Home
last modified time | relevance | path

Searched refs:flops (Results 1 – 9 of 9) sorted by relevance

/libCEED/interface/ !
H A Dceed-qfunction.c642 int CeedQFunctionGetFlopsEstimate(CeedQFunction qf, CeedSize *flops) { in CeedQFunctionGetFlopsEstimate() argument
643 *flops = qf->user_flop_estimate; in CeedQFunctionGetFlopsEstimate()
1036 int CeedQFunctionSetUserFlopsEstimate(CeedQFunction qf, CeedSize flops) { in CeedQFunctionSetUserFlopsEstimate() argument
1037 …CeedCheck(flops >= 0, CeedQFunctionReturnCeed(qf), CEED_ERROR_INCOMPATIBLE, "Must set non-negative… in CeedQFunctionSetUserFlopsEstimate()
1038 qf->user_flop_estimate = flops; in CeedQFunctionSetUserFlopsEstimate()
H A Dceed-basis.c938 CeedSize *flops) { in CeedBasisGetFlopsEstimate() argument
982 *flops = 0; in CeedBasisGetFlopsEstimate()
985 …*flops = tensor_flops + num_points * num_comp * (point_tensor_flops + (t_mode == CEED_TRANSPOSE ? … in CeedBasisGetFlopsEstimate()
987 *flops += num_points * Q_1d * in CeedBasisGetFlopsEstimate()
990 *flops += num_points * (is_gpu ? num_comp : 1) * dim * chebyshev_flops; in CeedBasisGetFlopsEstimate()
995 …*flops = tensor_flops + num_points * num_comp * (point_tensor_flops + (t_mode == CEED_TRANSPOSE ? … in CeedBasisGetFlopsEstimate()
1000 …*flops += num_points * Q_1d * (chebyshev_flops + d_chebyshev_flops + num_comp * (inner_flops + (t_… in CeedBasisGetFlopsEstimate()
1002 …*flops += num_points * (is_gpu ? num_comp : 1) * dim * (d_chebyshev_flops + (dim - 1) * chebyshev_… in CeedBasisGetFlopsEstimate()
1015 *flops = num_points; in CeedBasisGetFlopsEstimate()
1021 *flops = 0; in CeedBasisGetFlopsEstimate()
[all …]
H A Dceed-operator.c1828 int CeedOperatorGetFlopsEstimate(CeedOperator op, CeedSize *flops) { in CeedOperatorGetFlopsEstimate() argument
1833 *flops = 0; in CeedOperatorGetFlopsEstimate()
1847 *flops += suboperator_flops; in CeedOperatorGetFlopsEstimate()
1899 *flops += rstr_flops; in CeedOperatorGetFlopsEstimate()
1904 *flops += basis_flops * num_elem; in CeedOperatorGetFlopsEstimate()
1921 *flops += num_elem * num_qpts * qf_flops; in CeedOperatorGetFlopsEstimate()
1938 *flops += rstr_flops; in CeedOperatorGetFlopsEstimate()
1943 *flops += basis_flops * num_elem; in CeedOperatorGetFlopsEstimate()
H A Dceed-elemrestriction.c576 …mRestrictionGetFlopsEstimate(CeedElemRestriction rstr, CeedTransposeMode t_mode, CeedSize *flops) { in CeedElemRestrictionGetFlopsEstimate() argument
613 *flops = e_size * scale; in CeedElemRestrictionGetFlopsEstimate()
/libCEED/include/ceed/ !
H A Dbackend.h324 …emRestrictionGetFlopsEstimate(CeedElemRestriction rstr, CeedTransposeMode t_mode, CeedSize *flops);
351 CeedSize *flops);
397 CEED_EXTERN int CeedQFunctionGetFlopsEstimate(CeedQFunction qf, CeedSize *flops);
H A Dceed.h388 CEED_EXTERN int CeedQFunctionSetUserFlopsEstimate(CeedQFunction qf, CeedSize flops);
488 CEED_EXTERN int CeedOperatorGetFlopsEstimate(CeedOperator op, CeedSize *flops);
/libCEED/julia/LibCEED.jl/src/generated/ !
H A Dlibceed_bindings.jl558 function CeedQFunctionSetUserFlopsEstimate(qf, flops) argument
559 ccall((:CeedQFunctionSetUserFlopsEstimate, libceed), Cint, (CeedQFunction, CeedSize), qf, flops)
781 function CeedOperatorGetFlopsEstimate(op, flops) argument
782 ccall((:CeedOperatorGetFlopsEstimate, libceed), Cint, (CeedOperator, Ptr{CeedSize}), op, flops)
1084 function CeedElemRestrictionGetFlopsEstimate(rstr, t_mode, flops) argument
1085 …mate, libceed), Cint, (CeedElemRestriction, CeedTransposeMode, Ptr{CeedSize}), rstr, t_mode, flops)
1118 function CeedBasisGetFlopsEstimate(basis, t_mode, eval_mode, flops) argument
1119 …Cint, (CeedBasis, CeedTransposeMode, CeedEvalMode, Ptr{CeedSize}), basis, t_mode, eval_mode, flops)
1250 function CeedQFunctionGetFlopsEstimate(qf, flops) argument
1251 … ccall((:CeedQFunctionGetFlopsEstimate, libceed), Cint, (CeedQFunction, Ptr{CeedSize}), qf, flops)
/libCEED/doc/papers/joss/ !
H A Dpaper.md69 …ultiply and add for each nonzero entry, just 2 floating point operations (flops) per scalar that m…
70 Modern hardware is capable of nearly 100 flops per scalar streamed from memory [@kruppcomparison] s…
/libCEED/examples/solids/ !
H A Dindex.md523 For low-order elements where flops tend to be in surplus relative to memory bandwidth, recomputatio…