CuPy v7.0.0b4 を リリースしました

[Hiroyuki Yamazaki]

CuPy v7.0.0b4 をリリースしました！ リリースノートは以下の通りです。

This is the release note of v7.0.0b4. See here for the complete list of solved issues and merged PRs.

Note We're currently preparing wheels for CUDA 8.0 on Windows which will be published in a few days. Other wheels have already been published.

New Features

• Add cupy.cross (#2366, thanks @UmashankarTriforce!)
• Add cupy.may_share_memory (#2417)
• Add occupancy driver APIs (#2424, thanks @leofang!)
• Add scatter_max and scatter_min (#2427)
• Support texture memory in RawKernel (#2432, thanks @leofang!)

Enhancements

• Extend CuDNNError to have more debugging information (#2404)
• Add complex dtype support to cupy.std and cupy.var (#2411, thanks @grlee77!)
• Support complex dtypes in cupy.linalg.inv() (#2468, thanks @leofang!)

Performance Improvements

• Use CUB to speed up sum/min/max (#2090)
• Avoid creating empty numpy ndarray in common_type (#2307)
• cupy.linalg.norm: update docstring and improve performance for ord=2 cases (#2479, thanks @grlee77!)

Bug Fixes

• Fix bug in coosort (#2410, thanks @econtal!)
• Fix multithreading issue in cupy.cuda.cufft.get_current_plan() (#2435, thanks @leofang!)
• Normalize hidden layer strides in cuDNN RNN (#2442)
• Copy inputs in ufunc if they share the same memory with outputs (#2460)

Code Fixes

• Fix cuSOLVER devInfo dtype in inv and unify how those are specified (#2454)
• Fix code convention of statistics routines (#2459)
• Remove unnecessary ndarray private methods (#2465)

Documentation

• Document CUTENSOR_PATH environment variable (#2386)
• Add appropriate pointer types in Cython to the contribute guide (#2455, thanks @leofang!)
• Fix invalid escape sequence (#2470)

Installation

• Fix NumPy version in Dockerfile (#2430)

Examples

• Broadcast ValueError for n-clusters > 2 in k-means example (#2453, thanks @casheera!)

Tests

• Simplify cupy.fuse tests (#2339)
• Revert "Skip some ndarray-elementwise-op tests as temporary fix" (#2383)
• Drop Python 2 Travis CI configuration (#2428)
• Drop Python 2 PFN CI configuration (#2429)
• Change URL to place test assets (#2434)
• Add NumPy 1.17.1 to skip list of 0-length ifft test (#2441)
• Remove unnecessary skip in 0-length ifft test (#2443)
• Fix TestDLTensorMemory.test_delete (#2451)
• Fix data race in advanced indexing test (#2472)
• Fix pytest 5.x version errors (#2473)
• Remove global state from tests (#2475, thanks @leofang!)