Also applied the merged fix to nbnxn_cuda_kernel_fermi.cuh.
Change-Id: I79d0cb59214290097d6f356e35dd13c4eadaed75
* NOTE: No include fence as it is meant to be included multiple times.
*
* \author Szilárd Páll <pall.szilard@gmail.com>
+ * \author Berk Hess <hess@kth.se>
* \ingroup module_mdlib
*/
-#include "config.h"
+#include "gromacs/gpu_utils/cuda_arch_utils.cuh"
#include "gromacs/math/utilities.h"
#include "gromacs/pbcutil/ishift.h"
/* Note that floating-point constants in CUDA code should be suffixed
* code that is in double precision.
*/
-#if __CUDA_ARCH__ >= 300
-/* Note: convenience macros, need to be undef-ed at the end of the file. */
-#define REDUCE_SHUFFLE
-/* On Kepler pre-loading i-atom types to shmem gives a few %,
- but on Fermi it does not */
-#define IATYPE_SHMEM
-#ifdef HAVE_CUDA_TEXOBJ_SUPPORT
-#define USE_TEXOBJ
-#endif
+#if GMX_PTX_ARCH < 300
+#error "nbnxn_cuda_kernel.cuh included with GMX_PTX_ARCH < 300"
#endif
#if defined EL_EWALD_ANA || defined EL_EWALD_TAB
#define LJ_EWALD
#endif
+#if defined LJ_COMB_GEOM || defined LJ_COMB_LB
+#define LJ_COMB
+#endif
/*
Kernel launch parameters:
- #blocks = #pair lists, blockId = pair list Id
- - #threads = NTHREAD_Z * CL_SIZE^2
+ - #threads = NTHREAD_Z * c_clSize^2
- shmem = see nbnxn_cuda.cu:calc_shmem_required()
Each thread calculates an i force-component taking one pair of i-j atoms.
*/
-/* NTHREAD_Z controls the number of j-clusters processed concurrently on NTHREAD_Z
+/**@{*/
+/*! \brief Compute capability dependent definition of kernel launch configuration parameters.
+ *
+ * NTHREAD_Z controls the number of j-clusters processed concurrently on NTHREAD_Z
* warp-pairs per block.
*
* - On CC 2.0-3.5, 5.0, and 5.2, NTHREAD_Z == 1, translating to 64 th/block with 16
* shuffle-based reduction, hence CC >= 3.0.
*/
-/* Kernel launch bounds as function of NTHREAD_Z.
+/* Kernel launch bounds for different compute capabilities. The value of NTHREAD_Z
+ * determines the number of threads per block and it is chosen such that
+ * 16 blocks/multiprocessor can be kept in flight.
* - CC 3.0/3.5/5.x, >=6.1: NTHREAD_Z=1, (64, 16) bounds
* - CC 3.7, 6.0: NTHREAD_Z=2, (128, 16) bounds
*
* Note: convenience macros, need to be undef-ed at the end of the file.
*/
-#if __CUDA_ARCH__ == 370 || __CUDA_ARCH__ == 600
-#define NTHREAD_Z (2)
-#define MIN_BLOCKS_PER_MP (16)
+#if GMX_PTX_ARCH == 370 || GMX_PTX_ARCH == 600
+ #define NTHREAD_Z (2)
+ #define MIN_BLOCKS_PER_MP (16)
#else
-#define NTHREAD_Z (1)
-#define MIN_BLOCKS_PER_MP (16)
-#endif
-#define THREADS_PER_BLOCK (CL_SIZE*CL_SIZE*NTHREAD_Z)
+ #define NTHREAD_Z (1)
+ #define MIN_BLOCKS_PER_MP (16)
+#endif /* GMX_PTX_ARCH == 370 || GMX_PTX_ARCH == 600 */
+#define THREADS_PER_BLOCK (c_clSize*c_clSize*NTHREAD_Z)
-#if __CUDA_ARCH__ >= 350
+#if GMX_PTX_ARCH >= 350
+#if (GMX_PTX_ARCH <= 210) && (NTHREAD_Z > 1)
+ #error NTHREAD_Z > 1 will give incorrect results on CC 2.x
+#endif
+/**@}*/
__launch_bounds__(THREADS_PER_BLOCK, MIN_BLOCKS_PER_MP)
#else
__launch_bounds__(THREADS_PER_BLOCK)
-#endif
+#endif /* GMX_PTX_ARCH >= 350 */
#ifdef PRUNE_NBL
#ifdef CALC_ENERGIES
__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _VF_prune_cuda)
#else
__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_prune_cuda)
-#endif
+#endif /* CALC_ENERGIES */
#else
#ifdef CALC_ENERGIES
__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _VF_cuda)
#else
__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_cuda)
-#endif
-#endif
+#endif /* CALC_ENERGIES */
+#endif /* PRUNE_NBL */
(const cu_atomdata_t atdat,
const cu_nbparam_t nbparam,
const cu_plist_t plist,
bool bCalcFshift)
+#ifdef FUNCTION_DECLARATION_ONLY
+; /* Only do function declaration, omit the function body. */
+#else
{
/* convenience variables */
const nbnxn_sci_t *pl_sci = plist.sci;
#endif
nbnxn_cj4_t *pl_cj4 = plist.cj4;
const nbnxn_excl_t *excl = plist.excl;
+#ifndef LJ_COMB
const int *atom_types = atdat.atom_types;
int ntypes = atdat.ntypes;
+#else
+ const float2 *lj_comb = atdat.lj_comb;
+ float2 ljcp_i, ljcp_j;
+#endif
const float4 *xq = atdat.xq;
float3 *f = atdat.f;
const float3 *shift_vec = atdat.shift_vec;
unsigned int tidxz = threadIdx.z;
#endif
unsigned int bidx = blockIdx.x;
- unsigned int widx = tidx / WARP_SIZE; /* warp index */
+ unsigned int widx = tidx / warp_size; /* warp index */
- int sci, ci, cj, ci_offset,
+ int sci, ci, cj,
ai, aj,
- cij4_start, cij4_end,
- typei, typej,
- i, jm, j4, wexcl_idx;
+ cij4_start, cij4_end;
+#ifndef LJ_COMB
+ int typei, typej;
+#endif
+ int i, jm, j4, wexcl_idx;
float qi, qj_f,
- r2, inv_r, inv_r2, inv_r6,
- c6, c12,
- int_bit,
+ r2, inv_r, inv_r2;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ float inv_r6, c6, c12;
+#endif
+#ifdef LJ_COMB_LB
+ float sigma, epsilon;
+#endif
+ float int_bit,
F_invr;
#ifdef CALC_ENERGIES
float E_lj, E_el;
unsigned int wexcl, imask, mask_ji;
float4 xqbuf;
float3 xi, xj, rv, f_ij, fcj_buf;
- float3 fci_buf[NCL_PER_SUPERCL]; /* i force buffer */
+ float3 fci_buf[c_numClPerSupercl]; /* i force buffer */
nbnxn_sci_t nb_sci;
+ /*! i-cluster interaction mask for a super-cluster with all c_numClPerSupercl=8 bits set */
+ const unsigned superClInteractionMask = ((1U << c_numClPerSupercl) - 1U);
+
/* shmem buffer for i x+q pre-loading */
extern __shared__ float4 xqib[];
+
/* shmem buffer for cj, for each warp separately */
- int *cjs = ((int *)(xqib + NCL_PER_SUPERCL * CL_SIZE)) + tidxz * 2 * NBNXN_GPU_JGROUP_SIZE;
-#ifdef IATYPE_SHMEM
+ int *cjs = ((int *)(xqib + c_numClPerSupercl * c_clSize)) + tidxz * c_nbnxnGpuClusterpairSplit * c_nbnxnGpuJgroupSize;
+ int *cjs_end = ((int *)(xqib + c_numClPerSupercl * c_clSize)) + NTHREAD_Z * c_nbnxnGpuClusterpairSplit * c_nbnxnGpuJgroupSize;
+#ifndef LJ_COMB
/* shmem buffer for i atom-type pre-loading */
- int *atib = ((int *)(xqib + NCL_PER_SUPERCL * CL_SIZE)) + NTHREAD_Z * 2 * NBNXN_GPU_JGROUP_SIZE;
-#endif
-
-#ifndef REDUCE_SHUFFLE
- /* shmem j force buffer */
-#ifdef IATYPE_SHMEM
- float *f_buf = (float *)(atib + NCL_PER_SUPERCL * CL_SIZE);
+ int *atib = cjs_end;
#else
- float *f_buf = (float *)(cjs + NTHREAD_Z * 2 * NBNXN_GPU_JGROUP_SIZE);
-#endif
+ /* shmem buffer for i-atom LJ combination rule parameters */
+ float2 *ljcpib = (float2 *)cjs_end;
#endif
nb_sci = pl_sci[bidx]; /* my i super-cluster's index = current bidx */
if (tidxz == 0)
{
/* Pre-load i-atom x and q into shared memory */
- ci = sci * NCL_PER_SUPERCL + tidxj;
- ai = ci * CL_SIZE + tidxi;
- xqib[tidxj * CL_SIZE + tidxi] = xq[ai] + shift_vec[nb_sci.shift];
-#ifdef IATYPE_SHMEM
+ ci = sci * c_numClPerSupercl + tidxj;
+ ai = ci * c_clSize + tidxi;
+
+ xqbuf = xq[ai] + shift_vec[nb_sci.shift];
+ xqbuf.w *= nbparam.epsfac;
+ xqib[tidxj * c_clSize + tidxi] = xqbuf;
+
+#ifndef LJ_COMB
/* Pre-load the i-atom types into shared memory */
- atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+ atib[tidxj * c_clSize + tidxi] = atom_types[ai];
+#else
+ /* Pre-load the LJ combination parameters into shared memory */
+ ljcpib[tidxj * c_clSize + tidxi] = lj_comb[ai];
#endif
}
__syncthreads();
- for (ci_offset = 0; ci_offset < NCL_PER_SUPERCL; ci_offset++)
+ for (i = 0; i < c_numClPerSupercl; i++)
{
- fci_buf[ci_offset] = make_float3(0.0f);
+ fci_buf[i] = make_float3(0.0f);
}
#ifdef LJ_EWALD
