Exemplo n.º 1
0
// Euler method, can be used as solver.Step.
func (s *BackwardEuler) Step() {
	util.AssertMsg(MaxErr > 0, "Backward euler solver requires MaxErr > 0")

	t0 := Time

	y := M.Buffer()

	y0 := cuda.Buffer(VECTOR, y.Size())
	defer cuda.Recycle(y0)
	data.Copy(y0, y)

	dy0 := cuda.Buffer(VECTOR, y.Size())
	defer cuda.Recycle(dy0)
	if s.dy1 == nil {
		s.dy1 = cuda.Buffer(VECTOR, y.Size())
	}
	dy1 := s.dy1

	Dt_si = FixDt
	dt := float32(Dt_si * GammaLL)
	util.AssertMsg(dt > 0, "Backward Euler solver requires fixed time step > 0")

	// Fist guess
	Time = t0 + 0.5*Dt_si // 0.5 dt makes it implicit midpoint method

	// with temperature, previous torque cannot be used as predictor
	if Temp.isZero() {
		cuda.Madd2(y, y0, dy1, 1, dt) // predictor euler step with previous torque
		M.normalize()
	}

	torqueFn(dy0)
	cuda.Madd2(y, y0, dy0, 1, dt) // y = y0 + dt * dy
	M.normalize()

	// One iteration
	torqueFn(dy1)
	cuda.Madd2(y, y0, dy1, 1, dt) // y = y0 + dt * dy1
	M.normalize()

	Time = t0 + Dt_si

	err := cuda.MaxVecDiff(dy0, dy1) * float64(dt)

	// adjust next time step
	//if err < MaxErr || Dt_si <= MinDt || FixDt != 0 { // mindt check to avoid infinite loop
	// step OK
	NSteps++
	setLastErr(err)
	setMaxTorque(dy1)
	//} else {
	// undo bad step
	//	util.Assert(FixDt == 0)
	//	Time = t0
	//	data.Copy(y, y0)
	//	NUndone++
	//}
}
Exemplo n.º 2
0
func (mini *Minimizer) Step() {
	m := M.Buffer()
	size := m.Size()
	k := mini.k
	h := mini.h

	// save original magnetization
	m0 := cuda.Buffer(3, size)
	defer cuda.Recycle(m0)
	data.Copy(m0, m)

	// make descent
	cuda.Minimize(m, m0, k, h)

	// calculate new torque for next step
	k0 := cuda.Buffer(3, size)
	defer cuda.Recycle(k0)
	data.Copy(k0, k)
	torqueFn(k)
	setMaxTorque(k) // report to user

	// just to make the following readable
	dm := m0
	dk := k0

	// calculate step difference of m and k
	cuda.Madd2(dm, m, m0, 1., -1.)
	cuda.Madd2(dk, k, k0, -1., 1.) // reversed due to LLNoPrecess sign

	// get maxdiff and add to list
	max_dm := cuda.MaxVecNorm(dm)
	mini.lastDm.Add(max_dm)
	setLastErr(mini.lastDm.Max()) // report maxDm to user as LastErr

	// adjust next time step
	var nom, div float32
	if NSteps%2 == 0 {
		nom = cuda.Dot(dm, dm)
		div = cuda.Dot(dm, dk)
	} else {
		nom = cuda.Dot(dm, dk)
		div = cuda.Dot(dk, dk)
	}
	if div != 0. {
		mini.h = nom / div
	} else { // in case of division by zero
		mini.h = 1e-4
	}

	M.normalize()

	// as a convention, time does not advance during relax
	NSteps++
}
Exemplo n.º 3
0
// Euler method, can be used as solver.Step.
func (_ *Euler) Step() {
	y := M.Buffer()
	dy0 := cuda.Buffer(VECTOR, y.Size())
	defer cuda.Recycle(dy0)

	torqueFn(dy0)
	setMaxTorque(dy0)

