neorv32 VS picorv32

How about NEORV32?

GitHub - stnolting/neorv32: 🖥️ A tiny, customizable and highly extensible MCU-class 32-bit RISC-V soft-core CPU and microcontroller-like SoC written in platform-independent VHDL. this one is good but is written in VHDL though

NEORV32 is an open source soft core and very well documented. I would recommend you to take a look at it and play around a bit. And it is certainly possible to have a soft core running on only the PL side without PS interference.

Maybe not a classic (whatever that means...) project, but I am working (together with others) on a RISC-V microcontroller for FPGAs: https://github.com/stnolting/neorv32

In contrast to most people here saying you NEED to spend money. I disagree with that. You can implement and simulate a SPI master/slave fully on your computer, no FPGA or other hardware required. There are simulation models for SPI peripherals you could use. For example: https://github.com/YosysHQ/picorv32/blob/master/picosoc/spiflash.v

The Pico RV32 is pretty small, and can go as low as about 750 LUTs, with most features elided. I don't know how Xilinix LUTs translate to Lattice though.

Picorv32: https://github.com/YosysHQ/picorv32

There are loads of free RISC-V cores that you can read the source of and run on cheap FPGAs. Take a look at PicoRV32: https://github.com/YosysHQ/picorv32

picorv32 is written in Verilog.

In short: it works, though the implementation lacks the crystal clarity of FemtoRV32 and PicoRV32. The core is larger than SERV but has higher IPC and (very arguably) a more conventional implementation. The compressed instruction set is easier to expand into regular RV32I instructions than it is to execute directly.

That is, reducing the number of LUT required to implement a CPU of a given ISA.

A basic RV32 CPU is down to 500-700 LUT.

    https://github.com/YosysHQ/picorv32

I've bought a cheap FPGA board (Sipeed Tang Nano 9K) because I want to implement a little 8 or 16-bit CPU. The FPGA has plenty of BRAM for such a little CPU, so I wouldn't even need to implement an SPI controller initially, but I want to implement a von Neumann architecture, and was wondering if the only way of doing so using single port (or semi dual port) RAM would be to use 2 cycles or more for memory transfer operations (one for loading the instruction, one for executing the actual memory transfer), or if there was any technique that could be used to avoid this without having to implement instruction-level parallelism. Even if not, references to understandable code implementing a simple memory interface would be appreciated. I looked at PicoRV32 but couldn't really understand its inner workings.

Have you looked at PicoRV32? https://github.com/YosysHQ/picorv32

What I mean is that you use the FPGA fabric to implement a soft-core CPU, like MicroBlaze (Xilinx) or Nios II (Altera/Intel) or RISC-V or any other CPU you like. Then you can do the MP3 or WAV signal decoding in software, which will be orders of magnitude easier to do than to do it in hardware. For a media player, this is more than adequate.