/* TODO: we are trading registers with flops by keeping lje_coeff-s, try re-calculating it later */
lje_coeff2 = nbparam.ewaldcoeff_lj*nbparam.ewaldcoeff_lj;
- lje_coeff6_6 = lje_coeff2*lje_coeff2*lje_coeff2*ONE_SIXTH_F;
-#endif /* LJ_EWALD */
+ lje_coeff6_6 = lje_coeff2*lje_coeff2*lje_coeff2*c_oneSixth;
+#endif
#ifdef CALC_ENERGIES
E_lj = 0.0f;
E_el = 0.0f;
-#if defined EXCLUSION_FORCES /* Ewald or RF */
- if (nb_sci.shift == CENTRAL && pl_cj4[cij4_start].cj[0] == sci*NCL_PER_SUPERCL)
+#ifdef EXCLUSION_FORCES /* Ewald or RF */
+ if (nb_sci.shift == CENTRAL && pl_cj4[cij4_start].cj[0] == sci*c_numClPerSupercl)
{
/* we have the diagonal: add the charge and LJ self interaction energy term */
- for (i = 0; i < NCL_PER_SUPERCL; i++)
+ for (i = 0; i < c_numClPerSupercl; i++)
{
#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
- qi = xqib[i * CL_SIZE + tidxi].w;
+ qi = xqib[i * c_clSize + tidxi].w;
E_el += qi*qi;
#endif
-#if defined LJ_EWALD
-#ifdef USE_TEXOBJ
- E_lj += tex1Dfetch<float>(nbparam.nbfp_texobj, atom_types[(sci*NCL_PER_SUPERCL + i)*CL_SIZE + tidxi]*(ntypes + 1)*2);
-#else
- E_lj += tex1Dfetch(nbfp_texref, atom_types[(sci*NCL_PER_SUPERCL + i)*CL_SIZE + tidxi]*(ntypes + 1)*2);
-#endif /* USE_TEXOBJ */
-#endif /* LJ_EWALD */
-
+#ifdef LJ_EWALD
+ E_lj += tex1Dfetch<float>(nbparam.nbfp_texobj, atom_types[(sci*c_numClPerSupercl + i)*c_clSize + tidxi]*(ntypes + 1)*2);
+#endif
}
/* divide the self term(s) equally over the j-threads, then multiply with the coefficients. */
#ifdef LJ_EWALD
- E_lj /= CL_SIZE*NTHREAD_Z;
- E_lj *= 0.5f*ONE_SIXTH_F*lje_coeff6_6;
-#endif /* LJ_EWALD */
+ E_lj /= c_clSize*NTHREAD_Z;
+ E_lj *= 0.5f*c_oneSixth*lje_coeff6_6;
+#endif
#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
- E_el /= CL_SIZE*NTHREAD_Z;
+ /* Correct for epsfac^2 due to adding qi^2 */
+ E_el /= nbparam.epsfac*c_clSize*NTHREAD_Z;
#if defined EL_RF || defined EL_CUTOFF
- E_el *= -nbparam.epsfac*0.5f*c_rf;
+ E_el *= -0.5f*c_rf;
#else
- E_el *= -nbparam.epsfac*beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
+ E_el *= -beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
#endif
-#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
+#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
}
-#endif /* EXCLUSION_FORCES */
-
-#endif /* CALC_ENERGIES */
+#endif /* EXCLUSION_FORCES */
- /* skip central shifts when summing shift forces */
- if (nb_sci.shift == CENTRAL)
- {
- bCalcFshift = false;
- }
+#endif /* CALC_ENERGIES */
/* loop over the j clusters = seen by any of the atoms in the current super-cluster */
for (j4 = cij4_start + tidxz; j4 < cij4_end; j4 += NTHREAD_Z)
{
wexcl_idx = pl_cj4[j4].imei[widx].excl_ind;
imask = pl_cj4[j4].imei[widx].imask;
- wexcl = excl[wexcl_idx].pair[(tidx) & (WARP_SIZE - 1)];
+ wexcl = excl[wexcl_idx].pair[(tidx) & (warp_size - 1)];
#ifndef PRUNE_NBL
if (imask)
#endif
{
/* Pre-load cj into shared memory on both warps separately */
- if ((tidxj == 0 || tidxj == 4) && tidxi < NBNXN_GPU_JGROUP_SIZE)
+ if ((tidxj == 0 || tidxj == 4) && tidxi < c_nbnxnGpuJgroupSize)
{
- cjs[tidxi + tidxj * NBNXN_GPU_JGROUP_SIZE / 4] = pl_cj4[j4].cj[tidxi];
+ cjs[tidxi + tidxj * c_nbnxnGpuJgroupSize/c_splitClSize] = pl_cj4[j4].cj[tidxi];
}
/* Unrolling this loop
- with pruning leads to register spilling;
- - on Kepler is much slower;
- - doesn't work on CUDA <v4.1
- Tested with nvcc 3.2 - 5.0.7 */
-#if !defined PRUNE_NBL && __CUDA_ARCH__ < 300 && GMX_CUDA_VERSION >= 4010
-#pragma unroll 4
-#endif
- for (jm = 0; jm < NBNXN_GPU_JGROUP_SIZE; jm++)
+ - on Kepler and later it is much slower;
+ Tested with up to nvcc 7.5 */
+ for (jm = 0; jm < c_nbnxnGpuJgroupSize; jm++)
{
- /* ((1U << NCL_PER_SUPERCL) - 1U) is the i-cluster interaction
- * mask for a super-cluster with all NCL_PER_SUPERCL bits set.
- */
- if (imask & (((1U << NCL_PER_SUPERCL) - 1U) << (jm * NCL_PER_SUPERCL)))
+ if (imask & (superClInteractionMask << (jm * c_numClPerSupercl)))
{
- mask_ji = (1U << (jm * NCL_PER_SUPERCL));
+ mask_ji = (1U << (jm * c_numClPerSupercl));
- cj = cjs[jm + (tidxj & 4) * NBNXN_GPU_JGROUP_SIZE / 4];
- aj = cj * CL_SIZE + tidxj;
+ cj = cjs[jm + (tidxj & 4) * c_nbnxnGpuJgroupSize/c_splitClSize];
+ aj = cj * c_clSize + tidxj;
/* load j atom data */
xqbuf = xq[aj];
xj = make_float3(xqbuf.x, xqbuf.y, xqbuf.z);
- qj_f = nbparam.epsfac * xqbuf.w;
+ qj_f = xqbuf.w;
+#ifndef LJ_COMB
typej = atom_types[aj];
+#else
+ ljcp_j = lj_comb[aj];
+#endif
fcj_buf = make_float3(0.0f);
- /* The PME and RF kernels don't unroll with CUDA <v4.1. */
-#if !defined PRUNE_NBL && !(GMX_CUDA_VERSION < 4010 && defined EXCLUSION_FORCES)
+#if !defined PRUNE_NBL
#pragma unroll 8
#endif
- for (i = 0; i < NCL_PER_SUPERCL; i++)
+ for (i = 0; i < c_numClPerSupercl; i++)
{
if (imask & mask_ji)
{
- ci_offset = i; /* i force buffer offset */
-
- ci = sci * NCL_PER_SUPERCL + i; /* i cluster index */
- ai = ci * CL_SIZE + tidxi; /* i atom index */
+ ci = sci * c_numClPerSupercl + i; /* i cluster index */
/* all threads load an atom from i cluster ci into shmem! */
- xqbuf = xqib[i * CL_SIZE + tidxi];
+ xqbuf = xqib[i * c_clSize + tidxi];
xi = make_float3(xqbuf.x, xqbuf.y, xqbuf.z);
/* distance between i and j atoms */
{
/* load the rest of the i-atom parameters */
qi = xqbuf.w;
-#ifdef IATYPE_SHMEM
- typei = atib[i * CL_SIZE + tidxi];
-#else
- typei = atom_types[ai];
-#endif
+#ifndef LJ_COMB
/* LJ 6*C6 and 12*C12 */
-#ifdef USE_TEXOBJ
+ typei = atib[i * c_clSize + tidxi];
c6 = tex1Dfetch<float>(nbparam.nbfp_texobj, 2 * (ntypes * typei + typej));
c12 = tex1Dfetch<float>(nbparam.nbfp_texobj, 2 * (ntypes * typei + typej) + 1);
#else
- c6 = tex1Dfetch(nbfp_texref, 2 * (ntypes * typei + typej));
- c12 = tex1Dfetch(nbfp_texref, 2 * (ntypes * typei + typej) + 1);
-#endif /* USE_TEXOBJ */
-
+ ljcp_i = ljcpib[i * c_clSize + tidxi];
+#ifdef LJ_COMB_GEOM
+ c6 = ljcp_i.x * ljcp_j.x;
+ c12 = ljcp_i.y * ljcp_j.y;
+#else
+ /* LJ 2^(1/6)*sigma and 12*epsilon */
+ sigma = ljcp_i.x + ljcp_j.x;
+ epsilon = ljcp_i.y * ljcp_j.y;
+#if defined CALC_ENERGIES || defined LJ_FORCE_SWITCH || defined LJ_POT_SWITCH
+ convert_sigma_epsilon_to_c6_c12(sigma, epsilon, &c6, &c12);
+#endif
+#endif /* LJ_COMB_GEOM */
+#endif /* LJ_COMB */
- /* avoid NaN for excluded pairs at r=0 */
- r2 += (1.0f - int_bit) * NBNXN_AVOID_SING_R2_INC;
+ // Ensure distance do not become so small that r^-12 overflows
+ r2 = max(r2, NBNXN_MIN_RSQ);
inv_r = rsqrt(r2);
inv_r2 = inv_r * inv_r;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
inv_r6 = inv_r2 * inv_r2 * inv_r2;
-#if defined EXCLUSION_FORCES
+#ifdef EXCLUSION_FORCES
/* We could mask inv_r2, but with Ewald
* masking both inv_r6 and F_invr is faster */
inv_r6 *= int_bit;
F_invr = inv_r6 * (c12 * inv_r6 - c6) * inv_r2;
#if defined CALC_ENERGIES || defined LJ_POT_SWITCH
- E_lj_p = int_bit * (c12 * (inv_r6 * inv_r6 + nbparam.repulsion_shift.cpot)*ONE_TWELVETH_F -
- c6 * (inv_r6 + nbparam.dispersion_shift.cpot)*ONE_SIXTH_F);
+ E_lj_p = int_bit * (c12 * (inv_r6 * inv_r6 + nbparam.repulsion_shift.cpot)*c_oneTwelveth -
+ c6 * (inv_r6 + nbparam.dispersion_shift.cpot)*c_oneSixth);
#endif
+#else /* !LJ_COMB_LB || CALC_ENERGIES */
+ float sig_r = sigma*inv_r;
+ float sig_r2 = sig_r*sig_r;
+ float sig_r6 = sig_r2*sig_r2*sig_r2;
+#ifdef EXCLUSION_FORCES
