|
SMAUG
Simulating Machine Learning Applications on gem5-Aladdin
|
| ▼Nsmaug | The smaug namespace is the parent namespace of all C++ code in SMAUG |
| ►Ngem5 | Contains utility functions for interacting with gem5 |
| ►Nsmv | The smv namespace contains all code specific to the Smv backend |
| CAvgPoolingOp | Implements the arithmetic-average-pooling operator |
| CBatchNormOp | Implements the batch normalization layer |
| CConcatOp | Concatenates N Tensors along a specified axis |
| CConvolutionOp | The base class for all 4D spatial convolution operators |
| CDataflowGraphWriter | DataflowGraphWriter writes the current network as a dot-graph file to the given ostream |
| CDataOp | Exposes a Tensor as its only output |
| CDebugStream | An stream class to consume debug logs |
| CDepthwiseConvolutionOp | Implements the depthwise convolution operator |
| CEltwiseAddOp | Adds two Tensors elementwise |
| CEltwiseMulOp | Multiplies two Tensors elementwise |
| CEltwiseOp | The base class of all elementwise operators |
| CEluOp | Implements the exponential linear unit function |
| CFlattenOp | Flattens each batch of a Tensor |
| CFromDataType | Provides compile-time conversion from SMAUG DataType to C type |
| CFromDataType< Bool > | |
| CFromDataType< Float16 > | |
| CFromDataType< Float32 > | |
| CFromDataType< Float64 > | |
| CFromDataType< Int32 > | |
| CFromDataType< Int64 > | |
| CFusedActivationOp | An Operator fused with an activation function |
| CGreaterEqualOp | Implements an elementwise greater than or equal to operator |
| CGreaterOp | Implements an elementwise greater than operator |
| CHardTanhOp | Implements the hard tanh operator, which bounds the min and max value of the tanh operator |
| CInnerProductOp | Implements the inner product operator |
| CLessEqualOp | Implements an elementwise less-than-or-equal-to operator |
| CLessOp | Implements an elementwise less-than operator |
| CMaxPoolingOp | Implements the max-pooling operator |
| CMergeOp | Forwards the first live input to its output |
| ►CNetwork | Network encapsulates all of the information SMAUG will use during execution: the overall computation graph of the model, all the operators and tensors, various housekeeping structures, and simulation information |
| COperator | Operator is the base class for all graph operators supported by SMAUG |
| CPaddingOp | Pad a given tensor in any number of dimensions with arbitrary size |
| CPoolingOp | Implements a pooling operator |
| CReferenceBackend | ReferenceBackend provides reference implementations of all operators supported by SMAUG |
| CReluOp | Implements the rectified linear unit operator: max(slope * x, 0) |
| CReorderOp | Implements a Tensor reordering operation to convert between different DataLayouts |
| CRepeatOp | Replicates a Tensor's data among all dimensions |
| CReshapeOp | Changes the Tensor's shape while retaining the number of elements |
| CScheduler | Scheduler is responsible for running the Network |
| CSeluOp | Implements the scaled exponential linear unit function |
| CSigmoidOp | Implements the sigmoid operator, defined as 1/(1 + exp(-input)) |
| CSmaugTest | The Catch2 test fixture used by all C++ unit tests |
| CSmvAcceleratorPool | Implements a pool of worker accelerators |
| CSmvAvgPoolingOp | Average pooling operator on SMV |
| CSmvBackend | SmvBackend implements a set of models of optimized DL kernels that were taped out on a machine learning SoC by the Harvard Architecture, Circuits, and Compilers |
| CSmvBatchNormOp | SMV backend implementation of batch normalization |
| CSmvConvolutionOp | SMV backend implementation of convolution |
| CSmvEltwiseAddOp | Elementwise addition on SMV |
| CSmvEltwiseMulOp | Elementwise multiplication on SMV |
| CSmvEluOp | Elementwise exponential linear unit on SMV |
| CSmvGreaterEqualOp | Elementwise greater-than-or-equal-to operator on SMV |
| CSmvGreaterOp | Elementwise greater-than operator on SMV |
| CSmvHardTanhOp | Hard tanh operator on SMV |
| CSmvInnerProductOp | Inner product operator on SMV |
| CSmvLessEqualOp | Elementwise less-than-or-equal-to operator on SMV |
| CSmvLessOp | Elementwise less-than operator on SMV |
| CSmvMaxPoolingOp | Max-pooling operator on SMV |
| CSmvPoolingOp | Base class for SMV pooling oeprators |
| CSmvReluOp | Rectified linear-unit operator on SMV |
| CSmvSeluOp | Elementwise scaled exponential linear unit on SMV |
| CSmvSigmoidOp | Sigmoid linear-unit operator on SMV |
| CSmvSoftmaxOp | Softmax operator on SMV |
| CSmvTanhOp | Tanh operator on SMV |
| CSoftmaxOp | Implements the softmax operator |
| CSplitOp | Implements the split operator, which divides a Tensor into N output Tensors along a specified dimension |
| CSwitchOp | Conditionally forwards an input to one of two outputs |
| CTanhOp | Implements the tanh operator |
| CTensor | Tensor represents a single multi-dimensional array of data |
| CTensorBase | The base class of all Tensor objects |
| CTensorIndexIterator | An iterator over a multidimensional tensor's indices, accounting for data alignment padding |
| CTensorIndices | Additional metadata for edges in the graph |
| CTensorRegionIndexIterator | A tensor index iterator that stays within a specified rectangular region |
| CTensorShape | TensorShape describes the shape of a Tensor |
| ►CThreadPool | A user-space cooperatve thread pool implementation designed for gem5 in SE mode |
| ►CTiledTensor | A multidimensional container of Tensors |
| CToDataType | Provides compile-time conversion from C types to SMAUG DataTypes |
| CToDataType< bool > | |
| CToDataType< double > | |
| CToDataType< float > | |
| CToDataType< float16 > | |
| CToDataType< int32_t > | |
| CToDataType< int64_t > | |
| CToDataType< uint32_t > | |
| CToDataType< uint64_t > | |
| CUnaryOp | Base class for all operators with one input |
| CWorkspace | Workspace is the container and owner of all Tensors and Operators in the Network |
| C_activation_param_t | Parameters to the activation function hardware |
| C_SamplingInfo | Simulation sampling information maintained by the Operator and passed to the accelerated kernel |
| CActivationInfo | Specifies an activation function and relevant parameters |
1.8.18