Hi,
I'm new to VHDL and am planning on using it for a school project. The project I'm doing will require implementing some fairly basic matrix math on hardware. What would be the best way to tackle accessing a block of BRAM that's used to allocate a matrix of data? I know that VHDL has the capability for using loops and being that I have more experience in the software realm, I'm drawn towards that naturally. For the purposes of simplicity, let's say all I'm trying to do is analyze a 20x20 matrix of 1-bit values stored in BRAM and create from that two output arrays of size 20, one will be a sum of the bits in each row and one would be a sum of the bits in each column. If I implement this with a loop in VHDL, will this be highly inefficient and run like a software loop does, where only one loop iteration runs per clock cycle? If so, is there a better way to do math of this sort in hardware. Let me know if more clarification is needed.
Thanks in advance for your help,
-Daniel
I'm new to VHDL and am planning on using it for a school project. The project I'm doing will require implementing some fairly basic matrix math on hardware. What would be the best way to tackle accessing a block of BRAM that's used to allocate a matrix of data? I know that VHDL has the capability for using loops and being that I have more experience in the software realm, I'm drawn towards that naturally. For the purposes of simplicity, let's say all I'm trying to do is analyze a 20x20 matrix of 1-bit values stored in BRAM and create from that two output arrays of size 20, one will be a sum of the bits in each row and one would be a sum of the bits in each column. If I implement this with a loop in VHDL, will this be highly inefficient and run like a software loop does, where only one loop iteration runs per clock cycle? If so, is there a better way to do math of this sort in hardware. Let me know if more clarification is needed.
Thanks in advance for your help,
-Daniel