+ sig_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = epsilon * sig_r6 * (sig_r6 - 1.0f) * inv_r2;
+#endif /* !LJ_COMB_LB || CALC_ENERGIES */
#ifdef LJ_FORCE_SWITCH
#ifdef CALC_ENERGIES
#endif /* LJ_EWALD_COMB_GEOM */
#endif /* LJ_EWALD */
+ #ifdef LJ_POT_SWITCH
+ #ifdef CALC_ENERGIES
+ calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #else
+ calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #endif /* CALC_ENERGIES */
+ #endif /* LJ_POT_SWITCH */
+
#ifdef VDW_CUTOFF_CHECK
/* Separate VDW cut-off check to enable twin-range cut-offs
* (rvdw < rcoulomb <= rlist)
#endif
#endif /* VDW_CUTOFF_CHECK */
- #ifdef LJ_POT_SWITCH
- #ifdef CALC_ENERGIES
- calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #else
- calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #endif /* CALC_ENERGIES */
- #endif /* LJ_POT_SWITCH */
-
#ifdef CALC_ENERGIES
E_lj += E_lj_p;
#endif
F_invr += qi * qj_f * (int_bit*inv_r2*inv_r + pmecorrF(beta2*r2)*beta3);
#elif defined EL_EWALD_TAB
F_invr += qi * qj_f * (int_bit*inv_r2 -
-#ifdef USE_TEXOBJ
- interpolate_coulomb_force_r(nbparam.coulomb_tab_texobj, r2 * inv_r, coulomb_tab_scale)
-#else
- interpolate_coulomb_force_r(r2 * inv_r, coulomb_tab_scale)
-#endif /* USE_TEXOBJ */
- ) * inv_r;
-#endif /* EL_EWALD_ANA/TAB */
+ interpolate_coulomb_force_r(nbparam.coulomb_tab_texobj, r2 * inv_r, coulomb_tab_scale)) * inv_r;
+#endif /* EL_EWALD_ANA/TAB */
#ifdef CALC_ENERGIES
#ifdef EL_CUTOFF
fcj_buf -= f_ij;
/* accumulate i forces in registers */
- fci_buf[ci_offset] += f_ij;
+ fci_buf[i] += f_ij;
}
}
}
/* reduce j forces */
-#ifdef REDUCE_SHUFFLE
reduce_force_j_warp_shfl(fcj_buf, f, tidxi, aj);
-#else
- /* store j forces in shmem */
- f_buf[ tidx] = fcj_buf.x;
- f_buf[ FBUF_STRIDE + tidx] = fcj_buf.y;
- f_buf[2 * FBUF_STRIDE + tidx] = fcj_buf.z;
-
- reduce_force_j_generic(f_buf, f, tidxi, tidxj, aj);
-#endif
}
}
#ifdef PRUNE_NBL
}
}
+ /* skip central shifts when summing shift forces */
+ if (nb_sci.shift == CENTRAL)
+ {
+ bCalcFshift = false;
+ }
+
float fshift_buf = 0.0f;
/* reduce i forces */
- for (ci_offset = 0; ci_offset < NCL_PER_SUPERCL; ci_offset++)
+ for (i = 0; i < c_numClPerSupercl; i++)
{
- ai = (sci * NCL_PER_SUPERCL + ci_offset) * CL_SIZE + tidxi;
-#ifdef REDUCE_SHUFFLE
- reduce_force_i_warp_shfl(fci_buf[ci_offset], f,
+ ai = (sci * c_numClPerSupercl + i) * c_clSize + tidxi;
+ reduce_force_i_warp_shfl(fci_buf[i], f,
&fshift_buf, bCalcFshift,
tidxj, ai);
-#else
- f_buf[ tidx] = fci_buf[ci_offset].x;
- f_buf[ FBUF_STRIDE + tidx] = fci_buf[ci_offset].y;
- f_buf[2 * FBUF_STRIDE + tidx] = fci_buf[ci_offset].z;
- __syncthreads();
- reduce_force_i(f_buf, f,
- &fshift_buf, bCalcFshift,
- tidxi, tidxj, ai);
- __syncthreads();
-#endif
}
/* add up local shift forces into global mem, tidxj indexes x,y,z */
-#ifdef REDUCE_SHUFFLE
if (bCalcFshift && (tidxj & 3) < 3)
{
- atomicAdd(&(atdat.fshift[nb_sci.shift].x) + (tidxj & ~4), fshift_buf);
+ atomicAdd(&(atdat.fshift[nb_sci.shift].x) + (tidxj & 3), fshift_buf);
}
-#else
- if (bCalcFshift && tidxj < 3)
- {
- atomicAdd(&(atdat.fshift[nb_sci.shift].x) + tidxj, fshift_buf);
- }
-#endif
#ifdef CALC_ENERGIES
-#ifdef REDUCE_SHUFFLE
/* reduce the energies over warps and store into global memory */
reduce_energy_warp_shfl(E_lj, E_el, e_lj, e_el, tidx);
-#else
- /* flush the energies to shmem and reduce them */
- f_buf[ tidx] = E_lj;
- f_buf[FBUF_STRIDE + tidx] = E_el;
- reduce_energy_pow2(f_buf + (tidx & WARP_SIZE), e_lj, e_el, tidx & ~WARP_SIZE);
-#endif
#endif
}
-
-#undef REDUCE_SHUFFLE
-#undef IATYPE_SHMEM
-#undef USE_TEXOBJ
+#endif /* FUNCTION_DECLARATION_ONLY */
#undef NTHREAD_Z
#undef MIN_BLOCKS_PER_MP
#undef EL_EWALD_ANY
#undef EXCLUSION_FORCES
#undef LJ_EWALD
+
+#undef LJ_COMB
--- /dev/null
- #ifdef LJ_POT_SWITCH
- #ifdef CALC_ENERGIES
- calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #else
- calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #endif /* CALC_ENERGIES */
- #endif /* LJ_POT_SWITCH */
-
+/*
+ * This file is part of the GROMACS molecular simulation package.
+ *
+ * Copyright (c) 2012,2013,2014,2015,2016, by the GROMACS development team, led by
+ * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
+ * and including many others, as listed in the AUTHORS file in the
+ * top-level source directory and at http://www.gromacs.org.
+ *
+ * GROMACS is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public License
+ * as published by the Free Software Foundation; either version 2.1
+ * of the License, or (at your option) any later version.
+ *
+ * GROMACS is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with GROMACS; if not, see
+ * http://www.gnu.org/licenses, or write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * If you want to redistribute modifications to GROMACS, please
+ * consider that scientific software is very special. Version
+ * control is crucial - bugs must be traceable. We will be happy to
+ * consider code for inclusion in the official distribution, but
+ * derived work must not be called official GROMACS. Details are found
+ * in the README & COPYING files - if they are missing, get the
+ * official version at http://www.gromacs.org.
+ *
+ * To help us fund GROMACS development, we humbly ask that you cite
+ * the research papers on the package. Check out http://www.gromacs.org.
+ */
+
+/*! \internal \file
+ * \brief
+ * CUDA non-bonded kernel used through preprocessor-based code generation
+ * of multiple kernel flavors for CC 2.x, see nbnxn_cuda_kernels.cuh.
+ *
+ * NOTE: No include fence as it is meant to be included multiple times.
+ *
+ * \author Szilárd Páll <pall.szilard@gmail.com>
+ * \author Berk Hess <hess@kth.se>
+ * \ingroup module_mdlib
+ */
+
+#include "gromacs/gpu_utils/cuda_arch_utils.cuh"
+#include "gromacs/math/utilities.h"
+#include "gromacs/pbcutil/ishift.h"
+/* Note that floating-point constants in CUDA code should be suffixed
+ * with f (e.g. 0.5f), to stop the compiler producing intermediate
+ * code that is in double precision.
+ */
+
+#if GMX_PTX_ARCH >= 300
+#error "nbnxn_cuda_kernel_fermi.cuh included with GMX_PTX_ARCH >= 300"
+#endif
+
+#if defined EL_EWALD_ANA || defined EL_EWALD_TAB
+/* Note: convenience macro, needs to be undef-ed at the end of the file. */
+#define EL_EWALD_ANY
+#endif
+
+#if defined EL_EWALD_ANY || defined EL_RF || defined LJ_EWALD || (defined EL_CUTOFF && defined CALC_ENERGIES)
+/* Macro to control the calculation of exclusion forces in the kernel
+ * We do that with Ewald (elec/vdw) and RF. Cut-off only has exclusion
+ * energy terms.
+ *
+ * Note: convenience macro, needs to be undef-ed at the end of the file.
+ */
+#define EXCLUSION_FORCES
+#endif
+
+#if defined LJ_EWALD_COMB_GEOM || defined LJ_EWALD_COMB_LB
+/* Note: convenience macro, needs to be undef-ed at the end of the file. */
+#define LJ_EWALD
+#endif
+
+#if defined LJ_COMB_GEOM || defined LJ_COMB_LB
+#define LJ_COMB
+#endif
+
+/*
+ Kernel launch parameters:
+ - #blocks = #pair lists, blockId = pair list Id
+ - #threads = c_clSize^2
+ - shmem = see nbnxn_cuda.cu:calc_shmem_required()
+
+ Each thread calculates an i force-component taking one pair of i-j atoms.
+ */
+
+/**@{*/
+/*! \brief Definition of kernel launch configuration parameters for CC 2.x.
+ */
+
+/* Kernel launch bounds, 16 blocks/multiprocessor can be kept in flight. */
+#define THREADS_PER_BLOCK (c_clSize*c_clSize)
+
+__launch_bounds__(THREADS_PER_BLOCK)
+#ifdef PRUNE_NBL
+#ifdef CALC_ENERGIES
+__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _VF_prune_cuda)
+#else
+__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_prune_cuda)
+#endif /* CALC_ENERGIES */
+#else
+#ifdef CALC_ENERGIES
+__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _VF_cuda)
+#else
+__global__ void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_cuda)
+#endif /* CALC_ENERGIES */
+#endif /* PRUNE_NBL */
+(const cu_atomdata_t atdat,
+ const cu_nbparam_t nbparam,
+ const cu_plist_t plist,
+ bool bCalcFshift)
+#ifdef FUNCTION_DECLARATION_ONLY
+; /* Only do function declaration, omit the function body. */
+#else
+{
+ /* convenience variables */
+ const nbnxn_sci_t *pl_sci = plist.sci;
+#ifndef PRUNE_NBL
+ const
+#endif
+ nbnxn_cj4_t *pl_cj4 = plist.cj4;
+ const nbnxn_excl_t *excl = plist.excl;
+#ifndef LJ_COMB
+ const int *atom_types = atdat.atom_types;
+ int ntypes = atdat.ntypes;
+#else
+ const float2 *lj_comb = atdat.lj_comb;
+ float2 ljcp_i, ljcp_j;
+#endif
+ const float4 *xq = atdat.xq;
+ float3 *f = atdat.f;
+ const float3 *shift_vec = atdat.shift_vec;
+ float rcoulomb_sq = nbparam.rcoulomb_sq;
+#ifdef VDW_CUTOFF_CHECK
+ float rvdw_sq = nbparam.rvdw_sq;
+ float vdw_in_range;
+#endif
+#ifdef LJ_EWALD
+ float lje_coeff2, lje_coeff6_6;
+#endif
+#ifdef EL_RF
+ float two_k_rf = nbparam.two_k_rf;
+#endif
+#ifdef EL_EWALD_TAB
+ float coulomb_tab_scale = nbparam.coulomb_tab_scale;
+#endif
+#ifdef EL_EWALD_ANA
+ float beta2 = nbparam.ewald_beta*nbparam.ewald_beta;
+ float beta3 = nbparam.ewald_beta*nbparam.ewald_beta*nbparam.ewald_beta;
+#endif
+#ifdef PRUNE_NBL
+ float rlist_sq = nbparam.rlist_sq;
+#endif
+
+#ifdef CALC_ENERGIES
+#ifdef EL_EWALD_ANY
+ float beta = nbparam.ewald_beta;
+ float ewald_shift = nbparam.sh_ewald;
+#else
+ float c_rf = nbparam.c_rf;
+#endif /* EL_EWALD_ANY */
+ float *e_lj = atdat.e_lj;
+ float *e_el = atdat.e_el;
+#endif /* CALC_ENERGIES */
+
+ /* thread/block/warp id-s */
+ unsigned int tidxi = threadIdx.x;
+ unsigned int tidxj = threadIdx.y;
+ unsigned int tidx = threadIdx.y * blockDim.x + threadIdx.x;
+ unsigned int bidx = blockIdx.x;
+ unsigned int widx = tidx / warp_size; /* warp index */
+
+ int sci, ci, cj,
+ ai, aj,
+ cij4_start, cij4_end;
+#ifndef LJ_COMB
+ int typei, typej;
+#endif
+ int i, jm, j4, wexcl_idx;
+ float qi, qj_f,
+ r2, inv_r, inv_r2;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ float inv_r6, c6, c12;
+#endif
+#ifdef LJ_COMB_LB
+ float sigma, epsilon;
+#endif
+ float int_bit,
+ F_invr;
+#ifdef CALC_ENERGIES
+ float E_lj, E_el;
+#endif
+#if defined CALC_ENERGIES || defined LJ_POT_SWITCH
+ float E_lj_p;
+#endif
+ unsigned int wexcl, imask, mask_ji;
+ float4 xqbuf;
+ float3 xi, xj, rv, f_ij, fcj_buf;
+ float3 fci_buf[c_numClPerSupercl]; /* i force buffer */
+ nbnxn_sci_t nb_sci;
+
+ /*! i-cluster interaction mask for a super-cluster with all c_numClPerSupercl=8 bits set */
+ const unsigned superClInteractionMask = ((1U << c_numClPerSupercl) - 1U);
+
+ /* shmem buffer for i x+q pre-loading */
+ extern __shared__ float4 xqib[];
+
+ /* shmem buffer for cj, for each warp separately */
+ int *cjs = ((int *)(xqib + c_numClPerSupercl * c_clSize));
+ /* shmem j force buffer */
+ float *f_buf = (float *)(cjs + c_nbnxnGpuClusterpairSplit * c_nbnxnGpuJgroupSize);
+
+ nb_sci = pl_sci[bidx]; /* my i super-cluster's index = current bidx */
+ sci = nb_sci.sci; /* super-cluster */
+ cij4_start = nb_sci.cj4_ind_start; /* first ...*/
+ cij4_end = nb_sci.cj4_ind_end; /* and last index of j clusters */
+
+ {
+ /* Pre-load i-atom x and q into shared memory */
+ ci = sci * c_numClPerSupercl + tidxj;
+ ai = ci * c_clSize + tidxi;
+
+ xqbuf = xq[ai] + shift_vec[nb_sci.shift];
+ xqbuf.w *= nbparam.epsfac;
+ xqib[tidxj * c_clSize + tidxi] = xqbuf;
+ }
+ __syncthreads();
+
+ for (i = 0; i < c_numClPerSupercl; i++)
+ {
+ fci_buf[i] = make_float3(0.0f);
+ }
+
+#ifdef LJ_EWALD
+ /* TODO: we are trading registers with flops by keeping lje_coeff-s, try re-calculating it later */
+ lje_coeff2 = nbparam.ewaldcoeff_lj*nbparam.ewaldcoeff_lj;
+ lje_coeff6_6 = lje_coeff2*lje_coeff2*lje_coeff2*c_oneSixth;
+#endif
+
+
+#ifdef CALC_ENERGIES
+ E_lj = 0.0f;
+ E_el = 0.0f;
+
+#ifdef EXCLUSION_FORCES /* Ewald or RF */
+ if (nb_sci.shift == CENTRAL && pl_cj4[cij4_start].cj[0] == sci*c_numClPerSupercl)
+ {
+ /* we have the diagonal: add the charge and LJ self interaction energy term */
+ for (i = 0; i < c_numClPerSupercl; i++)
+ {
+#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
+ qi = xqib[i * c_clSize + tidxi].w;
+ E_el += qi*qi;
+#endif
+
+#ifdef LJ_EWALD
+ E_lj += tex1Dfetch(nbfp_texref, atom_types[(sci*c_numClPerSupercl + i)*c_clSize + tidxi]*(ntypes + 1)*2);
+#endif
+ }
+
+ /* divide the self term(s) equally over the j-threads, then multiply with the coefficients. */
+#ifdef LJ_EWALD
+ E_lj /= c_clSize;
+ E_lj *= 0.5f*c_oneSixth*lje_coeff6_6;
+#endif
+
+#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
+ /* Correct for epsfac^2 due to adding qi^2 */
+ E_el /= nbparam.epsfac*c_clSize;
+#if defined EL_RF || defined EL_CUTOFF
+ E_el *= -0.5f*c_rf;
+#else
+ E_el *= -beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
+#endif
+#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
+ }
+#endif /* EXCLUSION_FORCES */
+
+#endif /* CALC_ENERGIES */
+
+ /* loop over the j clusters = seen by any of the atoms in the current super-cluster */
+ for (j4 = cij4_start; j4 < cij4_end; j4++)
+ {
+ wexcl_idx = pl_cj4[j4].imei[widx].excl_ind;
+ imask = pl_cj4[j4].imei[widx].imask;
+ wexcl = excl[wexcl_idx].pair[(tidx) & (warp_size - 1)];
+
+#ifndef PRUNE_NBL
+ if (imask)
+#endif
+ {
+ /* Pre-load cj into shared memory on both warps separately */
+ if ((tidxj == 0 || tidxj == 4) && tidxi < c_nbnxnGpuJgroupSize)
+ {
+ cjs[tidxi + tidxj * c_nbnxnGpuJgroupSize/c_splitClSize] = pl_cj4[j4].cj[tidxi];
+ }
+
+ /* Unrolling this loop with pruning leads to register spilling;
+ Tested with up to nvcc 7.5 */
+#if !defined PRUNE_NBL
+#pragma unroll 4
+#endif
+ for (jm = 0; jm < c_nbnxnGpuJgroupSize; jm++)
+ {
+ if (imask & (superClInteractionMask << (jm * c_numClPerSupercl)))
+ {
+ mask_ji = (1U << (jm * c_numClPerSupercl));
+
+ cj = cjs[jm + (tidxj & 4) * c_nbnxnGpuJgroupSize/c_splitClSize];
+ aj = cj * c_clSize + tidxj;
+
+ /* load j atom data */
+ xqbuf = xq[aj];
+ xj = make_float3(xqbuf.x, xqbuf.y, xqbuf.z);
+ qj_f = xqbuf.w;
+#ifndef LJ_COMB
+ typej = atom_types[aj];
+#else
+ ljcp_j = lj_comb[aj];
+#endif
+
+ fcj_buf = make_float3(0.0f);
+
+#if !defined PRUNE_NBL
+#pragma unroll 8
+#endif
+ for (i = 0; i < c_numClPerSupercl; i++)
+ {
+ if (imask & mask_ji)
+ {
+ ci = sci * c_numClPerSupercl + i; /* i cluster index */
+ ai = ci * c_clSize + tidxi; /* i atom index */
+
+ /* all threads load an atom from i cluster ci into shmem! */
+ xqbuf = xqib[i * c_clSize + tidxi];
+ xi = make_float3(xqbuf.x, xqbuf.y, xqbuf.z);
+
+ /* distance between i and j atoms */
+ rv = xi - xj;
+ r2 = norm2(rv);
+
+#ifdef PRUNE_NBL
+ /* If _none_ of the atoms pairs are in cutoff range,
+ the bit corresponding to the current
+ cluster-pair in imask gets set to 0. */
+ if (!__any(r2 < rlist_sq))
+ {
+ imask &= ~mask_ji;
+ }
+#endif
+
+ int_bit = (wexcl & mask_ji) ? 1.0f : 0.0f;
+
+ /* cutoff & exclusion check */
+#ifdef EXCLUSION_FORCES
+ if (r2 < rcoulomb_sq *
+ (nb_sci.shift != CENTRAL || ci != cj || tidxj > tidxi))
+#else
+ if (r2 < rcoulomb_sq * int_bit)
+#endif
+ {
+ /* load the rest of the i-atom parameters */
+ qi = xqbuf.w;
+
+#ifndef LJ_COMB
+ /* LJ 6*C6 and 12*C12 */
+ typei = atom_types[ai];
+ c6 = tex1Dfetch(nbfp_texref, 2 * (ntypes * typei + typej));
+ c12 = tex1Dfetch(nbfp_texref, 2 * (ntypes * typei + typej) + 1);
+#else
+ ljcp_i = lj_comb[ai];
+#ifdef LJ_COMB_GEOM
+ c6 = ljcp_i.x * ljcp_j.x;
+ c12 = ljcp_i.y * ljcp_j.y;
+#else
+ /* LJ 2^(1/6)*sigma and 12*epsilon */
+ sigma = ljcp_i.x + ljcp_j.x;
+ epsilon = ljcp_i.y * ljcp_j.y;
+#if defined CALC_ENERGIES || defined LJ_FORCE_SWITCH || defined LJ_POT_SWITCH
+ convert_sigma_epsilon_to_c6_c12(sigma, epsilon, &c6, &c12);
+#endif
+#endif /* LJ_COMB_GEOM */
+#endif /* LJ_COMB */
+
+ // Ensure distance do not become so small that r^-12 overflows
+ r2 = max(r2, NBNXN_MIN_RSQ);
+
+ inv_r = rsqrt(r2);
+ inv_r2 = inv_r * inv_r;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ inv_r6 = inv_r2 * inv_r2 * inv_r2;
+#ifdef EXCLUSION_FORCES
+ /* We could mask inv_r2, but with Ewald
+ * masking both inv_r6 and F_invr is faster */
+ inv_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = inv_r6 * (c12 * inv_r6 - c6) * inv_r2;
+#if defined CALC_ENERGIES || defined LJ_POT_SWITCH
+ E_lj_p = int_bit * (c12 * (inv_r6 * inv_r6 + nbparam.repulsion_shift.cpot)*c_oneTwelveth -
+ c6 * (inv_r6 + nbparam.dispersion_shift.cpot)*c_oneSixth);
+#endif
+#else /* !LJ_COMB_LB || CALC_ENERGIES */
+ float sig_r = sigma*inv_r;
+ float sig_r2 = sig_r*sig_r;
+ float sig_r6 = sig_r2*sig_r2*sig_r2;
+#ifdef EXCLUSION_FORCES
+ sig_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = epsilon * sig_r6 * (sig_r6 - 1.0f) * inv_r2;
+#endif /* !LJ_COMB_LB || CALC_ENERGIES */
+
+#ifdef LJ_FORCE_SWITCH
+#ifdef CALC_ENERGIES
+ calculate_force_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+#else
+ calculate_force_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr);
+#endif /* CALC_ENERGIES */
+#endif /* LJ_FORCE_SWITCH */
+
+
+#ifdef LJ_EWALD
+#ifdef LJ_EWALD_COMB_GEOM
+#ifdef CALC_ENERGIES
+ calculate_lj_ewald_comb_geom_F_E(nbparam, typei, typej, r2, inv_r2, lje_coeff2, lje_coeff6_6, int_bit, &F_invr, &E_lj_p);
