Notepad/enter/Machine Tips (Quantum)/Resources/Code & Circuit Operations/Languages/More Misc/Julia.md

Julia is another high-performance fast programming language that could be great for future quantum programmers as an option to code in. There are several resources available online for learning quantum through julia as well.

• Error correction also exists for the Julia program - HOQST which is mainly an annealing tool but with the recent Nature paper proves that it may be better than qiskit pulse One great source is this article that explains how to build a quantum circuit simulator in one day!

using LinearAlgebra
function naive_broutine!(r::AbstractVector, U::AbstractMatrix, loc::Int)
    n = Int(log2(length(r))) # get the number of qubits
    return kron(I(1<<(n-loc+1)), U), I(1<<loc)
end

log2i(x::Int64) = !signbit(x) ? (63 - leading_zeros(x)) : throw(ErrorException("nonnegative expected ($x)"))
log2i(x::UInt64) = 63 - leading_zeros(x)

for N in [8, 16, 32, 64, 128]
    T = Symbol(:Int, N)
    UT = Symbol(:UInt, N)
    @eval begin
        log2i(x::$T) =
            !signbit(x) ? ($(N - 1) - leading_zeros(x)) :
            throw(ErrorException("nonnegative expected ($x)"))
        log2i(x::$UT) = $(N - 1) - leading_zeros(x)
    end
end


function broutine!(r::AbstractVector, U::AbstractMatrix, locs::NTuple{N, Int}) where N
end


function bmask(itr)
    isempty(itr) && return 0
    ret = 0
    for b in itr
        ret += 1 << (b - 1)
    end
    return ret
end


function bmask(range::UnitRange{Int})
    ((1 << (range.stop - range.start + 1)) - 1) << (range.start - 1)
end

function bmask(::Type{T}, itr) where {T<:Integer}
    isempty(itr) && return 0
    ret = zero(T)
    for b in itr
        ret += one(T) << (b - 1)
    end
    return ret
end

function bmask(::Type{T}, range::UnitRange{Int})::T where {T<:Integer}
    ((one(T) << (range.stop - range.start + 1)) - one(T)) << (range.start - 1)
end

bmask(args...) = bmask(Int, args...)
# this is for removing the infinity call of the later function
bmask(::Type{T}) where {T<:Integer} = zero(T)
bmask(::Type{T}, positions::Int...) where {T<:Integer} = bmask(T, positions)

bmask(args...) = bmask(Int, args...)
bmask(::Type{T}) where {T<:Integer} = zero(T)
bmask(::Type{T}, positions::Int...) where {T<:Integer} = bmask(T, positions)

function bmask(::Type{T}, itr) where {T<:Integer}
    isempty(itr) && return 0
    ret = zero(T)
    for b in itr
        ret += one(T) << (b - 1)
    end
    return ret
end

function bmask(::Type{T}, range::UnitRange{Int})::T where {T<:Integer}
    ((one(T) << (range.stop - range.start + 1)) - one(T)) << (range.start - 1)
end

(xxx & ~0b001) << 1 + (xxx & 0b001) # = xx00x

@inline lmove(b::Int, mask::Int, k::Int)::Int = (b & ~mask) << k + (b & mask)



function group_shift(locations)
    masks = Int[]
    region_lens = Int[]
    k_prv = -1
    for k in locations
        # if current position in the contiguous region
        # since these bits will be moved together with
        # the first one, we don't need to generate a
        # new mask
        if k == k_prv + 1
            region_lens[end] += 1
        else
            # we generate a bit mask where the 1st to k-th bits are 1
            push!(masks, bmask(0:k-1))
            push!(region_lens, 1)
        end
        k_prv = k
    end
    return masks, region_lens
end

• An interesting package written in Julia on Quantum Walks
  • was utilized in  Impact of the malicious input data modification on the efficiency of quantum algorithms, DOI:10.1007/s11128-019-2459-3, arXiv:1802.10041 (2018).