Merge from taichi-dev

CMakeLists.txt

@@ -8,7 +8,7 @@ project(taichi)
 
 SET(TI_VERSION_MAJOR 0)
 SET(TI_VERSION_MINOR 5)
-SET(TI_VERSION_PATCH 0)
+SET(TI_VERSION_PATCH 1)
 
 execute_process(
   WORKING_DIRECTORY ${CMAKE_SOURCE_DIR} 

README.md

@@ -20,23 +20,13 @@ python3 -m pip install taichi-nightly-cuda-10-0
 python3 -m pip install taichi-nightly-cuda-10-1
 ```
 
+## [Contribution Guidelines](https://taichi.readthedocs.io/en/latest/contributor_guide.html)
+
 || **Linux (CUDA)** | **OS X (10.14+)** | **Windows** |
 |:------|:-----|:-----|:-----|
 |**Build**|[![Build Status](http://f11.csail.mit.edu:8080/job/taichi/badge/icon)](http://f11.csail.mit.edu:8080/job/taichi/)| [![Build Status](https://travis-ci.com/taichi-dev/taichi.svg?branch=master)](https://travis-ci.com/taichi-dev/taichi) | [![Build status](https://ci.appveyor.com/api/projects/status/yxm0uniin8xty4j7/branch/master?svg=true)](https://ci.appveyor.com/project/yuanming-hu/taichi/branch/master)|
 |**PyPI**|[![Build Status](https://travis-ci.com/yuanming-hu/taichi-wheels-test.svg?branch=master)](https://travis-ci.com/yuanming-hu/taichi-wheels-test)|[![Build Status](https://travis-ci.com/yuanming-hu/taichi-wheels-test.svg?branch=master)](https://travis-ci.com/yuanming-hu/taichi-wheels-test)|[![Build status](https://ci.appveyor.com/api/projects/status/39ar9wa8yd49je7o?svg=true)](https://ci.appveyor.com/project/IteratorAdvance/taichi-wheels-test)|
 
-## [Contribution Guidelines](https://taichi.readthedocs.io/en/latest/contributor_guide.html)
-
-## Related papers
-- [**(SIGGRAPH Asia 2019) High-Performance Computation on Sparse Data Structures**](http://taichi.graphics/wp-content/uploads/2019/09/taichi_lang.pdf) [[Video]](https://youtu.be/wKw8LMF3Djo) [[BibTex]](https://raw.githubusercontent.com/yuanming-hu/taichi/master/misc/taichi_bibtex.txt)
-  - by *Yuanming Hu, Tzu-Mao Li, Luke Anderson, Jonathan Ragan-Kelley, and Frédo Durand*
-- [**(ICLR 2020) Differentiable Programming for Physical Simulation**](https://arxiv.org/abs/1910.00935) [[Video]](https://www.youtube.com/watch?v=Z1xvAZve9aE) [[BibTex]](https://raw.githubusercontent.com/yuanming-hu/taichi/master/misc/difftaichi_bibtex.txt) [[Code]](https://github.com/yuanming-hu/difftaichi)
-  - by *Yuanming Hu, Luke Anderson, Tzu-Mao Li, Qi Sun, Nathan Carr, Jonathan Ragan-Kelley, and Frédo Durand*
-
-<div align="center">
-  <img width="800px" src="https://github.com/taichi-dev/taichi/blob/master/docs/life_of_kernel_lowres.jpg">
-</div>        
-
 ## 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:
@@ -46,6 +36,14 @@ python3 -m pip install taichi-nightly-cuda-10-1
 - (WIP) Tune the performance of the LLVM backend to match that of the legacy source-to-source backends (Hopefully by mid Feb, 2020. Current progress: setting up/tuning for final benchmarks)
 
 ## Updates
+- (Feb  16, 2020) v0.5.1 released
+   - Keyboard and mouse events supported in the GUI system. Check out [mpm128.py](https://github.com/taichi-dev/taichi/blob/master/examples/mpm128.py) for a interactive demo! (by **Yubin Peng [archibate] and Ye Kuang [k-ye]**)
+   - Basic algebraic simplification passes (by **Mingkuan Xu [xumingkuan]**)
+   - (For developers) `ti` (`ti.exe`) command supported on Windows after setting `%PATH%` correctly (by **Mingkuan Xu [xumingkuan]**)
+   - General power operator `x ** y` now supported in Taichi kernels (by **Yubin Peng [archibate]**)
+   - `.dense(...).pointer()` now abbreviated as `.pointer(...)`. `pointer` now stands for a dense pointer array. This leads to cleaner code and better performance. (by **Kenneth Lozes [KLozes]**)
+   - (Advanced struct-fors only) `for i in X` now iterates all child instances of `X` instead of `X` itself. Skip this if you only use `X=leaf node` such as `ti.f32/i32/Vector/Matrix`.