+#else
+ calculate_lj_ewald_comb_geom_F(nbparam, typei, typej, r2, inv_r2, lje_coeff2, lje_coeff6_6, &F_invr);
+#endif /* CALC_ENERGIES */
+#elif defined LJ_EWALD_COMB_LB
+ calculate_lj_ewald_comb_LB_F_E(nbparam, typei, typej, r2, inv_r2, lje_coeff2, lje_coeff6_6,
+#ifdef CALC_ENERGIES
+ int_bit, &F_invr, &E_lj_p
+#else
+ 0, &F_invr, NULL
+#endif /* CALC_ENERGIES */
+ );
+#endif /* LJ_EWALD_COMB_GEOM */
+#endif /* LJ_EWALD */
+
++#ifdef LJ_POT_SWITCH
++#ifdef CALC_ENERGIES
++ calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
++#else
++ calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
++#endif /* CALC_ENERGIES */
++#endif /* LJ_POT_SWITCH */
++
+#ifdef VDW_CUTOFF_CHECK
+ /* Separate VDW cut-off check to enable twin-range cut-offs
+ * (rvdw < rcoulomb <= rlist)
+ */
+ vdw_in_range = (r2 < rvdw_sq) ? 1.0f : 0.0f;
+ F_invr *= vdw_in_range;
+#ifdef CALC_ENERGIES
+ E_lj_p *= vdw_in_range;
+#endif
+#endif /* VDW_CUTOFF_CHECK */
+
+#ifdef CALC_ENERGIES
+ E_lj += E_lj_p;
+#endif
+
+
+#ifdef EL_CUTOFF
+#ifdef EXCLUSION_FORCES
+ F_invr += qi * qj_f * int_bit * inv_r2 * inv_r;
+#else
+ F_invr += qi * qj_f * inv_r2 * inv_r;
+#endif
+#endif
+#ifdef EL_RF
+ F_invr += qi * qj_f * (int_bit*inv_r2 * inv_r - two_k_rf);
+#endif
+#if defined EL_EWALD_ANA
+ F_invr += qi * qj_f * (int_bit*inv_r2*inv_r + pmecorrF(beta2*r2)*beta3);
+#elif defined EL_EWALD_TAB
+ F_invr += qi * qj_f * (int_bit*inv_r2 -
+ interpolate_coulomb_force_r(r2 * inv_r, coulomb_tab_scale)) * inv_r;
+#endif /* EL_EWALD_ANA/TAB */
+
+#ifdef CALC_ENERGIES
+#ifdef EL_CUTOFF
+ E_el += qi * qj_f * (int_bit*inv_r - c_rf);
+#endif
+#ifdef EL_RF
+ E_el += qi * qj_f * (int_bit*inv_r + 0.5f * two_k_rf * r2 - c_rf);
+#endif
+#ifdef EL_EWALD_ANY
+ /* 1.0f - erff is faster than erfcf */
+ E_el += qi * qj_f * (inv_r * (int_bit - erff(r2 * inv_r * beta)) - int_bit * ewald_shift);
+#endif /* EL_EWALD_ANY */
+#endif
+ f_ij = rv * F_invr;
+
+ /* accumulate j forces in registers */
+ fcj_buf -= f_ij;
+
+ /* accumulate i forces in registers */
+ fci_buf[i] += f_ij;
+ }
+ }
+
+ /* shift the mask bit by 1 */
+ mask_ji += mask_ji;
+ }
+
+ /* reduce j forces */
+ /* store j forces in shmem */
+ f_buf[ tidx] = fcj_buf.x;
+ f_buf[ c_fbufStride + tidx] = fcj_buf.y;
+ f_buf[2 * c_fbufStride + tidx] = fcj_buf.z;
+
+ reduce_force_j_generic(f_buf, f, tidxi, tidxj, aj);
+ }
+ }
+#ifdef PRUNE_NBL
+ /* Update the imask with the new one which does not contain the
+ out of range clusters anymore. */
+ pl_cj4[j4].imei[widx].imask = imask;
+#endif
+ }
+ }
+
+ /* skip central shifts when summing shift forces */
+ if (nb_sci.shift == CENTRAL)
+ {
+ bCalcFshift = false;
+ }
+
+ float fshift_buf = 0.0f;
+
+ /* reduce i forces */
+ for (i = 0; i < c_numClPerSupercl; i++)
+ {
+ ai = (sci * c_numClPerSupercl + i) * c_clSize + tidxi;
+ f_buf[ tidx] = fci_buf[i].x;
+ f_buf[ c_fbufStride + tidx] = fci_buf[i].y;
+ f_buf[2 * c_fbufStride + tidx] = fci_buf[i].z;
+ __syncthreads();
+ reduce_force_i(f_buf, f,
+ &fshift_buf, bCalcFshift,
+ tidxi, tidxj, ai);
+ __syncthreads();
+ }
+
+ /* add up local shift forces into global mem, tidxj indexes x,y,z */
+ if (bCalcFshift && tidxj < 3)
+ {
+ atomicAdd(&(atdat.fshift[nb_sci.shift].x) + tidxj, fshift_buf);
+ }
+
+#ifdef CALC_ENERGIES
+ /* flush the energies to shmem and reduce them */
+ f_buf[ tidx] = E_lj;
+ f_buf[c_fbufStride + tidx] = E_el;
+ reduce_energy_pow2(f_buf + (tidx & warp_size), e_lj, e_el, tidx & ~warp_size);
+#endif
+}
+#endif /* FUNCTION_DECLARATION_ONLY */
+
+#undef THREADS_PER_BLOCK
+
+#undef EL_EWALD_ANY
+#undef EXCLUSION_FORCES
+#undef LJ_EWALD
+
+#undef LJ_COMB
* the research papers on the package. Check out http://www.gromacs.org.
*/
+/*! \internal \file
+ * \brief OpenCL non-bonded kernel for AMD GPUs.
+ *
+ * OpenCL 1.2 support is expected.
+ *
+ * \author Anca Hamuraru <anca@streamcomputing.eu>
+ * \author Szilárd Páll <pall.szilard@gmail.com>
+ * \ingroup module_mdlib
+ */
+
+
#include "nbnxn_ocl_kernel_utils.clh"
/////////////////////////////////////////////////////////////////////////////////////////////////
#define LJ_EWALD
#endif
+#if defined LJ_COMB_GEOM || defined LJ_COMB_LB
+/* Note: convenience macro, needs to be undef-ed at the end of the file. */
+#define LJ_COMB
+#endif
+
/*
Kernel launch parameters:
- #blocks = #pair lists, blockId = pair list Id
- shmem = CL_SIZE^2 * sizeof(float)
Each thread calculates an i force-component taking one pair of i-j atoms.
+
+ TODO: implement 128 threads/wavefront by porting over the NTHREAD_Z/j4 loop
+ "horizontal splitting" over threads.
*/
-//#if __CUDA_ARCH__ >= 350
-//__launch_bounds__(64, 16)
-//#endif
+
/* NOTE:
NB_KERNEL_FUNC_NAME differs from the CUDA equivalent as it is not a variadic macro due to OpenCL not having a support for them, this version only takes exactly 2 arguments.
Thus if more strings need to be appended a new macro must be written or it must be directly appended here.
__kernel void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_opencl)
#endif
#endif
-(int ntypes, /* IN */
+(
+#ifndef LJ_COMB
+ int ntypes, /* IN */
+#endif
cl_nbparam_params_t nbparam_params, /* IN */
const __global float4 *restrict xq, /* IN */
__global float *restrict f, /* stores float3 values */ /* OUT */
__global float *restrict e_lj, /* OUT */
__global float *restrict e_el, /* OUT */
__global float *restrict fshift, /* stores float3 values */ /* OUT */
+#ifdef LJ_COMB
+ const __global float2 *restrict lj_comb, /* stores float2 values */ /* IN */
+#else
const __global int *restrict atom_types, /* IN */
+#endif
const __global float *restrict shift_vec, /* stores float3 values */ /* IN */
__constant float* nbfp_climg2d, /* IN */
__constant float* nbfp_comb_climg2d, /* IN */
cl_nbparam_params_t *nbparam = &nbparam_params;
float rcoulomb_sq = nbparam->rcoulomb_sq;
-
+#ifdef LJ_COMB
+ float2 ljcp_i, ljcp_j;
+#endif
#ifdef VDW_CUTOFF_CHECK
- float rvdw_sq = nbparam_params.rvdw_sq;//nbparam->rvdw_sq;
+ float rvdw_sq = nbparam_params.rvdw_sq;
float vdw_in_range;
#endif
#ifdef LJ_EWALD
unsigned int widx = tidx / WARP_SIZE; /* warp index */
int sci, ci, cj, ci_offset,
ai, aj,
- cij4_start, cij4_end,
- typei, typej,
- i, jm, j4, wexcl_idx;
+ cij4_start, cij4_end;
+#ifndef LJ_COMB
+ int typei, typej;
+#endif
+ int i, jm, j4, wexcl_idx;
float qi, qj_f,
- r2, inv_r, inv_r2, inv_r6,
- c6, c12,
- int_bit,
+ r2, inv_r, inv_r2;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ float inv_r6, c6, c12;
+#endif
+#if defined LJ_COMB_LB
+ float sigma, epsilon;
+#endif
+ float int_bit,
F_invr;
#ifdef CALC_ENERGIES
float3 fci_buf[NCL_PER_SUPERCL]; /* i force buffer */
nbnxn_sci_t nb_sci;
+ /*! i-cluster interaction mask for a super-cluster with all NCL_PER_SUPERCL=8 bits set */
+ const unsigned superClInteractionMask = ((1U << NCL_PER_SUPERCL) - 1U);
+
/* shmem buffer for cj, for both warps separately */
- __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
+ __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
#define LOCAL_OFFSET cjs + 2 * NBNXN_GPU_JGROUP_SIZE
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* shmem buffer for i atom-type pre-loading */
- __local int *atib = (__local int *)(LOCAL_OFFSET);
+ __local int *atib = (__local int *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET atib + NCL_PER_SUPERCL * CL_SIZE
+#else
+ __local float2 *ljcpib = (__local float2 *)(LOCAL_OFFSET);
+ #undef LOCAL_OFFSET
+ #define LOCAL_OFFSET ljcpib + NCL_PER_SUPERCL * CL_SIZE
+#endif
#endif
#ifndef REDUCE_SHUFFLE
/* shmem j force buffer */
- __local float *f_buf = (__local float *)(LOCAL_OFFSET);
+ __local float *f_buf = (__local float *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET f_buf + CL_SIZE * CL_SIZE * 3
#endif
ci = sci * NCL_PER_SUPERCL + tidxj;
ai = ci * CL_SIZE + tidxi;
- xqib[tidxj * CL_SIZE + tidxi] = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf.w *= nbparam->epsfac;
+ xqib[tidxj * CL_SIZE + tidxi] = xqbuf;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* Pre-load the i-atom types into shared memory */
- atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+ atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+#else
+ ljcpib[tidxj * CL_SIZE + tidxi] = lj_comb[ai];
+#endif
#endif
/* Initialise warp vote. (8x8 block) 2 warps for nvidia */
if(tidx==0 || tidx==32)
#endif /* LJ_EWALD */
#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
- E_el /= CL_SIZE;
+ /* Correct for epsfac^2 due to adding qi^2 */
+ E_el /= nbparam->epsfac*CL_SIZE;
#if defined EL_RF || defined EL_CUTOFF
- E_el *= -nbparam->epsfac*0.5f*c_rf;
+ E_el *= -0.5f*c_rf;
#else
- E_el *= -nbparam->epsfac*beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
+ E_el *= -beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
#endif
-#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
- }
-#endif /* EXCLUSION_FORCES */
-
-#endif /* CALC_ENERGIES */
-
- /* skip central shifts when summing shift forces */
- if (nb_sci.shift == CENTRAL)
- {
- bCalcFshift = false;
+#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
}
+#endif /* EXCLUSION_FORCES */
- fshift_buf = 0.0f;
+#endif /* CALC_ENERGIES */
/* loop over the j clusters = seen by any of the atoms in the current super-cluster */
for (j4 = cij4_start; j4 < cij4_end; j4++)
cjs[tidxi + tidxj * NBNXN_GPU_JGROUP_SIZE / 4] = pl_cj4[j4].cj[tidxi];
}
- /* Unrolling this loop
- - with pruning leads to register spilling;
- - on Kepler is much slower;
- - doesn't work on CUDA <v4.1
- Tested with nvcc 3.2 - 5.0.7 */
-#if !defined PRUNE_NBL //&& __CUDA_ARCH__ < 300 && CUDA_VERSION >= 4010
-//#pragma unroll 4
+ /* Unrolling this loop improves performance without pruning but
+ * with pruning it leads to slowdown.
