data).PNG1. ILA Capture Full.PNG2. ILA Capture Zoomed in on Operation.PNGAccumulator Config Capture.PNG

0 Kudos