Docs | Tutorial | DiffTaichi | Examples | FAQ | Forum

Documentations	Chat	taichi-nightly	taichi-nightly-cuda-10-0	taichi-nightly-cuda-10-1

# Python 3.6/3.7 needed

# CPU only. No GPU/CUDA needed. (Linux, OS X and Windows)
python3 -m pip install taichi-nightly

# With GPU (CUDA 10.0) support (Linux only)
python3 -m pip install taichi-nightly-cuda-10-0

# With GPU (CUDA 10.1) support (Linux only)
python3 -m pip install taichi-nightly-cuda-10-1

	Linux (CUDA)	OS X (10.14)	Windows
Build
PyPI

Short-term goals

(Done) Fully implement the LLVM backend to replace the legacy source-to-source C++/CUDA backends (By Dec 2019)
- The only missing features compared to the old source-to-source backends:
  - Vectorization on CPUs. Given most users who want performance are using GPUs (CUDA), this is given low priority.
  - Automatic shared memory utilization. Postponed until Feb/March 2020.
(WIP) Tune the performance of the LLVM backend to match that of the legacy source-to-source backends (By the end of Jan 2020)
(WIP) Redesign memory allocator

Updates

(Jan 3, 2020) v0.3.20 released.
- Support for loops with ti.static(ti.grouped(ti.ndrange(...)))
(Jan 2, 2020) v0.3.19 released.
- Added ti.atan2(y, x)
- Improved error msg when using float point numbers as tensor indices
(Jan 1, 2020) v0.3.18 released.
- Added ti.GUI class
- Improved the performance of performance ti.Matrix.fill
(Dec 31, 2019) v0.3.17 released.
- Fixed cuda context conflict with PyTorch (thanks to @Xingzhe He for reporting)
- Support ti.Matrix.T() for transposing a matrix
- Iteratable ti.static(ti.ndrange)
- Fixed ti.Matrix.identity()
- Added ti.Matrix.one() (create a matrix with 1 as all the entries)
- Improved ir_printer on SNodes
- Better support for dynamic SNodes.
  - Struct-for's on dynamic nodes supported
  - ti.length and ti.append to query and manipulate dynamic nodes
(Dec 29, 2019) v0.3.16 released.
- Fixed ndrange-fors with local variables (thanks to Xingzhe He for reporting this issue)
(Dec 28, 2019) v0.3.15 released.
- Multi-dimensional parallel range-for using ti.ndrange:

  @ti.kernel
  def fill_3d():
    # Parallelized for all 3 <= i < 8, 1 <= j < 6, 0 <= k < 9
    for i, j, k in ti.ndrange((3, 8), (1, 6), 9):
      x[i, j, k] = i + j + k

