v2.1.1

v2.1.1 (2019-12-19)

Diff since v2.1.0

v2.1.0

v2.1.0 (2019-12-10)

Diff since v2.0.1

v2.0.1

v2.0.1 (2019-12-06)

Diff since v2.0.0

v2.0.0

v2.0.0 (2019-11-20)

Diff since v1.3.1

Closed issues:

• Multi-dimensional array contraction code not working (#73)

Merged pull requests:

• cuarray support via cutensor (#76) (Jutho)
• ncon docs (#75) (Jutho)
• add ncon function and macro, add optional contraction order (#72) (Jutho)

v1.3.1

v1.3.1 (2019-10-22)

Diff since v1.3.0

v1.3.0

v1.3.0 (2019-10-08)

Diff since v1.2.0

Closed issues:

• gpu support (#61)

Merged pull requests:

• completely rewritten parser (#71) (Jutho)

v1.2.0

• switch to thread safe LRU cache for temporaries
• use fixed size default for maximal memory of cache
• refactor part of the index notation processing to make it easier to inspect and more robust and efficient for complicated tensor objects
• introduce an overloadable memsize function to measure the size of the temporaries in the cache

v1.1.0

• Tensoropt support for large number of tensors, thanks @GiggleLiu
• Switched to new registration system

v1.0.4

simplify deindexify

v1.0.3

Support for blocks inside @tensor, small changes in precompilation/compilation time ?