+ *
+ * Tested with driver 1800.5
+ */
+#if !defined PRUNE_NBL
+#pragma unroll 4
#endif
for (jm = 0; jm < NBNXN_GPU_JGROUP_SIZE; jm++)
{
- if (imask & (supercl_interaction_mask << (jm * NCL_PER_SUPERCL)))
+ if (imask & (superClInteractionMask << (jm * NCL_PER_SUPERCL)))
{
mask_ji = (1U << (jm * NCL_PER_SUPERCL));
/* load j atom data */
xqbuf = xq[aj];
xj = (float3)(xqbuf.xyz);
- qj_f = nbparam->epsfac * xqbuf.w;
+ qj_f = xqbuf.w;
+#ifndef LJ_COMB
typej = atom_types[aj];
+#else
+ ljcp_j = lj_comb[aj];
+#endif
fcj_buf = (float3)(0.0f);
- /* The PME and RF kernels don't unroll with CUDA <v4.1. */
-#if !defined PRUNE_NBL //&& !(CUDA_VERSION < 4010 && defined EXCLUSION_FORCES)
-//#pragma unroll 8
+#if !defined PRUNE_NBL
+#pragma unroll 8
#endif
for (i = 0; i < NCL_PER_SUPERCL; i++)
{
imask &= ~mask_ji;
warp_any[widx]=0;
-
#endif
int_bit = (wexcl & mask_ji) ? 1.0f : 0.0f;
{
/* load the rest of the i-atom parameters */
qi = xqbuf.w;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
typei = atib[i * CL_SIZE + tidxi];
#else
+ ljcp_i = ljcpib[i * CL_SIZE + tidxi];
+#endif
+#else /* IATYPE_SHMEM */
+#ifndef LJ_COMB
typei = atom_types[ai];
+#else
+ ljcp_i = lj_comb[ai];
#endif
+#endif /* IATYPE_SHMEM */
/* LJ 6*C6 and 12*C12 */
+#ifndef LJ_COMB
c6 = nbfp_climg2d[2 * (ntypes * typei + typej)];
c12 = nbfp_climg2d[2 * (ntypes * typei + typej)+1];
+#else /* LJ_COMB */
+#ifdef LJ_COMB_GEOM
+ c6 = ljcp_i.x * ljcp_j.x;
+ c12 = ljcp_i.y * ljcp_j.y;
+#else
+ /* LJ 2^(1/6)*sigma and 12*epsilon */
+ sigma = ljcp_i.x + ljcp_j.x;
+ epsilon = ljcp_i.y * ljcp_j.y;
+#if defined CALC_ENERGIES || defined LJ_FORCE_SWITCH || defined LJ_POT_SWITCH
+ convert_sigma_epsilon_to_c6_c12(sigma, epsilon, &c6, &c12);
+#endif
+#endif /* LJ_COMB_GEOM */
+#endif /* LJ_COMB */
- /* avoid NaN for excluded pairs at r=0 */
- r2 += (1.0f - int_bit) * NBNXN_AVOID_SING_R2_INC;
+ // Ensure distance do not become so small that r^-12 overflows
+ r2 = max(r2,NBNXN_MIN_RSQ);
inv_r = rsqrt(r2);
inv_r2 = inv_r * inv_r;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
inv_r6 = inv_r2 * inv_r2 * inv_r2;
#if defined EXCLUSION_FORCES
/* We could mask inv_r2, but with Ewald
c6 * (inv_r6 + nbparam->dispersion_shift.cpot)*ONE_SIXTH_F);
#endif
+#else /* ! LJ_COMB_LB || CALC_ENERGIES */
+ float sig_r = sigma*inv_r;
+ float sig_r2 = sig_r*sig_r;
+ float sig_r6 = sig_r2*sig_r2*sig_r2;
+#if defined EXCLUSION_FORCES
+ sig_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = epsilon * sig_r6 * (sig_r6 - 1.0f) * inv_r2;
+#endif /* ! LJ_COMB_LB || CALC_ENERGIES */
#ifdef LJ_FORCE_SWITCH
#endif /* LJ_EWALD_COMB_GEOM */
#endif /* LJ_EWALD */
+ #ifdef LJ_POT_SWITCH
+ #ifdef CALC_ENERGIES
+ calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #else
+ calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #endif /* CALC_ENERGIES */
+ #endif /* LJ_POT_SWITCH */
+
#ifdef VDW_CUTOFF_CHECK
/* Separate VDW cut-off check to enable twin-range cut-offs
* (rvdw < rcoulomb <= rlist)
#endif
#endif /* VDW_CUTOFF_CHECK */
- #ifdef LJ_POT_SWITCH
- #ifdef CALC_ENERGIES
- calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #else
- calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #endif /* CALC_ENERGIES */
- #endif /* LJ_POT_SWITCH */
-
#ifdef CALC_ENERGIES
E_lj += E_lj_p;
}
}
+ /* skip central shifts when summing shift forces */
+ if (nb_sci.shift == CENTRAL)
+ {
+ bCalcFshift = false;
+ }
+
+ fshift_buf = 0.0f;
+
/* reduce i forces */
for (ci_offset = 0; ci_offset < NCL_PER_SUPERCL; ci_offset++)
{
barrier(CLK_LOCAL_MEM_FENCE);
}
- /* add up local shift forces into global mem */
- //if (bCalcFshift && tidxj == 0)
- // atomicAdd_g_f3(&(fshift[3 * nb_sci.shift]),fshift_buf);
+ /* add up local shift forces into global mem */
if (bCalcFshift)
- {
+ {
/* Only threads with tidxj < 3 will update fshift.
The threads performing the update must be the same with the threads
which stored the reduction result in reduce_force_i function
#undef EL_EWALD_ANY
#undef EXCLUSION_FORCES
#undef LJ_EWALD
+
+#undef LJ_COMB
* the research papers on the package. Check out http://www.gromacs.org.
*/
+/*! \internal \file
+ * \brief OpenCL non-bonded "generic" kernel for architectures other than AMD/NVIDIA.
+ *
+ * This kernel is use by default on all architectures other than the ones we
+ * explicitly support/optimize for.
+ *
+ * OpenCL 1.1 support is expected.
+ *
+ * \author Anca Hamuraru <anca@streamcomputing.eu>
+ * \author Szilárd Páll <pall.szilard@gmail.com>
+ * \ingroup module_mdlib
+ */
+
+
#include "nbnxn_ocl_kernel_utils.clh"
/////////////////////////////////////////////////////////////////////////////////////////////////
#define LJ_EWALD
#endif
+#if defined LJ_COMB_GEOM || defined LJ_COMB_LB
+/* Note: convenience macro, needs to be undef-ed at the end of the file. */
+#define LJ_COMB
+#endif
+
/*
Kernel launch parameters:
- #blocks = #pair lists, blockId = pair list Id
- shmem = CL_SIZE^2 * sizeof(float)
Each thread calculates an i force-component taking one pair of i-j atoms.
+
+ TODO: implement 128 threads/wavefront by porting over the NTHREAD_Z/j4 loop
+ "horizontal splitting" over threads.
*/
-//#if __CUDA_ARCH__ >= 350
-//__launch_bounds__(64, 16)
-//#endif
+
/* NOTE:
NB_KERNEL_FUNC_NAME differs from the CUDA equivalent as it is not a variadic macro due to OpenCL not having a support for them, this version only takes exactly 2 arguments.
Thus if more strings need to be appended a new macro must be written or it must be directly appended here.
