diff --git a/python/src/docstrings.hpp b/python/src/docstrings.hpp index 37f3ff7..6b3a1dc 100644 --- a/python/src/docstrings.hpp +++ b/python/src/docstrings.hpp @@ -513,10 +513,10 @@ function not in the record function.)doc"; static const char *__doc_kp_OpTensorSyncDevice = R"doc(Operation that syncs tensor's device by mapping local data into the device memory. For TensorTypes::eDevice it will use a staging tensor -to perform the copy. For TensorTypes::eStaging it will only copy the +to perform the copy. For TensorTypes::eHost it will only copy the data and perform a map, which will be executed during the record (as opposed to during the sequence eval/submit). This function cannot be -carried out for TensorTypes::eStaging.)doc"; +carried out for TensorTypes::eHost.)doc"; static const char *__doc_kp_OpTensorSyncDevice_OpTensorSyncDevice = R"doc()doc"; @@ -533,7 +533,7 @@ queues @param device Vulkan logical device for passing to Algorithm static const char *__doc_kp_OpTensorSyncDevice_init = R"doc(Performs basic checks such as ensuring that there is at least one tensor provided, that they are initialized and that they are not of -type TensorTpes::eStaging. For staging tensors in host memory, the map +type TensorTpes::eHost. For staging tensors in host memory, the map is performed during the init function.)doc"; static const char *__doc_kp_OpTensorSyncDevice_mStagingTensors = R"doc()doc"; @@ -549,11 +549,11 @@ from the temporary staging tensor.)doc"; static const char *__doc_kp_OpTensorSyncLocal = R"doc(Operation that syncs tensor's local data by mapping the data from device memory into the local vector. For TensorTypes::eDevice it will -use a staging tensor to perform the copy. For TensorTypes::eStaging it +use a staging tensor to perform the copy. For TensorTypes::eHost it will only copy the data and perform a map, which will be executed during the postSubmit (there will be no copy during the sequence eval/submit). This function cannot be carried out for -TensorTypes::eStaging.)doc"; +TensorTypes::eHost.)doc"; static const char *__doc_kp_OpTensorSyncLocal_OpTensorSyncLocal = R"doc()doc"; @@ -570,7 +570,7 @@ queues @param device Vulkan logical device for passing to Algorithm static const char *__doc_kp_OpTensorSyncLocal_init = R"doc(Performs basic checks such as ensuring that there is at least one tensor provided, that they are initialized and that they are not of -type TensorTpes::eStaging.)doc"; +type TensorTpes::eHost.)doc"; static const char *__doc_kp_OpTensorSyncLocal_mStagingTensors = R"doc()doc"; @@ -719,7 +719,7 @@ shader storage).)doc"; static const char *__doc_kp_Tensor_TensorTypes_eDevice = R"doc(< Type is device memory, source and destination)doc"; -static const char *__doc_kp_Tensor_TensorTypes_eStaging = R"doc(< Type is host memory, source and destination)doc"; +static const char *__doc_kp_Tensor_TensorTypes_eHost = R"doc(< Type is host memory, source and destination)doc"; static const char *__doc_kp_Tensor_TensorTypes_eStorage = R"doc(< Type is Device memory (only))doc"; diff --git a/python/src/main.cpp b/python/src/main.cpp index 09368f2..6e795fa 100644 --- a/python/src/main.cpp +++ b/python/src/main.cpp @@ -26,7 +26,7 @@ PYBIND11_MODULE(kp, m) { py::enum_(m, "TensorTypes", DOC(kp, Tensor, TensorTypes)) .value("device", kp::Tensor::TensorTypes::eDevice, "Tensor holding data in GPU memory.") - .value("staging", kp::Tensor::TensorTypes::eStaging, "Tensor used for transfer of data to device.") + .value("host", kp::Tensor::TensorTypes::eHost, "Tensor used for CPU visible GPU data.") .value("storage", kp::Tensor::TensorTypes::eStorage, "Tensor with host visible gpu memory.") .export_values(); @@ -112,7 +112,7 @@ PYBIND11_MODULE(kp, m) { .def("record_tensor_sync_device", &kp::Sequence::record, "Records operation to sync tensor from local memory to GPU memory") .def("record_tensor_sync_local", &kp::Sequence::record, - "Records operation to sync tensor(s) from GPU memory to local memory using staging tensors") + "Records operation to sync tensor(s) from GPU memory to local memory") .def("record_algo_mult", &kp::Sequence::record, "Records operation to run multiplication compute shader to two input tensors and an output tensor") .def("record_algo_file", [](kp::Sequence &self, @@ -179,7 +179,7 @@ PYBIND11_MODULE(kp, m) { .def("eval_tensor_sync_device_def", &kp::Manager::evalOpDefault, "Evaluates operation to sync tensor from local memory to GPU memory with new anonymous Sequence") .def("eval_tensor_sync_local_def", &kp::Manager::evalOpDefault, - "Evaluates operation to sync tensor(s) from GPU memory to local memory using staging tensors with new anonymous Sequence") + "Evaluates operation to sync tensor(s) from GPU memory to local memory with new anonymous Sequence") .def("eval_algo_mult_def", &kp::Manager::evalOpDefault, "Evaluates operation to run multiplication compute shader to two input tensors and an output tensor with new anonymous Sequence") .def("eval_algo_file_def", &kp::Manager::evalOpDefault, @@ -216,7 +216,7 @@ PYBIND11_MODULE(kp, m) { .def("eval_tensor_sync_device", &kp::Manager::evalOp, "Evaluates operation to sync tensor from local memory to GPU memory with explicitly named Sequence") .def("eval_tensor_sync_local", &kp::Manager::evalOp, - "Evaluates operation to sync tensor(s) from GPU memory to local memory using staging tensors with explicitly named Sequence") + "Evaluates operation to sync tensor(s) from GPU memory to local memory with explicitly named Sequence") .def("eval_algo_mult", &kp::Manager::evalOp, "Evaluates operation to run multiplication compute shader to two input tensors and an output tensor with explicitly named Sequence") .def("eval_algo_file", &kp::Manager::evalOp, @@ -256,7 +256,7 @@ PYBIND11_MODULE(kp, m) { .def("eval_async_tensor_sync_device_def", &kp::Manager::evalOpAsyncDefault, "Evaluates asynchronously operation to sync tensor from local memory to GPU memory with anonymous Sequence") .def("eval_async_tensor_sync_local_def", &kp::Manager::evalOpAsyncDefault, - "Evaluates asynchronously operation to sync tensor(s) from GPU memory to local memory using staging tensors with anonymous Sequence") + "Evaluates asynchronously operation to sync tensor(s) from GPU memory to local memory with anonymous Sequence") .def("eval_async_algo_mult_def", &kp::Manager::evalOpAsyncDefault, "Evaluates asynchronously operation to run multiplication compute shader to two input tensors and an output tensor with anonymous Sequence") .def("eval_async_algo_file_def", &kp::Manager::evalOpAsyncDefault, @@ -293,7 +293,7 @@ PYBIND11_MODULE(kp, m) { .def("eval_async_tensor_sync_device", &kp::Manager::evalOpAsync, "Evaluates asynchronously operation to sync tensor from local memory to GPU memory with explicitly named Sequence") .def("eval_async_tensor_sync_local", &kp::Manager::evalOpAsync, - "Evaluates asynchronously operation to sync tensor(s) from GPU memory to local memory using staging tensors with explicitly named Sequence") + "Evaluates asynchronously operation to sync tensor(s) from GPU memory to local memory with explicitly named Sequence") .def("eval_async_algo_mult", &kp::Manager::evalOpAsync, "Evaluates asynchronously operation to run multiplication compute shader to two input tensors and an output tensor with explicitly named Sequence") .def("eval_async_algo_file", &kp::Manager::evalOpAsync,