coaster 0.2.0

high-performance computation on any hardware
Documentation
<a name="0.0.8"></a>
## 0.0.8 (2016-02-21)

#### Features

* **backend:**
  *  implement Backend::default() ([0d4f0cc2]https://github.com/autumnai/collenchyma/commit/0d4f0cc2061ba09bf17ba90a6ae5845150df0bb5)
  *  add ability to synchronize a backend ([e1bf0147]https://github.com/autumnai/collenchyma/commit/e1bf01476cf159d627de33962a3d02d92dbf25c9)
* **tensor:**  add `resize` to SharedTensor ([f3bb3b44]https://github.com/autumnai/collenchyma/commit/f3bb3b44992ca721adb1c59369d3729522592b34)
* **feature_flag:**  add alternative unstable method to allocate native memory (usable via `unstable_alloc` feature flag) ([ebaf3dac]https://github.com/autumnai/collenchyma/commit/ebaf3dace462303c8c88f61a2ab388cffb6a9af5)
* **export:**  add prelude module (import most of collenchyma via `use collenchyma::prelude::*;`) ([4aea78df]https://github.com/autumnai/collenchyma/commit/4aea78df867761c5963858856fd74ea7b5ea4661, [60c1a395]https://github.com/autumnai/collenchyma/commit/60c1a395e500b57cf66a5cb6bb02f912a14484de)
* **license:**  change license to dual MIT/Apache-2.0 ([af5c08a0]https://github.com/autumnai/collenchyma/commit/af5c08a0fa2b7eb4bf5d84592d83f1b447161a69)

#### Bug Fixes

* **cuda_context:**  fix repeating CUDA context creation ([0bf616f8]https://github.com/autumnai/collenchyma/commit/0bf616f8538de917b4ec2e2fdae154ee37bc26e0)


<a name="0.0.7"></a>
## 0.0.7 (2015-12-21)


#### Features

* **backend_device:**  add device() to IBackend ([f797a72a]https://github.com/autumnai/collenchyma/commit/f797a72a39a530f2d29f7dc25c0b1c11ec7cda87)
* **tensor:**  add reshape to SharedTensor ([9c8721ed]https://github.com/autumnai/collenchyma/commit/9c8721edd5acc66caa955253d9fdb403600f9b85)

#### Bug Fixes

* **lib:**  various fixes concerning impl of plugins ([bec27ca1]https://github.com/autumnai/collenchyma/commit/bec27ca1de82b1be21ca2295a65d56311533ccba)



<a name="0.0.6"></a>
## 0.0.6 (2015-12-16)


#### Features

* **tensor:**  rename SharedMemory -> Tensor ([acc3cbb7]https://github.com/autumnai/collenchyma/commit/acc3cbb7f0f850cf6414131f33c013f71a53a852)

#### Bug Fixes

* **travis:**  fix feature build flags ([f9861a73]https://github.com/autumnai/collenchyma/commit/f9861a731360f3a3dc7ad6f81a69be6cc05fe622)



<a name="0.0.5"></a>
## 0.0.5 (2015-12-09)


#### Bug Fixes

* **use-size-types:**  use target dependent size types ([4e4a5cd3]https://github.com/autumnai/collenchyma/commit/4e4a5cd3f1716c122a22bb0d008b06bb61f74bce)

#### Features

* **bench:**  add benchmarks for memory synchronization ([762b87ca]https://github.com/autumnai/collenchyma/commit/762b87ca2c1bf55e257803607d7a8463c07e77e3)
* **features:**  add native and opencl feature ([3609fea1]https://github.com/autumnai/collenchyma/commit/3609fea17d1cd196196d7dba3140ede53e681d41)



<a name="0.0.4"></a>
## 0.0.4 (2015-12-08)


#### Features

* **cuda:**
  *  add memory synchronization ([760e4c45]https://github.com/autumnai/collenchyma/commit/760e4c45f97312729770e51980ad1481906d33b4)
  *  add cuda device support ([62eef6d8]https://github.com/autumnai/collenchyma/commit/62eef6d8cfb8a60f6fac892b50301738adaaa65c)
  *  add memory allocation ([35f7f479]https://github.com/autumnai/collenchyma/commit/35f7f47916cebcea626e43313c3f42ebce2e4e21)
  *  add basic context and memory support ([b0a40d38]https://github.com/autumnai/collenchyma/commit/b0a40d38f46a3ef313495a0d4d0847db0821bb64)
* **cudnn:**  add cudnn ffi ([0bbbff83]https://github.com/autumnai/collenchyma/commit/0bbbff832632c868589f8905e6e9e70d003161c2)
* **library:**  remove last pieces of library ([38dcd6a6]https://github.com/autumnai/collenchyma/commit/38dcd6a68914ab070d76a89f19307ff58beaccf7)
* **opencl:**  implement shared_memory for OpenCL ([be47d6ba]https://github.com/autumnai/collenchyma/commit/be47d6ba6a5de216726bd4adc0dd2f99fbe7c31b)
* **perf:**  make error messages static strings ([430c4ed6]https://github.com/autumnai/collenchyma/commit/430c4ed657242dce21ebafad7084e94d9755fae3)
* **plugin:**  move library out; replace with thin plugin mod ([3bbebe9a]https://github.com/autumnai/collenchyma/commit/3bbebe9a7f95ce24a936b3824c5ce410c79e7214)
* **shared_memory:**  add dimensionality to shared_memory ([13cd0905]https://github.com/autumnai/collenchyma/commit/13cd090596358e523c70e98dcf32885b2b9271bd)