	// Adaptive time stepping: treat MaxErr as the maximum magnetization delta
	// (proportional to the error, but an overestimation for sure)
	var dt float32
	if FixDt != 0 {
		Dt_si = FixDt
		dt = float32(Dt_si * GammaLL)
	} else {
		dt = float32(MaxErr / LastTorque)
		Dt_si = float64(dt) / GammaLL
	}
	util.AssertMsg(dt > 0, "Euler solver requires fixed time step > 0")
	setLastErr(float64(dt) * LastTorque)

	cuda.Madd2(y, y, dy0, 1, dt) // y = y + dt * dy
	M.normalize()
	Time += Dt_si
	NSteps++
}
Exemplo n.º 4
0
Arquivo: heun.go Projeto: kyeongdong/3
// Adaptive Heun method, can be used as solver.Step
func (_ *Heun) Step() {
	y := M.Buffer()
	dy0 := cuda.Buffer(VECTOR, y.Size())
	defer cuda.Recycle(dy0)

	if FixDt != 0 {
		Dt_si = FixDt
	}

	dt := float32(Dt_si * GammaLL)
	util.Assert(dt > 0)

	// stage 1
	torqueFn(dy0)
	cuda.Madd2(y, y, dy0, 1, dt) // y = y + dt * dy

	// stage 2
	dy := cuda.Buffer(3, y.Size())
	defer cuda.Recycle(dy)
	Time += Dt_si
	torqueFn(dy)

	err := cuda.MaxVecDiff(dy0, dy) * float64(dt)

	// adjust next time step
	if err < MaxErr || Dt_si <= MinDt || FixDt != 0 { // mindt check to avoid infinite loop
		// step OK
		cuda.Madd3(y, y, dy, dy0, 1, 0.5*dt, -0.5*dt)
		M.normalize()
		NSteps++
		adaptDt(math.Pow(MaxErr/err, 1./2.))
		setLastErr(err)
		setMaxTorque(dy)
	} else {
		// undo bad step
		util.Assert(FixDt == 0)
		Time -= Dt_si
		cuda.Madd2(y, y, dy0, 1, -dt)
		NUndone++
		adaptDt(math.Pow(MaxErr/err, 1./3.))
	}
}
Exemplo n.º 5
0
func (e *excitation) AddTo(dst *data.Slice) {
	if !e.perRegion.isZero() {
		cuda.RegionAddV(dst, e.perRegion.gpuLUT(), regions.Gpu())
	}

	for _, t := range e.extraTerms {
		var mul float32 = 1
		if t.mul != nil {
			mul = float32(t.mul())
		}
		cuda.Madd2(dst, dst, t.mask, 1, mul)
	}
}
Exemplo n.º 6
0
func (b *thermField) AddTo(dst *data.Slice) {
	if !Temp.isZero() {
		b.update()
		cuda.Madd2(dst, dst, b.noise, 1, 1)
	}
}
Exemplo n.º 7
0
Arquivo: rk23.go Projeto: kyeongdong/3
func (rk *RK23) Step() {
	m := M.Buffer()
	size := m.Size()

	if FixDt != 0 {
		Dt_si = FixDt
	}

	// upon resize: remove wrongly sized k1
	if rk.k1.Size() != m.Size() {
		rk.Free()
	}

	// first step ever: one-time k1 init and eval
	if rk.k1 == nil {
		rk.k1 = cuda.NewSlice(3, size)
		torqueFn(rk.k1)
	}

	// FSAL cannot be used with temperature
	if !Temp.isZero() {
		torqueFn(rk.k1)
	}

	t0 := Time
	// backup magnetization
	m0 := cuda.Buffer(3, size)
	defer cuda.Recycle(m0)
	data.Copy(m0, m)

	k2, k3, k4 := cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size)
	defer cuda.Recycle(k2)
	defer cuda.Recycle(k3)
	defer cuda.Recycle(k4)

	h := float32(Dt_si * GammaLL) // internal time step = Dt * gammaLL

	// there is no explicit stage 1: k1 from previous step

	// stage 2
	Time = t0 + (1./2.)*Dt_si
	cuda.Madd2(m, m, rk.k1, 1, (1./2.)*h) // m = m*1 + k1*h/2
	M.normalize()
	torqueFn(k2)

	// stage 3
	Time = t0 + (3./4.)*Dt_si
	cuda.Madd2(m, m0, k2, 1, (3./4.)*h) // m = m0*1 + k2*3/4
	M.normalize()
	torqueFn(k3)

	// 3rd order solution
	madd4(m, m0, rk.k1, k2, k3, 1, (2./9.)*h, (1./3.)*h, (4./9.)*h)
	M.normalize()

	// error estimate
	Time = t0 + Dt_si
	torqueFn(k4)
	Err := k2 // re-use k2 as error
	// difference of 3rd and 2nd order torque without explicitly storing them first
	madd4(Err, rk.k1, k2, k3, k4, (7./24.)-(2./9.), (1./4.)-(1./3.), (1./3.)-(4./9.), (1. / 8.))

