lattice/latticegpu.jl
1
0
Fork 0
mirror of https://igit.ific.uv.es/alramos/latticegpu.jl.git synced 2025-06-29 05:19:27 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'fix/flow_obc' of https://igit.ific.uv.es/langn/lattice-gpu

Browse source
This commit is contained in:
Fernando P. Panadero 2025-05-27 10:25:56 +02:00
commit 4326c34621
5 changed files with 54 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

1
docs/make.jl
View file

@ -13,6 +13,7 @@ makedocs(sitename="LatticeGPU", modules=[LatticeGPU], doctest=true,
"Yang-Mills" => "ym.md",
"Gradient flow" => "flow.md",
"Schrödinger Functional" => "sf.md",
"Spinors" => "spinors.md",
"Dirac" => "dirac.md",
"Solvers" => "solvers.md",
"Input Output" => "io.md"

2
src/YM/YM.jl
View file

@ -165,7 +165,7 @@ include("YMfields.jl")
export randomize!, zero!, norm2
include("YMact.jl")
export krnl_plaq!, force_gauge, force_gauge_flw, force_wilson
export krnl_plaq!, force_gauge, force_gauge_flw, force_wilson, bnd_rescale_flw!
include("YMhmc.jl")
export gauge_action, hamiltonian, plaquette, HMC!, MD!

9
src/YM/YMfields.jl
View file

@ -9,16 +9,17 @@
### created: Thu Jul 15 15:16:47 2021
###
"""
function randomize!(f, lp::SpaceParm, ymws::YMworkspace)
function randomize!(f, lp::SpaceParm, ymws::YMworkspace; curng=CUDA.default_rng())
Given an algebra field with natural indexing, this routine sets the components to random Gaussian distributed values. If SF boundary conditions are used, the force at the boundaries is set to zero.
"""
function randomize!(f, lp::SpaceParm, ymws::YMworkspace)
function randomize!(f, lp::SpaceParm, ymws::YMworkspace; curng=CUDA.default_rng())
if ymws.ALG == SU2alg
@timeit "Randomize SU(2) algebra field" begin
m = CUDA.randn(ymws.PRC, lp.bsz,lp.ndim,3,lp.rsz)
m = Random.randn(curng, ymws.PRC, lp.bsz,lp.ndim,3,lp.rsz)
CUDA.@sync begin
CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_assign_SU2!(f,m,lp)
end
@ -28,7 +29,7 @@ function randomize!(f, lp::SpaceParm, ymws::YMworkspace)
if ymws.ALG == SU3alg
@timeit "Randomize SU(3) algebra field" begin
m = CUDA.randn(ymws.PRC, lp.bsz,lp.ndim,8,lp.rsz)
m = Random.randn(curng, ymws.PRC, lp.bsz,lp.ndim,8,lp.rsz)
CUDA.@sync begin
CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_assign_SU3!(f,m,lp)
end