__kernel void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_opencl)
#endif
#endif
-(int ntypes, /* IN */
+(
+#ifndef LJ_COMB
+ int ntypes, /* IN */
+#endif
cl_nbparam_params_t nbparam_params, /* IN */
const __global float4 *restrict xq, /* IN */
__global float *restrict f, /* stores float3 values */ /* OUT */
__global float *restrict e_lj, /* OUT */
__global float *restrict e_el, /* OUT */
__global float *restrict fshift, /* stores float3 values */ /* OUT */
+#ifdef LJ_COMB
+ const __global float2 *restrict lj_comb, /* stores float2 values */ /* IN */
+#else
const __global int *restrict atom_types, /* IN */
+#endif
const __global float *restrict shift_vec, /* stores float3 values */ /* IN */
__constant float* nbfp_climg2d, /* IN */
__constant float* nbfp_comb_climg2d, /* IN */
cl_nbparam_params_t *nbparam = &nbparam_params;
float rcoulomb_sq = nbparam->rcoulomb_sq;
-
+#ifdef LJ_COMB
+ float2 ljcp_i, ljcp_j;
+#endif
#ifdef VDW_CUTOFF_CHECK
- float rvdw_sq = nbparam_params.rvdw_sq;//nbparam->rvdw_sq;
+ float rvdw_sq = nbparam_params.rvdw_sq;
float vdw_in_range;
#endif
#ifdef LJ_EWALD
unsigned int widx = tidx / WARP_SIZE; /* warp index */
int sci, ci, cj, ci_offset,
ai, aj,
- cij4_start, cij4_end,
- typei, typej,
- i, jm, j4, wexcl_idx;
+ cij4_start, cij4_end;
+#ifndef LJ_COMB
+ int typei, typej;
+#endif
+ int i, jm, j4, wexcl_idx;
float qi, qj_f,
- r2, inv_r, inv_r2, inv_r6,
- c6, c12,
- int_bit,
+ r2, inv_r, inv_r2;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ float inv_r6, c6, c12;
+#endif
+#if defined LJ_COMB_LB
+ float sigma, epsilon;
+#endif
+ float int_bit,
F_invr;
#ifdef CALC_ENERGIES
float3 fci_buf[NCL_PER_SUPERCL]; /* i force buffer */
nbnxn_sci_t nb_sci;
+ /*! i-cluster interaction mask for a super-cluster with all NCL_PER_SUPERCL=8 bits set */
+ const unsigned superClInteractionMask = ((1U << NCL_PER_SUPERCL) - 1U);
+
/* shmem buffer for cj, for both warps separately */
- __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
+ __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
#define LOCAL_OFFSET cjs + 2 * NBNXN_GPU_JGROUP_SIZE
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* shmem buffer for i atom-type pre-loading */
- __local int *atib = (__local int *)(LOCAL_OFFSET);
+ __local int *atib = (__local int *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET atib + NCL_PER_SUPERCL * CL_SIZE
+#else
+ __local float2 *ljcpib = (__local float2 *)(LOCAL_OFFSET);
+ #undef LOCAL_OFFSET
+ #define LOCAL_OFFSET ljcpib + NCL_PER_SUPERCL * CL_SIZE
+#endif
#endif
#ifndef REDUCE_SHUFFLE
/* shmem j force buffer */
- __local float *f_buf = (__local float *)(LOCAL_OFFSET);
+ __local float *f_buf = (__local float *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET f_buf + CL_SIZE * CL_SIZE * 3
#endif
ci = sci * NCL_PER_SUPERCL + tidxj;
ai = ci * CL_SIZE + tidxi;
- xqib[tidxj * CL_SIZE + tidxi] = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf.w *= nbparam->epsfac;
+ xqib[tidxj * CL_SIZE + tidxi] = xqbuf;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* Pre-load the i-atom types into shared memory */
- atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+ atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+#else
+ ljcpib[tidxj * CL_SIZE + tidxi] = lj_comb[ai];
+#endif
#endif
/* Initialise warp vote. (8x8 block) 2 warps for nvidia */
if(tidx==0 || tidx==32)
#endif /* LJ_EWALD */
#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
- E_el /= CL_SIZE;
+ /* Correct for epsfac^2 due to adding qi^2 */
+ E_el /= nbparam->epsfac*CL_SIZE;
#if defined EL_RF || defined EL_CUTOFF
- E_el *= -nbparam->epsfac*0.5f*c_rf;
+ E_el *= -0.5f*c_rf;
#else
- E_el *= -nbparam->epsfac*beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
+ E_el *= -beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
#endif
-#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
+#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
}
-#endif /* EXCLUSION_FORCES */
+#endif /* EXCLUSION_FORCES */
-#endif /* CALC_ENERGIES */
-
- /* skip central shifts when summing shift forces */
- if (nb_sci.shift == CENTRAL)
- {
- bCalcFshift = false;
- }
-
- fshift_buf = 0.0f;
+#endif /* CALC_ENERGIES */
/* loop over the j clusters = seen by any of the atoms in the current super-cluster */
for (j4 = cij4_start; j4 < cij4_end; j4++)
cjs[tidxi + tidxj * NBNXN_GPU_JGROUP_SIZE / 4] = pl_cj4[j4].cj[tidxi];
}
- /* Unrolling this loop
- - with pruning leads to register spilling;
- - on Kepler is much slower;
- - doesn't work on CUDA <v4.1
- Tested with nvcc 3.2 - 5.0.7 */
-#if !defined PRUNE_NBL //&& __CUDA_ARCH__ < 300 && CUDA_VERSION >= 4010
-//#pragma unroll 4
+ /* Unrolling this loop improves performance without pruning but
+ * with pruning it leads to slowdown.
+ *
+ * Tested with driver 1800.5
+ */
+#if !defined PRUNE_NBL
+#pragma unroll 4
#endif
for (jm = 0; jm < NBNXN_GPU_JGROUP_SIZE; jm++)
{
- if (imask & (supercl_interaction_mask << (jm * NCL_PER_SUPERCL)))
+ if (imask & (superClInteractionMask << (jm * NCL_PER_SUPERCL)))
{
mask_ji = (1U << (jm * NCL_PER_SUPERCL));
/* load j atom data */
xqbuf = xq[aj];
xj = (float3)(xqbuf.xyz);
- qj_f = nbparam->epsfac * xqbuf.w;
+ qj_f = xqbuf.w;
+#ifndef LJ_COMB
typej = atom_types[aj];
+#else
+ ljcp_j = lj_comb[aj];
+#endif
fcj_buf = (float3)(0.0f);
- /* The PME and RF kernels don't unroll with CUDA <v4.1. */
-#if !defined PRUNE_NBL //&& !(CUDA_VERSION < 4010 && defined EXCLUSION_FORCES)
-//#pragma unroll 8
+#if !defined PRUNE_NBL
+#pragma unroll 8
#endif
for (i = 0; i < NCL_PER_SUPERCL; i++)
{
imask &= ~mask_ji;
warp_any[widx]=0;
-
#endif
int_bit = (wexcl & mask_ji) ? 1.0f : 0.0f;
{
/* load the rest of the i-atom parameters */
qi = xqbuf.w;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
typei = atib[i * CL_SIZE + tidxi];
#else
+ ljcp_i = ljcpib[i * CL_SIZE + tidxi];
+#endif
+#else /* IATYPE_SHMEM */
+#ifndef LJ_COMB
typei = atom_types[ai];
+#else
+ ljcp_i = lj_comb[ai];
#endif
+#endif /* IATYPE_SHMEM */
/* LJ 6*C6 and 12*C12 */
+#ifndef LJ_COMB
c6 = nbfp_climg2d[2 * (ntypes * typei + typej)];
c12 = nbfp_climg2d[2 * (ntypes * typei + typej)+1];
+#else /* LJ_COMB */
+#ifdef LJ_COMB_GEOM
+ c6 = ljcp_i.x * ljcp_j.x;
+ c12 = ljcp_i.y * ljcp_j.y;
+#else
+ /* LJ 2^(1/6)*sigma and 12*epsilon */
+ sigma = ljcp_i.x + ljcp_j.x;
+ epsilon = ljcp_i.y * ljcp_j.y;
+#if defined CALC_ENERGIES || defined LJ_FORCE_SWITCH || defined LJ_POT_SWITCH
+ convert_sigma_epsilon_to_c6_c12(sigma, epsilon, &c6, &c12);
+#endif
+#endif /* LJ_COMB_GEOM */
+#endif /* LJ_COMB */
- /* avoid NaN for excluded pairs at r=0 */
- r2 += (1.0f - int_bit) * NBNXN_AVOID_SING_R2_INC;
+ // Ensure distance do not become so small that r^-12 overflows
+ r2 = max(r2,NBNXN_MIN_RSQ);
inv_r = rsqrt(r2);
inv_r2 = inv_r * inv_r;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
inv_r6 = inv_r2 * inv_r2 * inv_r2;
#if defined EXCLUSION_FORCES
/* We could mask inv_r2, but with Ewald
c6 * (inv_r6 + nbparam->dispersion_shift.cpot)*ONE_SIXTH_F);
#endif
+#else /* ! LJ_COMB_LB || CALC_ENERGIES */
+ float sig_r = sigma*inv_r;
+ float sig_r2 = sig_r*sig_r;
+ float sig_r6 = sig_r2*sig_r2*sig_r2;
+#if defined EXCLUSION_FORCES
+ sig_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = epsilon * sig_r6 * (sig_r6 - 1.0f) * inv_r2;
+#endif /* ! LJ_COMB_LB || CALC_ENERGIES */
#ifdef LJ_FORCE_SWITCH
#endif /* LJ_EWALD_COMB_GEOM */
#endif /* LJ_EWALD */
+ #ifdef LJ_POT_SWITCH
+ #ifdef CALC_ENERGIES
+ calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #else
+ calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #endif /* CALC_ENERGIES */
+ #endif /* LJ_POT_SWITCH */
+
#ifdef VDW_CUTOFF_CHECK
/* Separate VDW cut-off check to enable twin-range cut-offs
* (rvdw < rcoulomb <= rlist)
#endif
#endif /* VDW_CUTOFF_CHECK */
- #ifdef LJ_POT_SWITCH
- #ifdef CALC_ENERGIES
- calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #else
- calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #endif /* CALC_ENERGIES */
- #endif /* LJ_POT_SWITCH */
-
#ifdef CALC_ENERGIES
E_lj += E_lj_p;
}
}
+ /* skip central shifts when summing shift forces */
+ if (nb_sci.shift == CENTRAL)
+ {
+ bCalcFshift = false;
+ }
+
+ fshift_buf = 0.0f;
+
/* reduce i forces */
for (ci_offset = 0; ci_offset < NCL_PER_SUPERCL; ci_offset++)
{
barrier(CLK_LOCAL_MEM_FENCE);
}
- /* add up local shift forces into global mem */
- //if (bCalcFshift && tidxj == 0)
- // atomicAdd_g_f3(&(fshift[3 * nb_sci.shift]),fshift_buf);
+ /* add up local shift forces into global mem */
if (bCalcFshift)
- {
+ {
/* Only threads with tidxj < 3 will update fshift.
The threads performing the update must be the same with the threads
which stored the reduction result in reduce_force_i function
#undef EL_EWALD_ANY
#undef EXCLUSION_FORCES
#undef LJ_EWALD
+
+#undef LJ_COMB
* the research papers on the package. Check out http://www.gromacs.org.
*/
+/*! \internal \file
+ * \brief OpenCL non-bonded kernel for NVIDIA GPUs.
+ *
+ * OpenCL 1.2 support is expected (CUDA driver API 7.5 and later).
+ *
+ * \author Anca Hamuraru <anca@streamcomputing.eu>
+ * \author Szilárd Páll <pall.szilard@gmail.com>
+ * \ingroup module_mdlib
+ */
+
+
#include "nbnxn_ocl_kernel_utils.clh"
/////////////////////////////////////////////////////////////////////////////////////////////////
#define LJ_EWALD
#endif
+#if defined LJ_COMB_GEOM || defined LJ_COMB_LB
+/* Note: convenience macro, needs to be undef-ed at the end of the file. */
+#define LJ_COMB
+#endif
+
/*
Kernel launch parameters:
- #blocks = #pair lists, blockId = pair list Id
Each thread calculates an i force-component taking one pair of i-j atoms.
*/
-//#if __CUDA_ARCH__ >= 350
-//__launch_bounds__(64, 16)
-//#endif
+
/* NOTE:
NB_KERNEL_FUNC_NAME differs from the CUDA equivalent as it is not a variadic macro due to OpenCL not having a support for them, this version only takes exactly 2 arguments.
Thus if more strings need to be appended a new macro must be written or it must be directly appended here.