	// determine error
	err := cuda.MaxVecNorm(Err) * float64(h)

	// adjust next time step
	if err < MaxErr || Dt_si <= MinDt || FixDt != 0 { // mindt check to avoid infinite loop
		// step OK
		setLastErr(err)
		setMaxTorque(k4)
		NSteps++
		Time = t0 + Dt_si
		adaptDt(math.Pow(MaxErr/err, 1./3.))
		data.Copy(rk.k1, k4) // FSAL
	} else {
		// undo bad step
		//util.Println("Bad step at t=", t0, ", err=", err)
		util.Assert(FixDt == 0)
		Time = t0
		data.Copy(m, m0)
		NUndone++
		adaptDt(math.Pow(MaxErr/err, 1./4.))
	}
}
Exemplo n.º 8
0
Arquivo: rk23.go Projeto: kyeongdong/3
// TODO: into cuda
func madd4(dst, src1, src2, src3, src4 *data.Slice, w1, w2, w3, w4 float32) {
	cuda.Madd3(dst, src1, src2, src3, w1, w2, w3)
	cuda.Madd2(dst, dst, src4, 1, w4)
}
Exemplo n.º 9
0
func madd6(dst, src1, src2, src3, src4, src5, src6 *data.Slice, w1, w2, w3, w4, w5, w6 float32) {
	madd5(dst, src1, src2, src3, src4, src5, w1, w2, w3, w4, w5)
	cuda.Madd2(dst, dst, src6, 1, w6)
}
Exemplo n.º 10
0
func (rk *RK45DP) Step() {
	m := M.Buffer()
	size := m.Size()

	if FixDt != 0 {
		Dt_si = FixDt
	}

	// upon resize: remove wrongly sized k1
	if rk.k1.Size() != m.Size() {
		rk.Free()
	}

	// first step ever: one-time k1 init and eval
	if rk.k1 == nil {
		rk.k1 = cuda.NewSlice(3, size)
		torqueFn(rk.k1)
	}

	// FSAL cannot be used with finite temperature
	if !Temp.isZero() {
		torqueFn(rk.k1)
	}

	t0 := Time
	// backup magnetization
	m0 := cuda.Buffer(3, size)
	defer cuda.Recycle(m0)
	data.Copy(m0, m)

	k2, k3, k4, k5, k6 := cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size)
	defer cuda.Recycle(k2)
	defer cuda.Recycle(k3)
	defer cuda.Recycle(k4)
	defer cuda.Recycle(k5)
	defer cuda.Recycle(k6)
	// k2 will be re-used as k7

	h := float32(Dt_si * GammaLL) // internal time step = Dt * gammaLL

	// there is no explicit stage 1: k1 from previous step

	// stage 2
	Time = t0 + (1./5.)*Dt_si
	cuda.Madd2(m, m, rk.k1, 1, (1./5.)*h) // m = m*1 + k1*h/5
	M.normalize()
	torqueFn(k2)

	// stage 3
	Time = t0 + (3./10.)*Dt_si
	cuda.Madd3(m, m0, rk.k1, k2, 1, (3./40.)*h, (9./40.)*h)
	M.normalize()
	torqueFn(k3)

	// stage 4
	Time = t0 + (4./5.)*Dt_si
	madd4(m, m0, rk.k1, k2, k3, 1, (44./45.)*h, (-56./15.)*h, (32./9.)*h)
	M.normalize()
	torqueFn(k4)

	// stage 5
	Time = t0 + (8./9.)*Dt_si
	madd5(m, m0, rk.k1, k2, k3, k4, 1, (19372./6561.)*h, (-25360./2187.)*h, (64448./6561.)*h, (-212./729.)*h)
	M.normalize()
	torqueFn(k5)