42
src/YM/YMflow.jl
View file

@ -127,6 +127,22 @@ function add_zth_term(ymws::YMworkspace, U, lp)
return nothing
end
"""
function rescale_bnd(ymws::YMworkspace, lp)
Rescales the time-boundary of ymws.frc1 by a factor of 2.0 (needed for open BC)
"""
function rescale_bnd(ymws::YMworkspace, lp)
CUDA.@sync begin
CUDA.@cuda threads=lp.bsz blocks=lp.rsz bnd_rescale_flw!(ymws.frc1,lp::SpaceParm)
end
return nothing
end
function krnl_add_zth!(frc, frc2::AbstractArray{TA}, U::AbstractArray{TG}, lp::SpaceParm{N,M,B,D}) where {TA,TG,N,M,B,D}
@inbounds begin
@ -155,12 +171,14 @@ function krnl_add_zth!(frc, frc2::AbstractArray{TA}, U::AbstractArray{TG}, lp::S
end
end
if OBC
if (it > 1) && (it < lp.iL[end])
if (id < N)
frc2[b,id,r] = (5/6)*frc[b,id,r] + (1/6)*(projalg(Ud\Y*Ud) +
projalg(U[b,id,r]*X/U[b,id,r]))
elseif ((it == lp.iL[end]) || (it == 1)) && (id < N)
elseif (it > 1) && (it < (lp.iL[end]-1))
frc2[b,id,r] = (5/6)*frc[b,id,r] + (1/6)*(projalg(Ud\Y*Ud) +
projalg(U[b,id,r]*X/U[b,id,r]))
elseif (it == 1) || (it == (lp.iL[end]-1))
frc2[b,id,r] = frc[b,id,r]
end
else
frc2[b,id,r] = (5/6)*frc[b,id,r] + (1/6)*(projalg(Ud\Y*Ud) +
@ -171,6 +189,8 @@ function krnl_add_zth!(frc, frc2::AbstractArray{TA}, U::AbstractArray{TG}, lp::S
return nothing
end
"""
function flw(U, int::FlowIntr{NI,T}, ns::Int64, gp::GaugeParm, lp::SpaceParm, ymws::YMworkspace)
@ -204,18 +224,20 @@ flw(U, int::FlowIntr{NI,T}, ns::Int64, gp::GaugeParm, lp::SpaceParm, ymws::YMwor
function flw(U, int::FlowIntr{NI,T}, ns::Int64, eps, gp::GaugeParm, lp::SpaceParm{N,M,BC_OPEN,D}, ymws::YMworkspace) where {NI,T,N,M,D}
@timeit "Integrating flow equations" begin
for i in 1:ns
force_gauge_flw(ymws, U, int.c0, 1, gp, lp)
force_gauge(ymws, U, int.c0, 1, gp, lp)
if int.add_zth
add_zth_term(ymws::YMworkspace, U, lp)
end
rescale_bnd(ymws, lp)
ymws.mom .= ymws.frc1
U .= expm.(U, ymws.mom, 2*eps*int.r)
for k in 1:NI
force_gauge_flw(ymws, U, int.c0, 1, gp, lp)
force_gauge(ymws, U, int.c0, 1, gp, lp)
if int.add_zth
add_zth_term(ymws::YMworkspace, U, lp)
end
rescale_bnd(ymws, lp)
ymws.mom .= int.e0[k].*ymws.mom .+ int.e1[k].*ymws.frc1
U .= expm.(U, ymws.mom, 2*eps)
end
@ -328,6 +350,10 @@ function Eoft_plaq(Eslc, U, gp::GaugeParm{T,G,NN}, lp::SpaceParm{N,M,B,D}, ymws:
for it in 1:lp.iL[end]
Eslc[it,ipl] = 2*Etmp[it]
end
if OBC ## Spatial plaquettes at time boundary count half (Luescher)
Eslc[1,ipl] = Etmp[1]
Eslc[end,ipl] = Etmp[end]
end
end
end
@ -426,16 +452,24 @@ function Eoft_clover(Eslc, U, gp::GaugeParm, lp::SpaceParm{4,M,B,D}, ymws::YMwor
end
Etmp .= reshape(Array(CUDA.reduce(+, ymws.rm;dims=tp)),lp.iL[end])/V3
for it in 1:lp.iL[end]
if (B == BC_OPEN) && (it == 1 || it == lp.iL[end])
Eslc[it,ipl1] = 0.0
else
Eslc[it,ipl1] = Etmp[it]/8
end
end
CUDA.@sync begin
CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_add_et!(ymws.rm, ymws.frc2, lp)
end
Etmp .= reshape(Array(CUDA.reduce(+, ymws.rm;dims=tp)),lp.iL[end])/V3
for it in 1:lp.iL[end]
if (B == BC_OPEN) && (it == 1 || it == lp.iL[end])
Eslc[it,ipl2] = 0.0
else
Eslc[it,ipl2] = Etmp[it]/8
end
end
return nothing
end

11
src/YM/YMhmc.jl
View file

@ -68,18 +68,19 @@ function hamiltonian(mom, U, lp, gp, ymws)
return K+V
end
"""
HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace; noacc=false)
HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace; noacc=false, rng=Random.default_rng(), curng=CUDA.default_rng())
Performs a HMC step (molecular dynamics integration and accept/reject step). The configuration `U` is updated and function returns the energy violation and if the configuration was accepted in a tuple.
"""
function HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace{T}; noacc=false) where T
function HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace{T}; noacc=false, rng=Random.default_rng(), curng=CUDA.default_rng()) where T
@timeit "HMC trayectory" begin
ymws.U1 .= U
randomize!(ymws.mom, lp, ymws)
randomize!(ymws.mom, lp, ymws; curng)
hini = hamiltonian(ymws.mom, U, lp, gp, ymws)
MD!(ymws.mom, U, int, lp, gp, ymws)
@ -93,7 +94,7 @@ function HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspac
end
if (pacc < 1.0)
r = rand()
r = rand(rng)
if (pacc < r)
U .= ymws.U1
acc = false
@ -105,7 +106,7 @@ function HMC!(U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspac
end
return dh, acc
end
HMC!(U, eps, ns, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace{T}; noacc=false) where T = HMC!(U, omf4(T, eps, ns), lp, gp, ymws; noacc=noacc)
HMC!(U, eps, ns, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace{T}; noacc=false, rng=Random.default_rng(), curng=CUDA.default_rng()) where T = HMC!(U, omf4(T, eps, ns), lp, gp, ymws; noacc=noacc, rng, curng)
"""
function MD!(mom, U, int::IntrScheme, lp::SpaceParm, gp::GaugeParm, ymws::YMworkspace)