__kernel void NB_KERNEL_FUNC_NAME(nbnxn_kernel, _F_opencl)
#endif
#endif
-(int ntypes, /* IN */
+(
+#ifndef LJ_COMB
+ int ntypes, /* IN */
+#endif
cl_nbparam_params_t nbparam_params, /* IN */
const __global float4 *restrict xq, /* IN */
__global float *restrict f, /* stores float3 values */ /* OUT */
__global float *restrict e_lj, /* OUT */
__global float *restrict e_el, /* OUT */
__global float *restrict fshift, /* stores float3 values */ /* OUT */
+#ifdef LJ_COMB
+ const __global float2 *restrict lj_comb, /* stores float2 values */ /* IN */
+#else
const __global int *restrict atom_types, /* IN */
+#endif
const __global float *restrict shift_vec, /* stores float3 values */ /* IN */
__constant float* nbfp_climg2d, /* IN */
__constant float* nbfp_comb_climg2d, /* IN */
cl_nbparam_params_t *nbparam = &nbparam_params;
float rcoulomb_sq = nbparam->rcoulomb_sq;
-
+#ifdef LJ_COMB
+ float2 ljcp_i, ljcp_j;
+#endif
#ifdef VDW_CUTOFF_CHECK
- float rvdw_sq = nbparam_params.rvdw_sq;//nbparam->rvdw_sq;
+ float rvdw_sq = nbparam_params.rvdw_sq;
float vdw_in_range;
#endif
#ifdef LJ_EWALD
unsigned int widx = tidx / WARP_SIZE; /* warp index */
int sci, ci, cj, ci_offset,
ai, aj,
- cij4_start, cij4_end,
- typei, typej,
- i, jm, j4, wexcl_idx;
+ cij4_start, cij4_end;
+#ifndef LJ_COMB
+ int typei, typej;
+#endif
+ int i, jm, j4, wexcl_idx;
float qi, qj_f,
- r2, inv_r, inv_r2, inv_r6,
- c6, c12,
- int_bit,
+ r2, inv_r, inv_r2;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
+ float inv_r6, c6, c12;
+#endif
+#if defined LJ_COMB_LB
+ float sigma, epsilon;
+#endif
+ float int_bit,
F_invr;
#ifdef CALC_ENERGIES
float3 fci_buf[NCL_PER_SUPERCL]; /* i force buffer */
nbnxn_sci_t nb_sci;
+ /*! i-cluster interaction mask for a super-cluster with all NCL_PER_SUPERCL=8 bits set */
+ const unsigned superClInteractionMask = ((1U << NCL_PER_SUPERCL) - 1U);
+
/* shmem buffer for cj, for both warps separately */
- __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
+ __local int *cjs = (__local int *)(xqib + NCL_PER_SUPERCL * CL_SIZE);
#define LOCAL_OFFSET cjs + 2 * NBNXN_GPU_JGROUP_SIZE
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* shmem buffer for i atom-type pre-loading */
- __local int *atib = (__local int *)(LOCAL_OFFSET);
+ __local int *atib = (__local int *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET atib + NCL_PER_SUPERCL * CL_SIZE
+#else
+ __local float2 *ljcpib = (__local float2 *)(LOCAL_OFFSET);
+ #undef LOCAL_OFFSET
+ #define LOCAL_OFFSET ljcpib + NCL_PER_SUPERCL * CL_SIZE
+#endif
#endif
#ifndef REDUCE_SHUFFLE
/* shmem j force buffer */
- __local float *f_buf = (__local float *)(LOCAL_OFFSET);
+ __local float *f_buf = (__local float *)(LOCAL_OFFSET);
#undef LOCAL_OFFSET
#define LOCAL_OFFSET f_buf + CL_SIZE * CL_SIZE * 3
#endif
ci = sci * NCL_PER_SUPERCL + tidxj;
ai = ci * CL_SIZE + tidxi;
- xqib[tidxj * CL_SIZE + tidxi] = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf = xq[ai] + (float4)(shift_vec[3 * nb_sci.shift], shift_vec[3 * nb_sci.shift + 1], shift_vec[3 * nb_sci.shift + 2], 0.0f);
+ xqbuf.w *= nbparam->epsfac;
+ xqib[tidxj * CL_SIZE + tidxi] = xqbuf;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
/* Pre-load the i-atom types into shared memory */
- atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+ atib[tidxj * CL_SIZE + tidxi] = atom_types[ai];
+#else
+ ljcpib[tidxj * CL_SIZE + tidxi] = lj_comb[ai];
+#endif
#endif
/* Initialise warp vote. (8x8 block) 2 warps for nvidia */
if(tidx==0 || tidx==32)
#endif /* LJ_EWALD */
#if defined EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF
- E_el /= CL_SIZE;
+ /* Correct for epsfac^2 due to adding qi^2 */
+ E_el /= nbparam->epsfac*CL_SIZE;
#if defined EL_RF || defined EL_CUTOFF
- E_el *= -nbparam->epsfac*0.5f*c_rf;
+ E_el *= -0.5f*c_rf;
#else
- E_el *= -nbparam->epsfac*beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
+ E_el *= -beta*M_FLOAT_1_SQRTPI; /* last factor 1/sqrt(pi) */
#endif
-#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
+#endif /* EL_EWALD_ANY || defined EL_RF || defined EL_CUTOFF */
}
-#endif /* EXCLUSION_FORCES */
+#endif /* EXCLUSION_FORCES */
-#endif /* CALC_ENERGIES */
-
- /* skip central shifts when summing shift forces */
- if (nb_sci.shift == CENTRAL)
- {
- bCalcFshift = false;
- }
-
- fshift_buf = 0.0f;
+#endif /* CALC_ENERGIES */
/* loop over the j clusters = seen by any of the atoms in the current super-cluster */
for (j4 = cij4_start; j4 < cij4_end; j4++)
cjs[tidxi + tidxj * NBNXN_GPU_JGROUP_SIZE / 4] = pl_cj4[j4].cj[tidxi];
}
- /* Unrolling this loop
- - with pruning leads to register spilling;
- - on Kepler is much slower;
- - doesn't work on CUDA <v4.1
- Tested with nvcc 3.2 - 5.0.7 */
-#if !defined PRUNE_NBL //&& __CUDA_ARCH__ < 300 && CUDA_VERSION >= 4010
+ /* TODO: check loop unrolling with NVIDIA OpenCL */
+#if !defined PRUNE_NBL
//#pragma unroll 4
#endif
for (jm = 0; jm < NBNXN_GPU_JGROUP_SIZE; jm++)
{
- if (imask & (supercl_interaction_mask << (jm * NCL_PER_SUPERCL)))
+ if (imask & (superClInteractionMask << (jm * NCL_PER_SUPERCL)))
{
mask_ji = (1U << (jm * NCL_PER_SUPERCL));
/* load j atom data */
xqbuf = xq[aj];
xj = (float3)(xqbuf.xyz);
- qj_f = nbparam->epsfac * xqbuf.w;
+ qj_f = xqbuf.w;
+#ifndef LJ_COMB
typej = atom_types[aj];
+#else
+ ljcp_j = lj_comb[aj];
+#endif
fcj_buf = (float3)(0.0f);
- /* The PME and RF kernels don't unroll with CUDA <v4.1. */
-#if !defined PRUNE_NBL //&& !(CUDA_VERSION < 4010 && defined EXCLUSION_FORCES)
-//#pragma unroll 8
+#if !defined PRUNE_NBL
+#pragma unroll 8
#endif
for (i = 0; i < NCL_PER_SUPERCL; i++)
{
imask &= ~mask_ji;
warp_any[widx]=0;
-
#endif
int_bit = (wexcl & mask_ji) ? 1.0f : 0.0f;
{
/* load the rest of the i-atom parameters */
qi = xqbuf.w;
-#ifdef IATYPE_SHMEM //Should not be defined! CUDA > 300
+#ifdef IATYPE_SHMEM
+#ifndef LJ_COMB
typei = atib[i * CL_SIZE + tidxi];
#else
+ ljcp_i = ljcpib[i * CL_SIZE + tidxi];
+#endif
+#else /* IATYPE_SHMEM */
+#ifndef LJ_COMB
typei = atom_types[ai];
+#else
+ ljcp_i = lj_comb[ai];
#endif
+#endif /* IATYPE_SHMEM */
/* LJ 6*C6 and 12*C12 */
+#ifndef LJ_COMB
c6 = nbfp_climg2d[2 * (ntypes * typei + typej)];
c12 = nbfp_climg2d[2 * (ntypes * typei + typej)+1];
+#else /* LJ_COMB */
+#ifdef LJ_COMB_GEOM
+ c6 = ljcp_i.x * ljcp_j.x;
+ c12 = ljcp_i.y * ljcp_j.y;
+#else
+ /* LJ 2^(1/6)*sigma and 12*epsilon */
+ sigma = ljcp_i.x + ljcp_j.x;
+ epsilon = ljcp_i.y * ljcp_j.y;
+#if defined CALC_ENERGIES || defined LJ_FORCE_SWITCH || defined LJ_POT_SWITCH
+ convert_sigma_epsilon_to_c6_c12(sigma, epsilon, &c6, &c12);
+#endif
+#endif /* LJ_COMB_GEOM */
+#endif /* LJ_COMB */
- /* avoid NaN for excluded pairs at r=0 */
- r2 += (1.0f - int_bit) * NBNXN_AVOID_SING_R2_INC;
+ // Ensure distance do not become so small that r^-12 overflows
+ r2 = max(r2,NBNXN_MIN_RSQ);
inv_r = rsqrt(r2);
inv_r2 = inv_r * inv_r;
+#if !defined LJ_COMB_LB || defined CALC_ENERGIES
inv_r6 = inv_r2 * inv_r2 * inv_r2;
#if defined EXCLUSION_FORCES
/* We could mask inv_r2, but with Ewald
c6 * (inv_r6 + nbparam->dispersion_shift.cpot)*ONE_SIXTH_F);
#endif
+#else /* ! LJ_COMB_LB || CALC_ENERGIES */
+ float sig_r = sigma*inv_r;
+ float sig_r2 = sig_r*sig_r;
+ float sig_r6 = sig_r2*sig_r2*sig_r2;
+#if defined EXCLUSION_FORCES
+ sig_r6 *= int_bit;
+#endif /* EXCLUSION_FORCES */
+
+ F_invr = epsilon * sig_r6 * (sig_r6 - 1.0f) * inv_r2;
+#endif /* ! LJ_COMB_LB || CALC_ENERGIES */
#ifdef LJ_FORCE_SWITCH
#endif /* LJ_EWALD_COMB_GEOM */
#endif /* LJ_EWALD */
+ #ifdef LJ_POT_SWITCH
+ #ifdef CALC_ENERGIES
+ calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #else
+ calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
+ #endif /* CALC_ENERGIES */
+ #endif /* LJ_POT_SWITCH */
+
#ifdef VDW_CUTOFF_CHECK
/* Separate VDW cut-off check to enable twin-range cut-offs
* (rvdw < rcoulomb <= rlist)
#endif
#endif /* VDW_CUTOFF_CHECK */
- #ifdef LJ_POT_SWITCH
- #ifdef CALC_ENERGIES
- calculate_potential_switch_F_E(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #else
- calculate_potential_switch_F(nbparam, c6, c12, inv_r, r2, &F_invr, &E_lj_p);
- #endif /* CALC_ENERGIES */
- #endif /* LJ_POT_SWITCH */
-
#ifdef CALC_ENERGIES
E_lj += E_lj_p;
}
}
+ /* skip central shifts when summing shift forces */
+ if (nb_sci.shift == CENTRAL)
+ {
+ bCalcFshift = false;
+ }
+
+ fshift_buf = 0.0f;
+
/* reduce i forces */
for (ci_offset = 0; ci_offset < NCL_PER_SUPERCL; ci_offset++)
{
barrier(CLK_LOCAL_MEM_FENCE);
}
- /* add up local shift forces into global mem */
- //if (bCalcFshift && tidxj == 0)
- // atomicAdd_g_f3(&(fshift[3 * nb_sci.shift]),fshift_buf);
+ /* add up local shift forces into global mem */
if (bCalcFshift)
- {
+ {
/* Only threads with tidxj < 3 will update fshift.
The threads performing the update must be the same with the threads
which stored the reduction result in reduce_force_i function
#undef EL_EWALD_ANY
#undef EXCLUSION_FORCES
#undef LJ_EWALD
+
+#undef LJ_COMB
biod.pnpi.spb.ru / alexxy / gromacs.git / commitdiff