+   - Fixed cuda random number generator racing conditions
 - (Feb  14, 2020) **v0.5.0 released with a new Apple Metal GPU backend for Mac OS X users!** (by **Ye Kuang [k-ye]**)
    - Just initialize your program with `ti.init(..., arch=ti.metal)` and run Taichi on your Mac GPUs!
    - A few takeaways if you do want to use the Metal backend:
@@ -103,3 +101,14 @@ python3 -m pip install taichi-nightly-cuda-10-1
    - Doc updated
 
 - [Full changelog](changelog.md)
+
+
+## Related papers
+- [**(SIGGRAPH Asia 2019) High-Performance Computation on Sparse Data Structures**](http://taichi.graphics/wp-content/uploads/2019/09/taichi_lang.pdf) [[Video]](https://youtu.be/wKw8LMF3Djo) [[BibTex]](https://raw.githubusercontent.com/yuanming-hu/taichi/master/misc/taichi_bibtex.txt)
+  - by *Yuanming Hu, Tzu-Mao Li, Luke Anderson, Jonathan Ragan-Kelley, and Frédo Durand*
+- [**(ICLR 2020) Differentiable Programming for Physical Simulation**](https://arxiv.org/abs/1910.00935) [[Video]](https://www.youtube.com/watch?v=Z1xvAZve9aE) [[BibTex]](https://raw.githubusercontent.com/yuanming-hu/taichi/master/misc/difftaichi_bibtex.txt) [[Code]](https://github.com/yuanming-hu/difftaichi)
+  - by *Yuanming Hu, Luke Anderson, Tzu-Mao Li, Qi Sun, Nathan Carr, Jonathan Ragan-Kelley, and Frédo Durand*
+
+<div align="center">
+  <img width="800px" src="https://github.com/taichi-dev/taichi/blob/master/docs/life_of_kernel_lowres.jpg">
+</div>

appveyor.yml

@@ -33,14 +33,15 @@ skip_commits:
     - '.*'
 
 build_script:
+  - set TAICHI_REPO_DIR=C:\taichi
+  # - 'if %PYTHON% %TAICHI_REPO_DIR%\misc\appveyor_filter.py; then exit 0; fi'
   - cd C:\
   - curl --retry 10 --retry-delay 5 https://github.com/yuanming-hu/taichi_assets/releases/download/llvm8/taichi-llvm-8.0.1-msvc2017.zip -LO
   - 7z x taichi-llvm-8.0.1-msvc2017.zip -otaichi_llvm
   - curl --retry 10 --retry-delay 5 https://github.com/yuanming-hu/taichi_assets/releases/download/llvm8/clang-7.0.1-win.zip -LO
   - 7z x clang-7.0.1-win.zip -otaichi_clang
   - set PATH=C:\taichi_llvm\bin;%PATH%;
   - set PATH=C:\taichi_clang\bin;%PATH%
-  - set TAICHI_REPO_DIR=C:\taichi
   - set PYTHONPATH=%TAICHI_REPO_DIR%/python
   - set PATH=%TAICHI_REPO_DIR%\bin;%PATH%
   - clang --version

benchmarks/fill_sparse.py

@@ -6,7 +6,7 @@ def benchmark_nested_struct():
 
   @ti.layout
   def place():
-    ti.root.dense(ti.ij, [N, N]).pointer().dense(ti.ij, [8, 8]).place(a)
+    ti.root.pointer(ti.ij, [N, N]).dense(ti.ij, [8, 8]).place(a)
 
   @ti.kernel
   def fill():
@@ -24,7 +24,7 @@ def benchmark_nested_struct_fill_and_clear():
 
   @ti.layout
   def place():
-    ti.root.dense(ti.ij, [N, N]).pointer().dense(ti.ij, [8, 8]).place(a)
+    ti.root.pointer(ti.ij, [N, N]).dense(ti.ij, [8, 8]).place(a)
 
   @ti.kernel
   def fill():

ci_setup.py

@@ -47,7 +47,7 @@ def get_shell_rc_name():
 
 def get_username():
   if build_type == 'ci':
-    os.environ['TC_CI'] = '1'
+    os.environ['TI_CI'] = '1'
     username = 'travis'
   else:
     assert get_os_name() != 'win'

cmake/TaichiCXXFlags.cmake

@@ -1,6 +1,6 @@
 message("Using C++ compiler: " ${CMAKE_CXX_COMPILER})
 
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTC_ISE_NONE")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_ISE_NONE")
 
 option(BUILD_WITH_ADDRESS_SANITIZER "Build with clang address sanitizer" OFF)
 
@@ -49,19 +49,19 @@ if (USE_STDCPP)
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++")
 endif()
 
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTC_PASS_EXCEPTION_TO_PYTHON")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_PASS_EXCEPTION_TO_PYTHON")
 
