23.05
|
#include <performance_parameters.hpp>
Public Member Functions | |
PerformanceParameters (float k) | |
PerformanceParameters (float k, float p, float m) | |
Data Fields | |
float | kernel_macs_cycle |
float | prepare_bytes_cycle = 0.0f |
float | merge_bytes_cycle = 0.0f |
Definition at line 28 of file performance_parameters.hpp.
|
inline |
Definition at line 33 of file performance_parameters.hpp.
|
inline |
Definition at line 34 of file performance_parameters.hpp.
float kernel_macs_cycle |
Definition at line 29 of file performance_parameters.hpp.
Referenced by GemmHybrid< strategy, To, Tr >::estimate_cycles(), GemmInterleavedPretransposed2d< strategy, To, Tr >::estimate_cycles(), GemmHybridIndirect< strategy, To, Tr, OutputStage, SeparateQuantize, FixedFormat >::estimate_cycles(), and GemmInterleaved< strategy, To, Tr, OutputStage, MergeStep, FixedFormat, ForceThreadColumns, ForceFloatAccumulate >::estimate_cycles().
float merge_bytes_cycle = 0.0f |
Definition at line 31 of file performance_parameters.hpp.
Referenced by GemmInterleavedPretransposed2d< strategy, To, Tr >::estimate_cycles(), GemmHybridIndirect< strategy, To, Tr, OutputStage, SeparateQuantize, FixedFormat >::estimate_cycles(), and GemmInterleaved< strategy, To, Tr, OutputStage, MergeStep, FixedFormat, ForceThreadColumns, ForceFloatAccumulate >::estimate_cycles().
float prepare_bytes_cycle = 0.0f |
Definition at line 30 of file performance_parameters.hpp.
Referenced by GemmInterleavedPretransposed2d< strategy, To, Tr >::estimate_cycles(), GemmHybridIndirect< strategy, To, Tr, OutputStage, SeparateQuantize, FixedFormat >::estimate_cycles(), and GemmInterleaved< strategy, To, Tr, OutputStage, MergeStep, FixedFormat, ForceThreadColumns, ForceFloatAccumulate >::estimate_cycles().