(Dec 28, 2019) v0.3.14 released.
- GPU random number generator support for more than 1024x1024 threads
- Parallelized element list generation on GPUs. Struct-fors significantly sped up.
- ti and tid (debug mode) CLI commands
(Dec 26, 2019) v0.3.13 released.
- ti.append now returns the list length before appending
- Fixed for loops with 0 iterations
- Set ti.get_runtime().set_verbose_kernel_launch(True) to log kernel launches
- Distinguish / and // following the Python convention
- Allow using local variables as kernel argument type annotations
(Dec 25, 2019) v0.3.11 released.
- Support multiple kernels with the same name, especially in the OOP cases where multiple member kernels share the same name
- Basic dynamic node support (ti.append, ti.length) in the new LLVM backend
- Fixed struct-for loops on 0-D tensors
(Dec 24, 2019) v0.3.10 released.
- assert <condition> statement supported in Taichi kernels.
- Comparison operator chaining (e.g. 1 < x <3) supported in Taichi kernels.
(Dec 24, 2019) v0.3.9 released.
- ti.classfunc decorator for functions within a data_oriented class
- [Expr/Vector/Matrix].to_torch now has a extra argument device, which specifies the device placement for returned torch tensor, and should have type torch.device. Default=None.
- Cross-device (CPU/GPU) taichi/PyTorch interaction support, when using to_torch/from_torch.
- #kernels compiled during external array IO significantly reduced (from matrix size to 1)
(Dec 23, 2019) v0.3.8 released.
- Breaking change: ti.data_oriented decorator introduced. Please decorate all your Taichi data-oriented objects using this decorator. To invoke the gradient versions of classmethod, for example, A.forward, simply use A.forward.grad() instead of A.forward(__gradient=True) (obsolete).
(Dec 22, 2019) v0.3.5 released.
- Maximum tensor dimensionality is 8 now (used to be 4). I.e., you can now allocate up to 8-D tensors.
(Dec 22, 2019) v0.3.4 released.
- 2D and 3D polar decomposition (R, S = ti.polar_decompose(A, ti.f32)) and svd (U, sigma, V = ti.svd(A, ti.f32)) support. Note that sigma is a 3x3 diagonal matrix.
- Fixed documentation versioning
- Allow expr_init with ti.core.DataType as inputs, so that ti.core.DataType can be used as ti.func parameter
(Dec 20, 2019) v0.3.3 released.
- Loud failure message when calling nested kernels. Closed #310
- DiffTaichi examples moved to a standalone repo
- Fixed documentation versioning
- Correctly differentiating kernels with multiple offloaded statements
(Dec 18, 2019) v0.3.2 released
- Vector.norm now comes with a parameter eps (=0 by default), and returns sqrt(\sum_i(x_i ^ 2) + eps). A non-zero eps safe guards the operator's gradient on zero vectors during differentiable programming.
(Dec 17, 2019) v0.3.1 released.
- Removed dependency on glibc 2.27
(Dec 17, 2019) v0.3.0 released.
- Documentation significantly improved
- break statements supported in while loops
- CPU multithreading enabled by default
(Dec 16, 2019) v0.2.6 released.
- ti.GUI.set_image(np.ndarray/Taichi tensor)
- Inplace adds are atomic by default. E.g., x[i] += j is equivalent to ti.atomic_add(x[i], j)
- ti.func arguments are forced to pass by value
- min/max can now take more than two arguments, e.g. max(a, b, c, d)
- Matrix operators transposed, trace, polar_decompose, determinant promoted to ti scope. I.e., users can now use ti.transposed(M) instead of ti.Matrix.transposed(M)
- ti.get_runtime().set_verbose(False) to eliminate verbose outputs
- LLVM backend now supports multithreading on CPUs
- LLVM backend now supports random number generators (ti.random(ti.i32/i64/f32/f64)
(Dec 5, 2019) v0.2.3 released.
- Simplified interaction between Taichi, numpy and PyTorch
  - taichi_scalar_tensor.to_numpy()/from_numpy(numpy_array)
  - taichi_scalar_tensor.to_torch()/from_torch(torch_array)
(Dec 4, 2019) v0.2.2 released.
- Argument type ti.ext_arr() now takes PyTorch tensors
(Dec 3, 2019) v0.2.1 released.
- Improved type mismatch error message
- native min/max supprt
- Tensor access index dimensionality checking
- Matrix.to_numpy, Matrix.zero, Matrix.identity, Matrix.fill
- Warning instead of error on lossy stores
- Added some initial support for cross-referencing local variables in different offloaded blocks.
(Nov 28, 2019) v0.2.0 released.
- More friendly syntax error when passing non-compile-time-constant values to ti.static
- Systematically resolved the variable name resolution issue
- Better interaction with numpy:
  - numpy arrays passed as a ti.ext_arr() [examples]
    - i32/f32/i64/f64 data type support for numpy
    - Multidimensional numpy arrays now supported in Taichi kernels
  - Tensor.to_numpy() and Tensor.from_numpy(numpy.ndarray) supported [examples]
  - Corresponding PyTorch tensor interaction will be supported very soon. Now only 1D f32 PyTorch tensors supproted when using ti.ext_arr(). Please use numpy arrays as intermediate buffers for now
- Indexing arrays with an incorrect number of indices now results in a syntax error
- Tensor shape reflection: [examples]
  - Tensor.dim() to retrieve the dimensionality of a global tensor
  - Tensor.shape() to retrieve the shape of a global tensor
  - Note the above queries will cause data structures to be materialized
- struct-for (e.g. for i, j in x) now supports iterating over tensors with non power-of-two dimensions
- Handy tensor filling: [examples]
  - Tensor.fill(x) to set all entries to x
  - Matrix.fill(x) to set all entries to x, where x can be a scalar or ti.Matrix of the same size
- Reduced python package size
- struct-for with grouped indices for better metaprogramming, especially in writing dimensionality-independent code, in e.g. physical simulation: [examples]

for I in ti.grouped(x): # I is a vector of size x.dim() and data type i32
  x[I] = 0
  
