Repository navigation
Bump cuquantum to 24.03 and cutensor to 2.0 - #1524
Conversation
https://docs.nvidia.com/cuda/cuquantum/latest/cutensornet/release_notes.html#cutensornet-v2-4-0 - cutensornetStateApplyTensor -> cutensornetStateApplyTensorOperator (introducing the latter, deprecating the former) - Use the new cutensornetStateApplyControlledTensorOperator: multi-controlled single-target gates + Add special handling for cSWAP, which is the only two-target gate defined in NVQIR runtime. + Remove the need for manual decomposition for MPS. - Update version info as necessary. Tested by: running tensornet and tensornet-mps unit tests.
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
No, it's intended for 0.8 and should only be merged after #1493 and 0.7.1 release. |
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
schweitzpgi
left a comment
There was a problem hiding this comment.
Changes look good, but still some residual failures.
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
Description
Some updates according to the release notes:
cutensornetStateApplyTensor->cutensornetStateApplyTensorOperator(introducing the latter, deprecating the former)
Use the new
cutensornetStateApplyControlledTensorOperator, which supports multi-controlled single-target gatesAdd special handling for cSWAP, which is the only two-target gate defined in NVQIR runtime.
Remove the need for manual decomposition for MPS.
Update version info in Dockerfile/build script/etc. as necessary.
Tested by: running
tensornetandtensornet-mpsunit tests.