Gatery Challenge

In this coding challenge, the goal is to implement a stream module that sums integer numbers. A function receives a stream of unsigned integer numbers with ready-valid handshake signals. The function is to sum N consecutive numbers and output these sums as an output stream, again with ready-valid handshake signals. In other words, for every N transfers on the input stream there is to be one transfer (of the sum) on the output stream.

For details, find the function RvStream<UInt> sum_N_numbers(RvStream<UInt> &inStream, size_t N) in the source/main.cpp file.

Getting Started

The easiest way is to solve this project in a github codespace. You will need a github account and the automated process to set up the VM takes about 20 minutes.

Alternatively, you can clone this repository locally and follow the instructions in the main gatery repository to set up the dependencies.

Assuming you launched in a github codespace, you will eventually be presented with a browser based instance of VS Code. Any changes made to files will persist in the Codespace until it is deleted. A Codespace can be restarted from the "Codespaces" tab on github and these subsequent restarts do not require rebuilding the VM.

Files can be downloaded by right-clicking them in the Explorer view and selecting "Download". Additionally, changes can be committed and pushed from the Source Control view as in a regular git repository (in case you want to work on a cloned repository or add your repository as an additional remote).

Press F5 to compile and run. Any compilation errors will be shown in the Terminal tab, as well as (mixed with other less important "problems") in the Problems tab. In the absence of compilation errors, the code will run and simulate. The Output tab will show information such as detected failures.

A .vcd waveform of the simulation will be written to bin/linux-x86_64-Debug/waveform.vcd. It can be downloaded and inspected locally, or opened in VS Code by double clicking (does not work in firefox).

Gatery Syntax Cheat Sheet

This is a list of the most important syntax constructs in gatery. See the documentation or tutorial for details.

Elementary Signals

Type	Description	Width	VHDL	Verilog
Bit	single boolean or bit	1	std_logic	wire
BVec	untyped bit vector	variable, >= 0	std_logic_vector	-
SInt	signed integer	variable, >= 0	signed	wire signed [...]
UInt	unsigned integer	variable, >= 0	unsigned	wire[...]
Enum<c>	Enum type signal based on a C/C++ enum c	Depends on c	custom type	-

Initialization

	UInt undefined_8_wide_uint = 8_b;
	UInt undefined_10_wide_uint = BitWidth(10);
	UInt undefined_12_wide_uint(12_b);
	UInt undefined_16_wide_uint;
	undefined_16_wide_uint = BitWidth(16);
	
	UInt another_16_wide_uint = undefined_16_wide_uint.width(); // Only copies the width, but does not connect the signals.

Signal Naming

    UInt v = 32_b;
    v = ...;

    // name the current state of v "initial_v" in the VHDL code and in the waveform
    setName(v, "initial_v");

    v += 2;

    // name this new, incremented state of v "modified_v" in the VHDL code and in the waveform
    setName(v, "modified_v");

    UInt some_other_vector = v;

    // name some_other_vector "some_other_vector" in the VHDL code and in the waveform
    HCL_NAMED(some_other_vector);

Literals

	Bit b;

	b = '1'; // true
	b = '0'; // false
	b = 'X'; // undefined

	b = true;  // true
	b = false; // false

	UInt bv_1, bv_2, bv_3, bv_4, bv_5;

	bv_1 = "b1010"; // Binary, 4_b wide
	bv_2 = "xff0f0"; // Hex, 20_b wide
	bv_3 = "d42"; // Decimal, 6_b wide
	bv_4 = "64b0"; // 64 zero bits
	bv_5 = "6b00xx00"; // Mixture of zeros and undefined bits

Bit vector member functions

	UInt vec = "32b0";

Method	Type	Description
vec.width()	BitWidth	Width of the vector (i.e. 32).
vec.size()	size_t	Width of the vector (i.e. 32).
vec[i]	Bit	ith bit, assignable.
vec.lsb()	Bit	Least significant bit, assignable. Same as vec[0].
vec.msb()	Bit	Most significant bit, assignable. Same as vec[-1].
vec(offset, size)	UInt	Slice from bit offset to bit offset+size, assignable.
vec.lower(size)	UInt	Lower size bits, assignable. Same as vec(0, size).
vec.upper(size)	UInt	Upper size bits, assignable.