#### Performance

* **blas:**  reduce overhead of calling blas ([8b7a7aee]https://github.com/autumnai/collenchyma/commit/8b7a7aeeaf67482031da0fd712328f747be09e72)
* **shared_memory:**  use linear_map for SharedMemory.copies ([44ea377d]https://github.com/autumnai/collenchyma/commit/44ea377d08da066159a01d646fed65f5f7080f8f)

#### Bug Fixes

* **compilation:**  make cuda optional ([1f933977]https://github.com/autumnai/collenchyma/commit/1f9339771d1eec9a6c42bcbed1ff784f97220896)
* **windows:**
  *  use `size_t` instead of `u64` where necessary ([6e9fdfbb]https://github.com/autumnai/collenchyma/commit/6e9fdfbb73155927a91b7b31dff2b208c3e49624)
  *  add the link attribute to OpenCL external block ([2017a10f]https://github.com/autumnai/collenchyma/commit/2017a10fcf8597b83f9f4ae11a6396927406c81d)



<a name="0.0.3"></a>
## 0.0.3 (2015-11-30)


#### Features

* **blas:**
  *  add blas native level 1 support ([38273645]https://github.com/autumnai/collenchyma/commit/3827364549dfa5b79ef2bbbb0bd38f0096e267cc)
  *  add basic level 1 blas methods for native ([62cbc4c4]https://github.com/autumnai/collenchyma/commit/62cbc4c42757a7a489a358c2bd5e16bdb47938cd)
* **computation:**  add basic design for backend-agnostic computation ([a3f9534f]https://github.com/autumnai/collenchyma/commit/a3f9534f9483531be4eecd91610d2e72ae84cd07)
* **cuda:**
  *  add cuda structure ([d42430c1]https://github.com/autumnai/collenchyma/commit/d42430c10248f39abb665a7bb22d0b3e32e2f08d)
* **dot:**  add working dot computation ([6572c010]https://github.com/autumnai/collenchyma/commit/6572c01036a4ac07906852f5bfaee3d7709d8f8c)
* **extern:**  add backend traits for extern support ([f3d50172]https://github.com/autumnai/collenchyma/commit/f3d5017228ebc4195d593d0210e88c4ca431eaa4)
* **flatbox:**  provide slice accessors and give more allocation responsibility to shared_memory ([cfbb5b13]https://github.com/autumnai/collenchyma/commit/cfbb5b135691f58671f7e391ec16793921073198)
* **memory:**
  *  implement MemoryType unwrappers ([fbd26776]https://github.com/autumnai/collenchyma/commit/fbd26776685aa6e4d90253628099104df1759871)
  *  add SharedMemory.latest_device() ([d425fc6c]https://github.com/autumnai/collenchyma/commit/d425fc6c29b177f778d8f3ed1c39fe5328191537)

#### Bug Fixes

* **ci:**  change clippy to be optional ([753dfb02]https://github.com/autumnai/collenchyma/commit/753dfb0203d2ff9e063e04bf14d93b40fbc506b5)


<a name="0.0.2"></a>
## 0.0.2 (2015-11-27)


#### Features

* **collenchyma:**  outline design ([876ac4b0]https://github.com/autumnai/collenchyma/commit/876ac4b037cda06975f1c88dfb8d86bc62d7fe47)
* **computation:**  add basic design for backend-agnostic computation ([e43f947a]https://github.com/autumnai/collenchyma/commit/e43f947adc12f02b5d5f84a9a4d4cce0d902580c)
* **dot:**  add working dot computation ([f5c8fdaa]https://github.com/autumnai/collenchyma/commit/f5c8fdaa11a43ef5b3244e2480afabe2f7374248)
* **flatbox:**  provide slice accessors and give more allocation responsibility to shared_memory ([a31dd493]https://github.com/autumnai/collenchyma/commit/a31dd4936c7bc37b8184bd8a8194e153b6826ec7)
* **lib:**  add foundation of the library design ([f5fd0235]https://github.com/autumnai/collenchyma/commit/f5fd02352e07094a457dff3b058a6842b39f798f)
* **memory:**
  *  implement MemoryType unwrappers ([a78edebb]https://github.com/autumnai/collenchyma/commit/a78edebb0036c82617847a58e5c4cb8cc5995d5b)
  *  add SharedMemory.latest_device() ([f9a7465d]https://github.com/autumnai/collenchyma/commit/f9a7465d54808672a5c39cb4cf89d8a0253e8aec)

#### Bug Fixes

* **ci:**  change clippy to be optional ([db13da29]https://github.com/autumnai/collenchyma/commit/db13da299ee635a4201ff66015a6df2e65314e73)