You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, when I used the 'finn.transformation.fpgadataflow.create_stitched_ip.CreateStitchedIP(fpgapart、clk_ns、ip_name
='finn_design'、vitis=False、signature=[] )' API, and also when I followed the example to build 'Stitched IP, out-of-context synth and rtlsim Performance' for my own network, like this:
As I was doing both, I ran into this problem, like this:
It says my SIMD should >= MW/1024, and I set SIMD = 16 for 'MatrixVectorActivation_4' when I did not use the 'CreateStitchedIP',like this:
In this part, the implement succeeded, but I want to know more about how FINN works, see more code.
So I did what I said in the beginning, but I tried both of the model, one is SIMD=1 and another is SIMD=16, they can not pass the HLS synthesis.
How can I build stitched ip and fix this problem? I do have a solution in my mind that I can change the network to make MW < 1024, but that maybe cause other problems,
could you give me another solution?
I did the implement of my model on PYNQ-Z2, but I want to see the ip in Vivado, and I also want to know can I change the HLS code
which FINN wrote, so I can maybe do a little change to let the FPGA works faster?
Looking forward to your reply!
Thanks
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi, when I used the 'finn.transformation.fpgadataflow.create_stitched_ip.CreateStitchedIP(fpgapart、clk_ns、ip_name
='finn_design'、vitis=False、signature=[] )' API, and also when I followed the example to build 'Stitched IP, out-of-context synth and rtlsim Performance' for my own network, like this:
As I was doing both, I ran into this problem, like this:
It says my SIMD should >= MW/1024, and I set SIMD = 16 for 'MatrixVectorActivation_4' when I did not use the 'CreateStitchedIP',like this:
In this part, the implement succeeded, but I want to know more about how FINN works, see more code.
So I did what I said in the beginning, but I tried both of the model, one is SIMD=1 and another is SIMD=16, they can not pass the HLS synthesis.
How can I build stitched ip and fix this problem?
I do have a solution in my mind that I can change the network to make MW < 1024, but that maybe cause other problems,
could you give me another solution?
I did the implement of my model on PYNQ-Z2, but I want to see the ip in Vivado, and I also want to know can I change the HLS code
which FINN wrote, so I can maybe do a little change to let the FPGA works faster?
Looking forward to your reply!
Thanks
Beta Was this translation helpful? Give feedback.
All reactions