Logical and Arithmetic Operators

    Bit a, b;
    a = ...;
    b = ...;

    // Logical and bitwise negation both do the same
    Bit not_a = ~a;
    Bit also_not_a = !a;

    // And, or, xor as usual
    Bit a_and_b = a & b;
    Bit a_or_b = a | b;
    Bit a_xor_b = a ^ b;

    // Composition and bracketing as usual
    Bit a_nand_b = ~(a & b);
    Bit a_nor_b = ~(a | b);

    UInt a = 8_b;
    UInt b = 8_b;
    a = ...;
    b = ...;

    // Bitwise negation
    UInt not_a = ~a;

    // And, or, xor as usual
    UInt a_and_b = a & b;
    UInt a_or_b = a | b;
    UInt a_xor_b = a ^ b;

    // Composition and bracketing as usual
    UInt a_nand_b = ~(a & b);
    UInt a_nor_b = ~(a | b);

    UInt c = 5_b;
    c = ...;
    // UInt illegal = a & c; <-- Illegal because c is 5 bits and a is 8 bits

    UInt a_plus_b = a + b;
    UInt a_minus_b = a - b;
    UInt a_times_b = a * b;
    UInt a_div_b = a / b;

    // Less / greater
    Bit a_lt_b = a < b;
    Bit a_gt_b = a > b;
    
    // Less or equal / greater or equal
    Bit a_le_b = a <= b;
    Bit a_ge_b = a >= b;

    // Equal / not equal
    Bit a_eq_b = a == b;
    Bit a_ne_b = a != b;    

    UInt mantissa = 23_b;
    mantissa = ...;
    UInt exponent = 8_b;
    exponent = ...;
    Bit sign = ...;

    // Concatenates all arguments, putting the last
    // argument (mantissa) into the least significant bits.
    UInt ieee_float_32 = cat(sign, exponent, mantissa);

    // Packs all arguments, putting the first
    // argument (mantissa) into the least significant bits.
    UInt same_ieee_float_32 = pack(mantissa, exponent, sign);

    UInt value = 10_b;
    value = ...;

    UInt value_times_4 = value << 2;
    UInt value_div_4 = value >> 2;
    UInt value_rotated_left_2 = rotl(value, 2);
    UInt value_rotated_right_2 = rotr(value, 2);    

Vector extension

	UInt unsigned_8_wide = "8b0";
	// Zero extends by 2 bits
	UInt unsigned_10_wide = ext(unsigned_8_wide, +2_b);

	SInt signed_8_wide = (SInt) "8b0";
	// Sign extends by 2 bits
	SInt signed_10_wide = ext(signed_8_wide, +2_b);

	UInt unsigned_8_wide = "8b0";
	// Zero extend unsigned integer
	UInt unsigned_10_wide = zext(unsigned_8_wide, +2_b);

	UInt signed_8_wide = "8b0";
	// Sign extend integer
	UInt signed_10_wide = sext(signed_8_wide, +2_b);

	UInt mask_8_wide = "8b0";
	// Extend with ones
	UInt mask_10_wide_one_extended = oext(mask_8_wide, +2_b);

    Bit bit = '1';
    // Sign extends by 9 bits
    UInt ten_1 = sext(bit, +9_b);

    UInt a = "10b0";
    UInt b = "8b0";

    // This would be illegal because a nd b have different sizes:
    // UInt c = a & b;

    // This zero-extends b to the width of a (10-bits) and then performs the element wise or
    UInt a_or_b = a | zext(b);
    
    // The same works for sext and oext.
    UInt a_and_b = a & oext(b);

Condition Scopes / Multiplexer

    UInt value = 4_b;
    value = ...;

    Bit some_condition = ...;

    IF (some_condition) {
        value <<= 1;
    } ELSE {
        value += 1;
    }

Clock Scopes

void functionA()
{
	// registers, memory, and code created here can automatically
	// use clockA by accessing it from the innermost active ClockScope
}

void functionB()
{
	// Build a clock
	Clock clockA{{.absoluteFrequency = 1'000'000}}; // 1 MHz
	// Use this clock while the clockScope variable exists
	ClockScope clockScope{ clockA };

	functionA();
}

Registers

    UInt a = ...;
    a_delayed = reg(a);

    Bit b = ...;
    b_delayed_w_reset = reg(b, '0'); // reset to zero

    // Use loop semantics to build state:
    Bit enableCounter = ...;
    UInt counter_10b = 10_b;
    IF (enableCounter)
        counter_10b += 1;
    counter_10b = reg(counter_10b, 0);    

Further Information

For more information, visit the Gatery website, where a tutorial is also available.