Tensor Module - Aggregation of Tensor Field Infrastructure
This module re-exports all tensor-related submodules, providing a single import point for the full tensor field stack:
- globaltensor — TensorField[D,R,L,T]: distributed fields with ghost regions, GlobalShifter transport, and discrete Laplacian
- localtensor — LocalTensorField[D,R,L,T]: rank-local host-memory views of global fields
- tensorview — TensorFieldView[L,T]: device-side views for the each macro (OpenCL / SYCL / OpenMP)
- stencil — LatticeStencil[D]: unified stencil operations across all backends
Data flows through the stack as:
TensorField → LocalTensorField → TensorFieldView → Kernels (GA/MPI) (host RAM) (device buffers) (each macro)
For distributed-memory shifting and stencil operations that do not require device dispatch, use GlobalShifter and discreteLaplacian from globaltensor directly.
Exports
-
newLatticeStencil, GlobalShifter, localGrid, lexToCoords, coordsToPaddedFlat, accessGhosts, localLexIdx, updateGhosts, localSiteOffset, newTensorField, innerPaddedOffset, coordsToLex, updateAllGhosts, innerPaddedBlockSize, releaseLocal, localIdx, discreteLaplacian, ghostWidth, accessLocal, paddedGrid, innerBlockSize, apply, newGlobalShifters, TensorField, newGlobalShifter, hasGhosts, newGlobalBackwardShifters, applyStencilShift, numSites, [], []=, []=, tensorElementsPerSite, []=, HostStorage, numGlobalSites, LocalTensorField, []=, []=, numElements, newLocalTensorField, getSite, []=, enqueueReadBuffer, DEVICE_VENDOR_ID, buildProgram, neighborOffset, fmod, enqueueFillBuffer, StencilView, COMMAND_COPY_BUFFER, createCommandQueue, getImageInfo, DEVICE_AFFINITY_DOMAIN_NEXT_PARTITIONABLE, []=, KERNEL_ARG_ADDRESS_CONSTANT, nPoints, KERNEL_ARG_ADDRESS_GLOBAL, $, createUserEvent, Vec, CONTEXT_INTEROP_USER_SYNC, KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE, trace, PROGRAM_BINARY_TYPE_LIBRARY, FILTER_NEAREST, GLOBAL, transformAoSoAtoAoS, KERNEL_ARG_TYPE_RESTRICT, PROGRAM_DEVICES, createContextFromType, EVENT_COMMAND_QUEUE, EVENT_COMMAND_TYPE, Tmem_object_type, DEVICE_IMAGE_MAX_BUFFER_SIZE, DEVICE_PARTITION_BY_COUNTS, write_mem_fence_impl, EVENT_COMMAND_EXECUTION_STATUS, enqueueMapBuffer, ==, $, Tdevice_exec_capabilities, nSitesInner, round, CL_SIGNED_INT32, PLATFORM_NAME, COMPLETE, atomic_and_impl, *, finalizeCL, formatSiteData, transpose, UseSYCL, matvec, PLATFORM_PROFILE, vecSize, COMMAND_MAP_IMAGE, MEM_READ_WRITE, Tmem_info, MEM_OBJECT_IMAGE1D, matmul, KERNEL_CONTEXT, PROGRAM_BINARY_TYPE_EXECUTABLE, DEVICE_MAX_COMPUTE_UNITS, COMMAND_WRITE_BUFFER_RECT, pathToStencil, DEVICE_PARTITION_TYPE, getEventInfo, newSimdLatticeLayout, enqueueTask, enqueueReadBufferRect, bwd, getDeviceIDs, [], Vec3f, native_tan_impl, getCommandQueueInfo, retainMemObject, MEM_OBJECT_IMAGE1D_BUFFER, Tcommand_queue_properties, createProgramWithSource, forwardStencil, StencilPattern, laplacianStencil, newStencil, PROGRAM_KERNEL_NAMES, release, native_exp, DEVICE_AFFINITY_DOMAIN_L2_CACHE, PROGRAM_BINARY_TYPE, nearestNeighborStencil, DeviceStorage, plaquettePath, DEVICE_OPENCL_C_VERSION, native_log10, simdLanes, CL_A, COMMAND_FILL_IMAGE, =copy, MEM_HOST_WRITE_ONLY, readSiteData, StencilEntry, clamp, Tdevice_mem_cache_type, enqueueReadImage, args, DEVICE_MAX_CONSTANT_ARGS, MAP_READ, floor, nSites, native_sin, PROGRAM_BINARY_TYPE_COMPILED_OBJECT, MEM_HOST_READ_ONLY, EVENT_CONTEXT, Y, +, DEVICE_NATIVE_VECTOR_WIDTH_SHORT, COMMAND_WRITE_BUFFER, each, Timage_format, name, *, transformAoSoAtoAoSSimd, global, DEVICE_IMAGE3D_MAX_DEPTH, idx, PROGRAM_NUM_KERNELS, -, elementType, native_powr_impl, [], VectorWidth, DEVICE_MAX_WORK_ITEM_DIMENSIONS, COMMAND_MAP_BUFFER, getProgramBuildInfo, []=, compileProgram, setArg, retainProgram, DEVICE_PLATFORM, ADDRESS_MIRRORED_REPEAT, gpuBuffer, Vec3d, -, waitForEvents, -, MEM_OBJECT_IMAGE2D, enqueueUnmapMemObject, LocalBuffer, MEM_ALLOC_HOST_PTR, Tkernel_arg_info, StencilShift, KERNEL_ARG_TYPE_CONST, getDevices, Timage_info, CL_RG, coordsToLexicographic, DEVICE_TYPE_ALL, finish, newLatticeStencil, createProgramWithBinary, +, MIGRATE_MEM_OBJECT_CONTENT_UNDEFINED, DEPTH_STENCIL, Dim, IMAGE_BUFFER, CL_RGx, releaseContext, PathStep, DEVICE_PARTITION_MAX_SUB_DEVICES, Mat3, DEVICE_PARENT_DEVICE, enqueueMapImage, $, multipleDeviceDefaults, Tcommand_queue_info, DEVICE_MAX_READ_IMAGE_ARGS, createProgram, DEVICE_EXECUTION_CAPABILITIES, mem_fence, DEVICE_NATIVE_VECTOR_WIDTH_INT, $, DEVICE_PARTITION_BY_AFFINITY_DOMAIN, LocalAddResult, QUEUE_PROFILING_ENABLE, getKernelArgInfo, DEVICE_TYPE_CPU, DEVICE_GLOBAL_MEM_CACHE_TYPE, KERNEL_PRIVATE_MEM_SIZE, FP_ROUND_TO_ZERO, IOKind, Vec4d, native_exp2, PROGRAM_BUILD_STATUS, trunc, COMMAND_TASK, COMMAND_READ_BUFFER_RECT, ceil, retainDevice, isSiteTensor, createKernelsInProgram, +, -, run2d, DEVICE_PREFERRED_VECTOR_WIDTH_LONG, Tkernel_work_group_info, getSupportedImageFormats, Tcontext_properties, releaseDevice, TensorSiteProxy, computeStrides, buildOn, native_recip_impl, ScalarMulResult, MEM_SIZE, DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE, native_rsqrt, releaseProgram, newLatticeStencil, MEM_USE_HOST_PTR, Vec3, globalMemory, Tkernel_arg_type_qualifier, PROGRAM_REFERENCE_COUNT, createImage2D, Tprogram_info, CL_RGBA, MAP_WRITE, elementsPerSite, DEVICE_IMAGE2D_MAX_HEIGHT, []=, []=, forEachNeighbor, IMAGE_FORMAT, UseSycl, *, write, Vec4f, BLOCKING, NON_BLOCKING, KERNEL_ARG_ACCESS_READ_ONLY, defaultSimdGrid, newStencilPattern, Tmem_flags, DEVICE_IMAGE3D_MAX_WIDTH, Tbuild_status, native_log2_impl, release, EXEC_NATIVE_KERNEL, IMAGE_WIDTH, setArg, atomic_cmpxchg_impl, buildOn, DEVICE_VENDOR, CLK_LOCAL_MEM_FENCE, [], native_cos, check, KERNEL_ARG_ACCESS_READ_WRITE, fmin, FP_ROUND_TO_NEAREST, getContextInfo, KERNEL_PROGRAM, CL_SIGNED_INT8, CL_RGBx, +, getKernelWorkGroupInfo, MEM_WRITE_ONLY, DEPTH, Vec2, getEntry, +, DEVICE_EXTENSIONS, DEVICE_LOCAL_MEM_SIZE, DEVICE_QUEUE_PROPERTIES, UseOpenMP, IMAGE_ARRAY_SIZE, CL_HALF_FLOAT, KERNEL_WORK_GROUP_SIZE, PROGRAM_BUILD_LOG, *, splitLatticeSites, acos, Tchannel_order, computeTotalLatticeSites, Tdevice_partition_property, +, read_mem_fence, Tplatform_info, createSampler, IMAGE_NUM_MIP_LEVELS, raiseEOpenCL, enqueueCopyBufferToImage, IMAGE_ELEMENT_SIZE, native_sqrt_impl, flush, DEVICE_TYPE_GPU, COMMAND_MARKER, createAndBuild, computeLocalGeom, Pkernel, matadd, sitesPerDevice, +, FP_SOFT_FLOAT, DEVICE_AFFINITY_DOMAIN_L3_CACHE, atomic_max, MEM_OBJECT_BUFFER, DEVICE_PARTITION_PROPERTIES, get_work_dim, createImage, fwd, Tsampler_info, ADDRESS_CLAMP_TO_EDGE, DEVICE_NATIVE_VECTOR_WIDTH_HALF, enqueueBarrier, tanh, T, DEVICE_AFFINITY_DOMAIN_NUMA, Tbool, numElements, createKernel, []=, Mat, Tprofiling_info, -, totalElements, SAMPLER_FILTER_MODE, $, DEVICE_PARTITION_AFFINITY_DOMAIN, read, COMMAND_COPY_BUFFER_RECT, rsqrt_impl, *, transformAoStoAoSoA, clamp_impl, Tdevice_info, LatticeStencil, write, native_recip, getPlatformByName, RuntimeSiteData, DEVICE_PREFERRED_VECTOR_WIDTH_HALF, run, *, Tmap_flags, TensorElementProxy, DEVICE_IMAGE_MAX_ARRAY_SIZE, bwd, retainContext, paddedToLocal, defaultSimdGrid, setMemObjectDestructorCallback, getPlatformIDs, kernel, [], SAMPLER_CONTEXT, rectanglePath, BUILD_ERROR, TUserCb, COMMAND_WRITE_IMAGE, TMemObjectDestructorCb, DEVICE_TYPE_ACCELERATOR, Tprogram_build_info, MEM_HOST_PTR, atomic_xchg, release, numSites, newStencilPoint, DEVICE_IMAGE_BASE_ADDRESS_ALIGNMENT, native_sin_impl, newSimdLatticeLayout, TEventCb, COMMAND_USER, release, localToPadded, forwardStencil, PROGRAM_CONTEXT, BackendQueue, BUILD_IN_PROGRESS, CL_Rx, atomic_or_impl, run3d, numVectorGroups, +, COMMAND_MIGRATE_MEM_OBJECTS, elementType, nSitesOuter, maxWorkGroups, []=, setArg, Pdevice_id, GpuBuffer, []=, native_powr, DEVICE_IMAGE_PITCH_ALIGNMENT, TClResult, DEVICE_MAX_PARAMETER_SIZE, hasSimdLayout, enqueueCopyImage, QUEUE_DEVICE, aosoaIndex, CL_UNORM_INT_101010, Mat3f, fmax, DEVICE_PREFERRED_INTEROP_USER_SYNC, =destroy, newLatticeStencil, MEM_OBJECT_IMAGE2D_ARRAY, Stencil, ScalarAddResult, native_log, sinh, release, CL_BGRA, MEM_TYPE, barrier, +, DEVICE_MAX_WORK_GROUP_SIZE, CONTEXT_PLATFORM, atomic_dec, +, directions, $, +, DRIVER_VERSION