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTC_INCLUDED")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_INCLUDED")
 
-if ($ENV{TC_USE_DOUBLE})
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTC_USE_DOUBLE")
+if ($ENV{TI_USE_DOUBLE})
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_USE_DOUBLE")
     message("Using float64 (double) precision as real")
 else()
     message("Using float32 (single) precision as real")
 endif()
 
-if (TC_USE_MPI)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTC_USE_MPI")
+if (TI_USE_MPI)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_USE_MPI")
     message("Using MPI")
 endif ()
 

cmake/TaichiCore.cmake

@@ -40,7 +40,7 @@ if (TI_WITH_CUDA)
         message("Building with CUDA ${CUDA_VERSION}")
         set(CUDA_ARCH 61)
         message("Found CUDA. Arch = ${CUDA_ARCH}")
-        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DCUDA_FOUND -DTI_WITH_CUDA")
+        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTI_WITH_CUDA")
         if (MSVC)
             include_directories(${CUDA_TOOLKIT_ROOT_DIR}/include)
             target_link_libraries(${LIBRARY_NAME} ${CUDA_TOOLKIT_ROOT_DIR}/lib/x64/cudart.lib ${CUDA_TOOLKIT_ROOT_DIR}/lib/x64/cuda.lib)
@@ -83,7 +83,7 @@ llvm_map_components_to_libnames(llvm_libs
         )
 target_link_libraries(${LIBRARY_NAME} ${llvm_libs})
 
-if (CUDA_FOUND)
+if (TI_WITH_CUDA)
     llvm_map_components_to_libnames(llvm_ptx_libs NVPTX)
     target_link_libraries(${LIBRARY_NAME} ${llvm_ptx_libs})
 endif()

docs/cpp_style.rst

@@ -5,7 +5,7 @@ Naming
 --------------------------------------------------------------------------
 - Variable names should consist of lowercase words connected by underscores, e.g. ``llvm_context``.
 - Class and struct names should consist of words with first letters capitalized, e.g. ``CodegenLLVM``.
-- Macros should be capital start with ``TC``, such as ``TC_INFO``, ``TC_IMPLEMENTATION``.
+- Macros should be capital start with ``TC``, such as ``TI_INFO``, ``TI_IMPLEMENTATION``.
 
    - We do not encourage the use of macro, although there are cases where macros are inevitable.
 

docs/utilities.rst

@@ -47,45 +47,45 @@ Serialization
 
 The serialization module of taichi allows you to serialize/deserialize objects into/from binary strings.
 
-You can use ``TC_IO`` macros to explicit define fields necessary in Taichi.
+You can use ``TI_IO`` macros to explicit define fields necessary in Taichi.
 
 .. code-block:: cpp
 
-    // TC_IO_DEF
+    // TI_IO_DEF
     struct Particle {
         Vector3f position, velocity;
         real mass;
         string name;
 
-        TC_IO_DEF(position, velocity, mass, name);
+        TI_IO_DEF(position, velocity, mass, name);
     }
 
-    // TC_IO_DECL
+    // TI_IO_DECL
     struct Particle {
         Vector3f position, velocity;
         real mass;
         bool has_name
         string name;
 
-        TC_IO_DECL() {
-            TC_IO(position);
-            TC_IO(velocity);
-            TC_IO(mass);
-            TC_IO(has_name);
+        TI_IO_DECL() {
+            TI_IO(position);
+            TI_IO(velocity);
+            TI_IO(mass);
+            TI_IO(has_name);
             // More flexibility:
             if (has_name) {
-                TC_IO(name);
+                TI_IO(name);
             }
         }
     }
 
-    // TC_IO_DEF_VIRT();
+    // TI_IO_DEF_VIRT();
 
 
 Progress Notification
 ----------------------------------
 
-The taichi messager can send an email to ``$TC_MONITOR_EMAIL`` when the task finished or crashed.
