TDT4255/introduction.org
2022-09-13 08:37:27 +02:00

200 lines
8.7 KiB
Org Mode

* About RISCV-FiveStage
The task in this project is to implement a 5-stage pipelined processor for
the [[./instructions.org][RISCV32I instruction set]].
This project framework is used for the all of the milestones in the processor
design course TDT4255, however you are more than welcome to use this project
yourself, or to teach a class. Please reach out if you do!
In this project you will build a 5-stage RISCV32I processor that is able to run
real RISC-V programs as long as they only use the 32I instruction subset.
Since this is your first time building a processor, starting with a 5-stage design
presents a very difficult challenge, which is why this project is split into four
parts. In the first two parts the instructions will be interspersed with NOP instructions,
four NOPs for every real. This means that you do not need to take into account
dependencies and so forth, making things a lot easier for you.
For the last two parts the only difference is that NOP instructions will not be
inserted. You can read about this in the [[exercise2.org][ex2 guide]], and will not be discussed
further here.
In the project skeleton files ([[./src/main/scala/][Found here]]) you can see that a lot of code has
already been provided, which can make it difficult to get started.
Hopefully this document can help clear up at least some of the confusion.
The rest of this document gives an overview of the exercise framework and testing.
If you want to jump straight to something practical you can start following the
[[exercise.org][exercise guide]], however at some point you should read through the rest of this document.
** A tour of FiveStage
In order to orient yourself you first need a map, thus a high level overview of the
processor you're going to design is showed underneath:
Keep in mind that this is just a high level sketch, omitting many details as well
entire features (for instance branch logic)
#+CAPTION: A very high level processor schematic. Registers, Instruction and data memory are already implemented.
#+attr_html: :width 1000px
#+attr_latex: :width 1000px
[[./Images/FiveStage.png]]
Now that you have an idea of what you're building it is time to take inventory of
the files included in the skeleton, and what, if anything should be added.
+ [[./src/main/scala/Tile.scala]]
This is the top level module for the system as a whole. This is where the test
harness accessses your design, providing the necessary IO.
*You should not modify this module for other purposes than debugging.*
+ [[./src/main/scala/CPU.scala]]
This is the top level module for your processor.
In this module the various stages and barriers that make up your processor
should be declared and wired together.
Some of these modules have already been declared in order to wire up the
debugging logic for your test harness.
This file corresponds to the high-level overview in its entirety.
*This module is intended to be further fleshed out by you.*
As you work with this module, try keeping logic to a minimum to help readability.
If you end up with a lot of signal select logic, consider moving that to a separate
module.
+ [[./src/main/scala/IF.scala]]
This is the instruction fetch stage.
In this stage instruction fetching should happen, meaning you will have to
add logic for handling branches, jumps, and eventually, stalls.
The reason this module is already included is that it contains the instruction
memory, described next which is heavily coupled to the testing harness.
*This module is intended to be further fleshed out by you.*
+ [[./src/main/scala/IMem.scala]]
This module contains the instruction memory for your processor.
Upon testing the test harness loads your program into the instruction memory,
freeing you from the hassle.
*You should not modify this module for other purposes than maaaaybe debugging.*
+ [[./src/main/scala/ID.scala]]
The instruction decode stage.
The reason this module is included is that the registers reside here, thus
for the test harness to work it must be wired up to the register unit to
record its state updates.
*This module is intended to be further fleshed out by you.*
+ [[./src/main/scala/Registers.scala]]
Contains the registers for your processor. Note that the zero register is alredy
disabled, you do not need to do this yourself.
The test harness ensures that all register updates are recorded.
*You should not modify this module for other purposes than maaaaybe debugging.*
+ [[./src/main/scala/MEM.scala]]
Like ID and IF, the MEM skeleton module is included so that the test harness
can set up and monitor the data memory
*This module is intended to be further fleshed out by you.*
+ [[./src/main/scala/DMem.scala]]
Like the registers and Imem, the DMem is already implemented.
*You should not modify this module for other purposes than maaaaybe debugging.*
+ [[./src/main/scala/Const.scala]]
Contains helpful constants for decoding, used by the decoder which is provided.
*This module may be fleshed out further by you if you so choose.*
+ [[./src/main/scala/Decoder.scala]]
The decoder shows how to conveniently demux the instruction.
In the provided ID.scala file a decoder module has already been instantiated.
You should flesh it out further.
You may find it useful to alter this module, especially in exercise 2.
*This module should be further fleshed out by you.*
+ [[./src/main/scala/ToplevelSignals.scala]]
Contains helpful constants.
You should add your own constants here when you find the need for them.
You are not required to use it at all, but it is very helpful.
*This module can be further fleshed out by you.*
+ [[./src/main/scala/SetupSignals.scala]]
You should obviously not modify this file.
You may choose to create a similar file for debug signals, modeled on how
the test harness is built.
*You should not modify this module at all.*
** Tests
In addition to the skeleton files it's useful to take a look at how the tests work.
You will not need to alter anything here other than the [[./src/test/scala/Manifest.scala][test manifest]], but some
of these settings can be quite useful to alter.
The main attraction is the test options. By altering the verbosity settings you
may change what is output.
The settings are:
+ printIfSuccessful
Enables logging on tests that succeed.
You typically want this turned off, at least for the full test runner.
+ printErrors
Enables logging of errors. You obviously want this one on, at least on the single
test.
+ printParsedProgram
Prints the desugared program. Useful when the test asm contains instructions that
needs to be expanded or altered.
Unsure what "bnez" means? Turn this setting on and see!
+ printVMtrace
Enables printing of the VM trace, showing how the ideal machine executes a test
+ printVMfinal
Enables printing of the final VM state, showing how the registers look after
completion. Useful if you want to see what a program returns.
+ printMergedTrace
Enables printing of a merged trace. With this option enabled you get to see how
the VM and your processor executed the program side by side.
This setting is extremely helpful to track down where your program goes wrong!
This option attempts to synchronize the execution traces as best as it can, however
once your processor design derails this becomes impossible, leading to rather
nonsensical output.
The output should look like this (picture is from exercise 2, without NOP padding)
#+attr_html: :width 300px
#+attr_latex: :width 3000px
[[./Images/merged.png]]
Instructions that were only executed by either VM or Your design is colored red or
blue.
*IF YOU ARE COLOR BLIND YOU SHOULD ALTER THE DISPLAY COLORS!*
On some windows terminal emulators there exists a bug that causes colors to not display
correctly, giving your terminal a very.. rastafarian look as shown below:
#+attr_html: :width 300px
#+attr_latex: :width 3000px
[[./Images/rasta.png]]
+ nopPadded
Set this to false when you're ready to enter the big-boy league
+ breakPoints
Not implemented. It's there as a teaser, urging you to implement it so I don't have to.