# If tensor x is 2D 
for I in ti.grouped(x): # I is a vector of size x.dim() and data type i32
  y[I + ti.Vector([0, 1])] = I[0] + I[1]
# is equivalent to
for i, j in x:
  y[i, j + 1] = i + j

(Nov 27, 2019) v0.1.5 released.
- Better modular programming support
- Disalow the use of ti.static outside Taichi kernels
- Documentation improvements (WIP)
- Codegen bug fixes
- Special thanks to Andrew Spielberg and KLozes for bug report and feedback.
(Nov 22, 2019) v0.1.3 released.
- Object-oriented programming. [Example]
- native Python function translation in Taichi kernels:
  - Use print instead of ti.print
  - Use int() instead of ti.cast(x, ti.i32) (or ti.cast(x, ti.i64) if your default integer precision is 64 bit)
  - Use float() instead of ti.cast(x, ti.f32) (or ti.cast(x, ti.f64) if your default float-point precision is 64 bit)
  - Use abs instead of ti.abs
  - Use ti.static_print for compile-time printing
(Nov 16, 2019) v0.1.0 released. Fixed PyTorch interface.
(Nov 12, 2019) v0.0.87 released.
- Added experimental Windows support with a [known issue] regarding virtual memory allocation, which will potentially limit the scalability of Taichi programs (If you are a Windows expert, please let me know how to solve this. Thanks!). Most examples work on Windows now.
- CUDA march autodetection;
- Complex kernel to override autodiff.
(Nov 4, 2019) v0.0.85 released.
- ti.stop_grad for stopping gradients during backpropagation. [Example];
- Compatibility improvements on Linux and OS X;
- Minor bug fixes.
(Nov 1, 2019) v0.0.77 released.
- Python wheels now support OS X 10.14+;
- LLVM is now the default backend. No need to install gcc-7 or clang-7 anymore. To use legacy backends, export TI_LLVM=0;
- LLVM compilation speed is improved by 2x;
- More friendly syntax error messages.
(Oct 30, 2019) v0.0.72 released.
- LLVM GPU backend now as fast as the legacy (yet optimized) CUDA backend. To enable, export TI_LLVM=1;
- Bug fixes: LLVM struct for list generation.
(Oct 29, 2019) v0.0.71 released. LLVM GPU backend performance greatly improved. Frontend compiler now emits readable syntax error messages.
(Oct 28, 2019) v0.0.70 released. This version comes with experimental LLVM backends for x86_64 and CUDA (via NVVM/PTX). GPU kernel compilation speed is improved by 10x. To enable, update the taichi package and export TI_LLVM=1.
(Oct 24, 2019) Python wheels (v0.0.61) released for Python 3.6/3.7 and CUDA 10.0/10.1 on Ubuntu 16.04+. Contributors of this release include Yuanming Hu, robbertvc, Zhoutong Zhang, Tao Du, Srinivas Kaza, and Kenneth Lozes.
(Oct 22, 2019) Added support for kernel templates. Kernel templates allow users to pass in taichi tensors and compile-time constants as kernel parameters.
(Oct 9, 2019) Compatibility improvements. Added a basic PyTorch interface. [Example].

Notes:

You still need to clone this repo for demo scripts under examples. You do not need to execute install.py or dev_setup.py. After installation using pip you can simply go to examples and execute, e.g., python3 mpm_fluid.py.
Make sure you clear your legacy Taichi installation (if applicable) by cleaning the environment variables (delete TAICHI_REPO_DIR, and remove legacy taichi from PYTHONPATH) in your .bashrc or .zshrc. Or you can simply do this in your shell to temporarily clear them:

export PYTHONPATH=
export TAICHI_REPO_DIR=

The Taichi Library [Legacy branch]

Taichi is an open-source computer graphics library that aims to provide easy-to-use infrastructures for computer graphics R&D. It's written in C++14 and wrapped friendly with Python.

News

May 17, 2019: Giga-Voxel SPGrid Topology Optimization Solver is released!
March 4, 2019: MLS-MPM/CPIC solver is now MIT-licensed!
August 14, 2018: MLS-MPM/CPIC solver reloaded! It delivers 4-14x performance boost over the previous state of the art on CPUs.

Moolight/taichi

Docs | Tutorial | DiffTaichi | Examples | FAQ | Forum

Related papers

Short-term goals

Updates

The Taichi Library [Legacy branch]

News

Getting Started (Legacy)