, CL_SNORM_INT16, Tdevice_affinity_domain, writeSiteElement, Tevent_info, BUFFER_CREATE_TYPE_REGION, Tbuffer_create_type, VERSION_1_2, eachImpl, DEVICE_MAX_SAMPLERS, numGlobalSites, Tbuffer_region, CONTEXT_NUM_DEVICES, Tdevice_fp_config, points, DEVICE_PREFERRED_VECTOR_WIDTH_SHORT, PROGRAM_SOURCE, FP_INF_NAN, release, DEVICE_MAX_CONSTANT_BUFFER_SIZE, +, MEM_FLAGS, nOuter, BUILD_NONE, BackendBuffer, MIGRATE_MEM_OBJECT_HOST, CONTEXT_DEVICES, enqueueMigrateMemObjects, transformAoStoAoSoASimd, DEVICE_PROFILE, buildErrors, forward, nearestNeighborStencil, []=, LocalSiteProxy, CL_TRUE, [], DEVICE_PREFERRED_VECTOR_WIDTH_FLOAT, LocalScalarMulResult, hash, Tdevice_local_mem_type, createProgramWithBuiltInKernels, [], isGhostNeighbor, Tkernel_arg_address_qualifier, createSubDevices, READ_WRITE_CACHE, RUNNING, shift, setKernelArg, DEVICE_PREFERRED_VECTOR_WIDTH_INT, BUILD_SUCCESS, native_sqrt, QUEUE_REFERENCE_COUNT, DEVICE_ERROR_CORRECTION_SUPPORT, -, log10, INTENSITY, atomic_or, DEVICE_GLOBAL_MEM_CACHE_SIZE, matRows, []=, atomic_inc_impl, exp, SUBMITTED, PROGRAM_BUILD_OPTIONS, FP_DENORM, KERNEL_ARG_ADDRESS_LOCAL, VERSION_1_0, KERNEL_ARG_TYPE_NONE, version, Mat2, simdLayout, UseOpenMP, DEVICE_AVAILABLE, DEVICE_SINGLE_FP_CONFIG, KERNEL_NUM_ARGS, linkProgram, DEVICE_TYPE_DEFAULT, identity, PROGRAM_NUM_DEVICES, atan, createBuffer, commandQueueFor, CL_UNORM_INT16, [], Z, IMAGE_NUM_SAMPLES, CONTEXT_PROPERTIES, getExtensionFunctionAddressForPlatform, retainKernel, IMAGE_SLICE_PITCH, MatAddResult, asin, *, TCreateContextCb, addPoint, Pcommand_queue, neighborSimd, addPoint, MEM_HOST_NO_ACCESS, DEVICE_IMAGE2D_MAX_WIDTH, atan2, QUEUE_PROPERTIES, []=, step, run3d, Tcommand_type, backwardStencil, []=, enqueueBarrierWithWaitList, native_exp2_impl, SAMPLER_ADDRESSING_MODE, Tkernel_arg_access_qualifier, releaseSampler, $, DEVICE_MAX_CLOCK_FREQUENCY, enqueueWriteBufferRect, UseSycl, StencilPoint, +, CL_UNSIGNED_INT8, COMMAND_UNMAP_MEM_OBJECT, log, COMMAND_COPY_IMAGE_TO_BUFFER, CLK_GLOBAL_MEM_FENCE, SAMPLER_NORMALIZED_COORDS, Taddressing_mode, Pprogram, pow, native_log_impl, *, unloadCompiler, KERNEL_GLOBAL_WORK_SIZE, enqueueCopyBufferRect, localToOuterInner, createContext, neighbors, SAMPLER_REFERENCE_COUNT, DEVICE_IMAGE3D_MAX_HEIGHT, singleDeviceDefaults, get_group_id, ADDRESS_REPEAT, atomic_min, Pplatform_id, storageType, +, run2d, COMMAND_RELEASE_GL_OBJECTS, Pevent, DEVICE_TYPE_CUSTOM, []=, atomic_sub_impl, LocalScalarAddResult, [], createAndBuild, releaseKernel, unloadPlatformCompiler, createAndBuildBinary, get_num_groups, atomic_max_impl, atomic_sub, atomic_cmpxchg, KERNEL_ARG_ADDRESS_QUALIFIER, DEVICE_REFERENCE_COUNT, DEVICE_MEM_BASE_ADDR_ALIGN, $, name, CL_UNORM_INT24, run, *, getEventProfilingInfo, PLATFORM_EXTENSIONS, PLATFORM_VENDOR, FP_ROUND_TO_INF, atomic_xor_impl, CL_SNORM_INT8, +, ElementType, []=, []=, Vec2f, newLatticeStencil, COMMAND_ACQUIRE_GL_OBJECTS, Tchannel_type, []=, retainSampler, +, VecAddResult, DEVICE_MAX_WORK_ITEM_SIZES, TensorFieldView, get_local_id, DEVICE_PREFERRED_VECTOR_WIDTH_CHAR, CL_R, DEVICE_BUILT_IN_KERNELS, NONE, getPlatformInfo, write_mem_fence, -, COMMAND_COPY_IMAGE, Mat3d, CL_UNSIGNED_INT16, TDeviceType, EXEC_KERNEL, DEVICE_PARTITION_EQUALLY, enqueueWriteBuffer, enqueueNDRangeKernel, DEVICE_ENDIAN_LITTLE, setArg, fwd, ADDRESS_NONE, get_global_id, newStencilView, TProgramCb, buffers, enqueueCopyBuffer, CL_UNORM_SHORT_555, KERNEL_COMPILE_WORK_GROUP_SIZE, newStencilPoint, X, Mat4f, [], UseWorkGroups, COMMAND_READ_BUFFER, setEventCallback, VectorWidth, log2, COMMAND_NDRANGE_KERNEL, atomic_min_impl, DEVICE_AFFINITY_DOMAIN_L4_CACHE, Vec4, SimdLatticeLayout, UseOpenCL, MatVecResult, cosh, releaseEvent, SignedDirection, Tcontext_info, rsqrt, buffer, Tfilter_mode, generateCoordTable, KERNEL_ARG_TYPE_VOLATILE, DebugKernels, aosoaIndexFromLocal, COMMAND_BARRIER, -, []=, offsetBufferSize, DEVICE_NAME, Tprogram_binary_type, COMMAND_FILL_BUFFER, native_exp_impl, validateSimdGrid, UseOpenMP, numElements, enqueueWaitForEvents, PROGRAM_BINARY_TYPE_NONE, createKernel, retainEvent, enqueueWriteImage, outerInnerToLocal, localMemory, Tkernel_info, isMat, []=, CL_FALSE, StencilBackend, PLATFORM_VERSION, tan, DEVICE_MAX_MEM_ALLOC_SIZE, getSamplerInfo, retainCommandQueue, KERNEL_ARG_ADDRESS_PRIVATE, *, -, setUserEventStatus, enqueueFillImage, releaseMemObject, getProgramInfo, MEM_MAP_COUNT, Timage_desc, COMMAND_COPY_BUFFER_TO_IMAGE, KERNEL_ARG_ACCESS_WRITE_ONLY, fma, KERNEL_ARG_TYPE_QUALIFIER, bufferLike, device, computeElementsPerSite, matCols, DEVICE_GLOBAL_MEM_SIZE, getKernelInfo, DEVICE_ADDRESS_BITS, =copy, Mat2f, backward, laplacianStencil, createContext, -, gpuBufferLike, exp2, DEVICE_AFFINITY_DOMAIN_L1_CACHE, -, maxWorkItems, DEVICE_NATIVE_VECTOR_WIDTH_LONG, getMemObjectInfo, atomic_add, DEVICE_NATIVE_VECTOR_WIDTH_CHAR, sites, get_global_size, QUEUED, MEM_ASSOCIATED_MEMOBJECT, EVENT_REFERENCE_COUNT, computeProduct, Tmem_migration_flags, neighbor, -, DEVICE_PARTITION_BY_COUNTS_LIST_END, Vec2d, DEVICE_MAX_WRITE_IMAGE_ARGS, MEM_CONTEXT, write, atomic_inc, PROGRAM_BINARIES, CL_UNORM_SHORT_565, createSubBuffer, IMAGE_DEPTH, CL_UNSIGNED_INT32, atomic_and, *, getOffsetBuffer, KERNEL_ATTRIBUTES, Direction, [], Psampler, openclDefaults, read, DEVICE_PROFILING_TIMER_RESOLUTION, IMAGE_ROW_PITCH, sqrt, Mat2d, EOpenCL, VERSION_1_1, MatMulResult, DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE, mem_fence_impl, QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, MAP_WRITE_INVALIDATE_REGION, localToPadded, CONTEXT_REFERENCE_COUNT, FILTER_LINEAR, constant, Mat4d, MEM_COPY_HOST_PTR, DEVICE_GLOBAL_MEM_CACHELINE_SIZE, COMMAND_READ_IMAGE, isVec, atomic_add_impl, native_tan, MEM_OBJECT_IMAGE3D, aosoaDataPtr, getDeviceInfo, DEVICE_COMPILER_AVAILABLE, IMAGE_HEIGHT, DEVICE_IMAGE_SUPPORT, [], shift, releaseCommandQueue, DEVICE_LINKER_AVAILABLE, DEVICE_NATIVE_VECTOR_WIDTH_FLOAT, allDirections, backwardStencil, native_cos_impl, KERNEL_FUNCTION_NAME, barrier_impl, sin, firstPlatform, createProgramBinary, KERNEL_ARG_ACCESS_NONE, *, get_local_size, KERNEL_ARG_ACCESS_QUALIFIER, DEVICE_LOCAL_MEM_TYPE, native_log10_impl, *, enqueueNativeKernel, Tbitfield, Mat4, enqueueCopyImageToBuffer, *, CL_SIGNED_INT16, addPoint, enqueueMarker, LocalMulResult, MEM_OFFSET, DEVICE_PRINTF_BUFFER_SIZE, ADDRESS_CLAMP, newLatticeStencil, KERNEL_REFERENCE_COUNT, atomic_dec_impl, CL_ARGB, CL_FLOAT, nLanes, initCL, getExtensionFunctionAddress, createImage3D, setArg, enqueueMarkerWithWaitList, read_mem_fence_impl, []=, vecadd, paddedToLocal, lexicographicToCoords, MEM_READ_ONLY, DEVICE_HOST_UNIFIED_MEMORY, Pmem, FP_FMA, native_log2, FP_CORRECTLY_ROUNDED_DIVIDE_SQRT, COMMAND_NATIVE_KERNEL, atomic_xchg_impl, DEVICE_TYPE, MEM_OBJECT_IMAGE1D_ARRAY, oclName, VectorWidth, native_rsqrt_impl, CL_RA, dot, PROGRAM_BINARY_SIZES, newTensorFieldView, CL_UNORM_INT8, QUEUE_CONTEXT, read, READ_ONLY_CACHE, get_global_offset, [], DEVICE_MIN_DATA_TYPE_ALIGN_SIZE, LUMINANCE, CL_RGB, DEVICE_VERSION, KERNEL_ARG_NAME, *, fabs, cos, MEM_REFERENCE_COUNT, KERNEL_LOCAL_MEM_SIZE, DEVICE_DOUBLE_FP_CONFIG, vectorGroups, newTensorFieldView, LOCAL, atomic_xor, *, Pcontext, KERNEL_ARG_TYPE_NAME, local, newLatticeStencil, newLatticeStencil, UseOpenCL, LatticeStencil, newLatticeStencil, nLanes, Y, pathToStencil, addPoint, fwd, ==, localToPadded, newStencilView, backward, hash, nPoints, newLatticeStencil, idx, Z, bwd, $, $, directions, $, shift, allDirections, neighborOffset, offsetBufferSize, rectanglePath, forEachNeighbor, addPoint, nearestNeighborStencil, paddedToLocal, bwd, forwardStencil, VectorWidth, neighborSimd, Stencil, nearestNeighborStencil, StencilPattern, laplacianStencil, Direction, nOuter, SignedDirection, backwardStencil, StencilEntry, StencilPoint, neighbor, step, newLatticeStencil, T, plaquettePath, StencilBackend, paddedToLocal, forwardStencil, fwd, backwardStencil, forward, points, laplacianStencil, getEntry, newStencilPattern, UseSYCL, newStencilPoint, X, getOffsetBuffer, $, localToPadded, shift, StencilShift, StencilView, nSites, newStencil, neighbors, newStencilPoint, addPoint, isGhostNeighbor, PathStep, sites, UseOpenMP