+The taichi messager can send an email to ``$TI_MONITOR_EMAIL`` when the task finished or crashed.
 To enable:
 
 .. code-block:: python

docs/version

@@ -1 +1 @@
-0.5.0
+0.5.1

examples/cpp/cnn.cpp

@@ -3,7 +3,7 @@
 #include <taichi/visual/gui.h>
 #include <taichi/system/profiler.h>
 
-TC_NAMESPACE_BEGIN
+TI_NAMESPACE_BEGIN
 
 using namespace Tlang;
 
@@ -12,10 +12,10 @@ constexpr int n = 256;
 constexpr int num_ch1 = 16, num_ch2 = 16;
 
 auto cnn = [](std::vector<std::string> cli_param) {
-  TC_WARN(
+  TI_WARN(
       "After refactoring the Texture class from Taichi is removed. Need to "
       "read from the raw bunny binary.");
-  TC_NOT_IMPLEMENTED
+  TI_NOT_IMPLEMENTED
 #if (0)
   CoreState::set_trigger_gdb_when_crash(true);
   auto param = parse_param(cli_param);
@@ -24,7 +24,7 @@ auto cnn = [](std::vector<std::string> cli_param) {
   auto cache_l1 = param.get("cache_l1", true);
   auto use_dense = param.get("use_dense", false);
   auto write_input_voxel = param.get("write_input", true);
-  TC_P(use_dense);
+  TI_P(use_dense);
 
   Program prog(gpu ? Arch::gpu : Arch::x86_64);
   prog.config.simplify_before_lower_access = opt;
@@ -228,6 +228,6 @@ auto cnn = [](std::vector<std::string> cli_param) {
 #endif
 #endif
 };
-TC_REGISTER_TASK(cnn);
+TI_REGISTER_TASK(cnn);
 
-TC_NAMESPACE_END
+TI_NAMESPACE_END

examples/cpp/diff_conv.cpp

@@ -1,23 +1,23 @@
 #include <taichi/util.h>
 
-TC_NAMESPACE_BEGIN
+TI_NAMESPACE_BEGIN
 
 auto diff_conv = [](const std::vector<std::string> args) {
   int grid_resolution = 254;
-  TC_ASSERT(args.size() == 3);
+  TI_ASSERT(args.size() == 3);
   float th = std::stof(args[2]);
-  TC_P(th);
+  TI_P(th);
   auto f = fopen(args[0].c_str(), "rb");
 
   int n = pow<3>(grid_resolution);
-  TC_ASSERT(f);
+  TI_ASSERT(f);
 
   std::vector<float32> ret1(n);
   trash(std::fread(ret1.data(), sizeof(float32), ret1.size(), f));
   std::fclose(f);
 
   f = fopen(args[1].c_str(), "rb");
-  TC_ASSERT(f);
+  TI_ASSERT(f);
   std::vector<float32> ret2(n);
   trash(std::fread(ret2.data(), sizeof(float32), ret2.size(), f));
   std::fclose(f);
@@ -53,19 +53,19 @@ auto diff_conv = [](const std::vector<std::string> args) {
     // fprintf(stderr, "ret1:%f, ret2:%f\n", ret1[i], ret2[i]);
     //}
   }
-  TC_INFO("same {} {}%", counter[0], 100.0f * counter[0] / n);
-  TC_INFO("non zero same {} {}%", counter[0],
+  TI_INFO("same {} {}%", counter[0], 100.0f * counter[0] / n);
+  TI_INFO("non zero same {} {}%", counter[0],
           100.0f * counter[1] / total_non_zero);
-  TC_P(sum1 / n);
-  TC_P(sum2 / n);
-  TC_P(sum1 / total_non_zero);
-  TC_P(sum2 / total_non_zero);
-  TC_P(max1);
-  TC_P(max2);
-  TC_P(non_zero1);
-  TC_P(non_zero2);
+  TI_P(sum1 / n);
+  TI_P(sum2 / n);
+  TI_P(sum1 / total_non_zero);
+  TI_P(sum2 / total_non_zero);
+  TI_P(max1);
+  TI_P(max2);
+  TI_P(non_zero1);
+  TI_P(non_zero2);
 };
 
-TC_REGISTER_TASK(diff_conv);
+TI_REGISTER_TASK(diff_conv);
 
-TC_NAMESPACE_END
+TI_NAMESPACE_END