	// stage 6
	Time = t0 + (1.)*Dt_si
	madd6(m, m0, rk.k1, k2, k3, k4, k5, 1, (9017./3168.)*h, (-355./33.)*h, (46732./5247.)*h, (49./176.)*h, (-5103./18656.)*h)
	M.normalize()
	torqueFn(k6)

	// stage 7: 5th order solution
	Time = t0 + (1.)*Dt_si
	// no k2
	madd6(m, m0, rk.k1, k3, k4, k5, k6, 1, (35./384.)*h, (500./1113.)*h, (125./192.)*h, (-2187./6784.)*h, (11./84.)*h) // 5th
	M.normalize()
	k7 := k2     // re-use k2
	torqueFn(k7) // next torque if OK

	// error estimate
	Err := cuda.Buffer(3, size) //k3 // re-use k3 as error estimate
	defer cuda.Recycle(Err)
	madd6(Err, rk.k1, k3, k4, k5, k6, k7, (35./384.)-(5179./57600.), (500./1113.)-(7571./16695.), (125./192.)-(393./640.), (-2187./6784.)-(-92097./339200.), (11./84.)-(187./2100.), (0.)-(1./40.))

	// determine error
	err := cuda.MaxVecNorm(Err) * float64(h)

	// adjust next time step
	if err < MaxErr || Dt_si <= MinDt || FixDt != 0 { // mindt check to avoid infinite loop
		// step OK
		setLastErr(err)
		setMaxTorque(k7)
		NSteps++
		Time = t0 + Dt_si
		adaptDt(math.Pow(MaxErr/err, 1./5.))
		data.Copy(rk.k1, k7) // FSAL
	} else {
		// undo bad step
		//util.Println("Bad step at t=", t0, ", err=", err)
		util.Assert(FixDt == 0)
		Time = t0
		data.Copy(m, m0)
		NUndone++
		adaptDt(math.Pow(MaxErr/err, 1./6.))
	}
}
Exemplo n.º 11
0
Arquivo: rk4.go Projeto: kyeongdong/3
func (rk *RK4) Step() {
	m := M.Buffer()
	size := m.Size()

	if FixDt != 0 {
		Dt_si = FixDt
	}

	t0 := Time
	// backup magnetization
	m0 := cuda.Buffer(3, size)
	defer cuda.Recycle(m0)
	data.Copy(m0, m)

	k1, k2, k3, k4 := cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size), cuda.Buffer(3, size)

	defer cuda.Recycle(k1)
	defer cuda.Recycle(k2)
	defer cuda.Recycle(k3)
	defer cuda.Recycle(k4)

	h := float32(Dt_si * GammaLL) // internal time step = Dt * gammaLL

	// stage 1
	torqueFn(k1)

	// stage 2
	Time = t0 + (1./2.)*Dt_si
	cuda.Madd2(m, m, k1, 1, (1./2.)*h) // m = m*1 + k1*h/2
	M.normalize()
	torqueFn(k2)

	// stage 3
	cuda.Madd2(m, m0, k2, 1, (1./2.)*h) // m = m0*1 + k2*1/2
	M.normalize()
	torqueFn(k3)

	// stage 4
	Time = t0 + Dt_si
	cuda.Madd2(m, m0, k3, 1, 1.*h) // m = m0*1 + k3*1
	M.normalize()
	torqueFn(k4)

	err := cuda.MaxVecDiff(k1, k4) * float64(h)

	// adjust next time step
	if err < MaxErr || Dt_si <= MinDt || FixDt != 0 { // mindt check to avoid infinite loop
		// step OK
		// 4th order solution
		madd5(m, m0, k1, k2, k3, k4, 1, (1./6.)*h, (1./3.)*h, (1./3.)*h, (1./6.)*h)
		M.normalize()
		NSteps++
		adaptDt(math.Pow(MaxErr/err, 1./4.))
		setLastErr(err)
		setMaxTorque(k4)
	} else {
		// undo bad step
		//util.Println("Bad step at t=", t0, ", err=", err)
		util.Assert(FixDt == 0)
		Time = t0
		data.Copy(m, m0)
		NUndone++
		adaptDt(math.Pow(MaxErr/err, 1./5.))
	}
}