Skip to main content
Back to top
Ctrl
+
K
tilus documentation
Getting Started
1. Installation
2. Tutorials
2.1. Matrix multiplication
2.1.1. Naive Matmul
2.1.2. Use Shared Memory
2.1.3. Auto-tuning
2.1.4. Use Async Copy
2.1.5. Software Pipelining
2.1.6. Split-K
Programming Guides
1. Overview
2. Tilus Script
3. Type System
3.1. Scalar Types
3.2. Pointer Types
3.3. Register Tensor
3.4. Shared Tensor
3.5. Global Tensor
4. Instructions
5. Control Flow
6. Cache
7. Autotuning
8. Layout System
8.1. Global Layout
8.2. Shared Layout
8.3. Register Layout
8.4. Automatic Layout Inference
Python API
tilus
tilus.autotune
tilus.option
tilus.option.cache_dir
tilus.option.parallel_workers
tilus.option.debug.dump_ir
tilus.option.debug.launch_blocking
tilus.Script
tilus.Script.attrs
tilus.Script.blockIdx
tilus.Script.gridDim
tilus.Script.assume
tilus.Script.range
tilus.Script.abs
tilus.Script.add
tilus.Script.annotate_layout
tilus.Script.assign
tilus.Script.cast
tilus.Script.copy_async
tilus.Script.copy_async_commit_group
tilus.Script.copy_async_wait_all
tilus.Script.copy_async_wait_group
tilus.Script.dot
tilus.Script.exp
tilus.Script.exp2
tilus.Script.free_shared
tilus.Script.global_tensor
tilus.Script.global_view
tilus.Script.load_global
tilus.Script.load_shared
tilus.Script.lock_semaphore
tilus.Script.max
tilus.Script.maximum
tilus.Script.min
tilus.Script.print_tensor
tilus.Script.printf
tilus.Script.register_tensor
tilus.Script.release_semaphore
tilus.Script.repeat
tilus.Script.repeat_interleave
tilus.Script.round
tilus.Script.shared_tensor
tilus.Script.squeeze
tilus.Script.store_global
tilus.Script.store_shared
tilus.Script.sum
tilus.Script.sync
tilus.Script.transpose
tilus.Script.unsqueeze
tilus.Script.view
tilus.Script.where
tilus.lang.Attributes.blocks
tilus.lang.Attributes.warps
tilus.ir
tilus.ir.DataType
tilus.ir.PointerType
tilus.ir.Tensor
tilus.ir.RegisterTensor
tilus.ir.SharedTensor
tilus.ir.GlobalTensor
tilus.ir.RegisterLayout
tilus.ir.layout.auto_local_spatial
tilus.ir.layout.column_local
tilus.ir.layout.column_spatial
tilus.ir.layout.compose
tilus.ir.layout.concat
tilus.ir.layout.divide
tilus.ir.layout.flatten
tilus.ir.layout.local
tilus.ir.layout.permute
tilus.ir.layout.reduce
tilus.ir.layout.register_layout
tilus.ir.layout.reshape
tilus.ir.layout.spatial
tilus.ir.layout.squeeze
tilus.ir.layout.unsqueeze
tilus.ir.SharedLayout
tilus.ir.layout.shared_row_major
tilus.ir.layout.shared_column_major
tilus.ir.layout.shared_compose
tilus.ir.GlobalLayout
tilus.ir.layout.global_row_major
tilus.ir.layout.global_column_major
tilus.ir.layout.global_strides
tilus.ir.layout.global_compose
.rst
.pdf
Low-Precision Support
Low-Precision Support
¶