MyHDL: a Python-Based Hardware Description Language
Digital hardware design typically is done using a specialized language, called a hardware description language (HDL). This approach is based on the idea that hardware design has unique requirements. The mainstream HDLs are Verilog and VHDL.
The MyHDL Project challenges conventional wisdom by making it possible to use Python, a high-level, general-purpose language, for hardware design. This approach lets hardware designers benefit from a well-designed, widely used language and the open-source model behind it.
An HDL contains certain concepts that reflect the nature of hardware. The most characteristic concept is a model for massive concurrency. An HDL description consists of a large amount of tiny threads that communicate closely with one another. This design calls for an approach to threading that is as lightweight as possible. HDL descriptions are executed on a dedicated runtime environment called a simulator.
When designing MyHDL, I took a minimalistic approach, which is in line with the Python spirit and a good idea in general. Therefore, an important part of MyHDL is a usage model for Python. The other part consists of a Python package, called myhdl, that contains objects that implement HDL concepts. The following Python code imports some MyHDL objects that we are going to use shortly:
from myhdl import Signal, Simulation, delay, now
MyHDL models concurrency with generator functions, recently introduced in Python (see the on-line Resources). They are similar to classic functions, except they have a nonfatal return capability. When a generator function is called, it returns a generator, which is the object of interest. Generators are resumable and keep state between invocations, making them usable as ultra-lightweight threads.
The following example is a generator function that models a clock generator:
def clkgen(clk): """ Clock generator. clk -- clock signal """ while 1: yield delay(10) clk.next = not clk
This function looks similar to a classic function in Python. Notice that the functional code starts with a while 1 loop; this is the idiomatic way to keep a generator alive forever. The essential difference between a classic and a generator function is the yield statement. It behaves similarly to a return statement, except the generator remains alive after yielding and can be resumed from that point. Moreover, the yield statement returns a delay object. In MyHDL, this mechanism is used to pass control information to the simulator. In this case, the simulator is informed that it should resume the generator after a delay of ten time units.
The parameter clk represents a clock signal. In MyHDL, signals are used for communication among generators. The concept of a signal is inherited from VHDL. A signal is an object with two values: a read-only current value and a next value that can be modified by assigning it to the .next attribute. In the example, the clock signal is toggled by setting its next value to the inverse of its current value.
To simulate the clock generator, we first create a clock signal:
clk = Signal(bool(0))
The signal clk has a Boolean zero as its initial value. Now, we can create a clock generator by calling the generator function:
clkgen_inst = clkgen(clk)
To have a minimally useful simulation, let's create another generator that monitors and prints the changes of the clock signal over time:
def monitor(): print "time: clk" while 1: print "%4d: %s" % (now(), int(clk)) yield clk
The yield clk statement shows how a generator can wait on a change of the signal value.
In MyHDL, a simulator is created with the Simulation object constructor, which takes an arbitrary number of generators as parameters:
sim = Simulation(clkGen_inst, monitor())
To run the simulator, we call its run method:
This runs the simulation for 50 time units. The output is as follows:
$ python clkgen.py time: clk 0: 0 10: 1 20: 0 30: 1 40: 0 50: 1
At this point, we can describe how the simulator works. The simulation algorithm is inspired by VHDL, an HDL slightly less popular than Verilog but a better example to follow. The simulator coordinates the parallel execution of all generators using an event-driven algorithm. The object that a generator yields specifies the event for which it wants to wait before its next invocation.
Suppose that at a given simulation step, some generators become active because the event they were waiting on has occurred. In a first simulation phase, all active generators are run, using current signal values and assigning to next values. In a second phase, the current signal values are updated with the next values. As a result of signal value changes, some generators become active again, and the simulation cycle repeats. This mechanism guarantees determinism, because the order in which the active generators are run is irrelevant for the behavior of the model.
|Updates from LinuxCon and ContainerCon, Toronto, August 2016||Aug 23, 2016|
|NVMe over Fabrics Support Coming to the Linux 4.8 Kernel||Aug 22, 2016|
|What I Wish I’d Known When I Was an Embedded Linux Newbie||Aug 18, 2016|
|Pandas||Aug 17, 2016|
|Juniper Systems' Geode||Aug 16, 2016|
|Analyzing Data||Aug 15, 2016|
- Updates from LinuxCon and ContainerCon, Toronto, August 2016
- NVMe over Fabrics Support Coming to the Linux 4.8 Kernel
- What I Wish I’d Known When I Was an Embedded Linux Newbie
- New Version of GParted
- All about printf
- Analyzing Data
- Tor 0.2.8.6 Is Released
- Blender for Visual Effects
- Juniper Systems' Geode
With all the industry talk about the benefits of Linux on Power and all the performance advantages offered by its open architecture, you may be considering a move in that direction. If you are thinking about analytics, big data and cloud computing, you would be right to evaluate Power. The idea of using commodity x86 hardware and replacing it every three years is an outdated cost model. It doesn’t consider the total cost of ownership, and it doesn’t consider the advantage of real processing power, high-availability and multithreading like a demon.
This ebook takes a look at some of the practical applications of the Linux on Power platform and ways you might bring all the performance power of this open architecture to bear for your organization. There are no smoke and mirrors here—just hard, cold, empirical evidence provided by independent sources. I also consider some innovative ways Linux on Power will be used in the future.Get the Guide