Skip to content

Latest commit

 

History

History
260 lines (197 loc) · 7.49 KB

CHANGELOG.md

File metadata and controls

260 lines (197 loc) · 7.49 KB
Raw

CHANGELOGS

Instrument:

  • Add [Feature] prefix for new features
  • Add [Bugfix] prefix for bug fixes
  • Add [API] prefix for API changes

staging

please add your unreleased change here.

  • [Feature] Add minimax approximation for log
  • [Feature] Support jax.lax.top_k
  • [Improvement] Default log approximation to minmax
  • [Improvement] Improve median performance

20240306

  • [Feature] Support more generic Torch model inference
  • [Improvement] Optimize one-time setup for yacl ot
  • [Improvement] Optimize sort performance
  • [Infra] Python 3.8 is no longer suppoted
  • [Infra] Bump jax minimum version to 0.4.16

20240105

  • [Feature] Add Odd-Even Merge Sort to replace the bitonic sort
  • [Feature] Add radix sort support for ABY3
  • [Feature] Integrate with secretflow/psi
  • [Feature] Add Linux aarch64 support
  • [Feature] Add equal support for SEMI2K and ABY3
  • [Improvement] Optimize sort memory usage
  • [Improvement] Improve compatibility with latest Jax
  • [Bugfix] Fix compilation cache collision under certain cases
  • [Deprecated] macOS 11.x is no longer supported

20231108

  • [Bugfix] Fix compatibility with latest Jax
  • [Feature] Improve memory efficiency during encode/decode data
  • [Feature] Add radix sort support for SEMI2K
  • [Feature] Experimental: ABY3 matmul CUDA support
  • [Feature] Experimental: Private support under colocated mode
  • [Feature] Add yacl ot support for Cheetah

20230906

  • [SPU] 0.5.0 release
  • [API] Update IO interface to support splitting large data into chunks
  • [API] Add intrinsic function support
  • [Feature] Add half type support
  • [Feature] Add PSI progress report
  • [Feature] Add SineOp/CosineOp support
  • [Feature] Add complex support

20230705

  • [SPU] 0.4.1 release
  • [Improvement] Improve tanh performance

20230614

  • [SPU] 0.4.0 release
  • [Feature] Improve secret dynamic slicing performance
  • [Feature] Improve accuracy of tanh approximation
  • [Example] Add flax GPT2
  • [Example] Add ResNet
  • [Bugfix] Fix gen cache in unbalanced psi
  • [Bugfix] Fix possible compilation cache collision
  • [Improvement] Improve accuracy of mean with large size input
  • [Improvement] Improve support of Jax 64bit mode
  • [API] Add compiler options

20230406

  • [SPU] 0.3.2 release
  • [Feature] Add TrustedThirdParty beaver provider for semi2k
  • [Feature] Expose ssl/tls options
  • [Feature] Add EpsilonOp
  • [Feature] Support CaseOp
  • [Feature] Improve sort performance
  • [Feature] Improve shift performance
  • [Feature] Support shift by secret number of bits
  • [Feature] Support secret indexing
  • [Feature] Add PIR python binding
  • [bugfix] Fix boolean ConstantOp
  • [bugfix] Fix jnp.median
  • [bugfix] Fix jnp.sort on floating point inputs
  • [bugfix] Fix secret sort with public payloads
  • [3p] Relax TensorFlow version in requirements.txt
  • [3p] Move to OpenXLA
  • [API] Move C++ API from spu to libspu
  • [PSI] Add ecdh-oprf 2-party two stage mode.

20230104

  • [SPU] 0.3.1 release
  • [API] Add get_var_meta
  • [API] Change ValueProto to bytes in Runtime/IO binding API
  • [Feature] Add SPU logging
  • [Feature] ECDH-PSI supports white box interconnection mode
  • [Feature] Various performance improvements
  • [Feature] Lift IfOp condition limit
  • [bugfix] Fix various crashes
  • [3p] Build with Tensorflow 2.11.0
  • [3p] Update JAX to 0.4.1

20221116

  • [SPU] 0.3.0 release
  • [API] Add IR type for frontend compilation
  • [Feature] Support SignOp, Expm1Op, DotGeneralOp
  • [Feature] Improve convolution performance
  • [Feature] Link Send/Recv python binding
  • [bugfix] Relax iota type constraint
  • [bugfix] Fix type inference for whileOp/dynamicUpdateSliceOp
  • [bugfix] Relax clamp type constraint
  • [bugfix] Fix NotOp with non-pred integer types

20220905

  • [SPU] 0.2.4 release
  • [bugfix] Fix Tensorflow example
  • [miscs] Performance improvements

20220822

  • [SPU] 0.2.1 release
  • [Feature] Add compiler cache
  • [Feature] Support bc22-pcg psi
  • [API] Refactor psi api
  • [miscs] Various performance improvements

20220727

  • [SPU] 0.1.1 release
  • [API] Added rsqrt support
  • [bugfix] Fixed some crashes and wrong answers

20220527

  • [SPU] 0.1.0 release
  • [API] Add exp/log approximation configurations.
  • [API] Change SigmoidMode enum value.

20220513

  • [SPU] 0.0.10
  • [Infra] Set minimum macOS requirements to 11.0
  • [Feature] Performance improvements
  • [Feature] Support SelectAndScatter.

20220511

  • [SPU] 0.0.9 release
  • [Feature] Add spu.binding.distributed.
  • [API] Expose visibility to ValueProto.
  • [bugfix] Fixed several vtype and dtype mismatches.
  • [bugfix] Improve multiply accuracy under certain scenarios.
  • [bugfix] Ensure dynamic-slice & dynamic-update-slice OOB behavior is aligned with XLA semantics.

20220427

  • [SPU] 0.0.8.1 release
  • [bugfix] Fix iota type mismatch.

20220425

  • [SPU] 0.0.8 release
  • [Feature] Add tanh support.
  • [bugfix] Fix secret shift wrong answer.

20220424

  • [SPU] 0.0.7.1 release
  • [bugfix] Fix secret shift.

20220422

  • [SPU] 0.0.7 release
  • [Feature] Add python simulation module.
  • [Feature] Support general sort comparator.
  • [Feature] Support remainder operation.
  • [Feature] Support arithmetic right shift operation.
  • [API] Change DataType definition, support more accurate encoding.

20220412.1

  • [SPU] 0.0.6.1 release
  • [bugfix] Fixed several macOS build issues.

20220412

  • [SPU] 0.0.6 release
  • [API] Add storage_type/data_type, remove type_data.
  • [Experimental] Log(natural logarithm) is calculated with Pade approximation instead of Householder's approximation.
  • [API] Fix typo, goldschmdit.
  • [Improvement] fxp div, improvement accuracy with normal distributed inputs.

20220325

  • [SPU] 0.0.5.1 release
  • [Bugfix] Fix SEMI2K division wrong answer

20220324

  • [SPU] 0.0.5 release
  • [Improvement] Lift Convolution limitations.
  • [Improvement] Optimize maximum/minimum speed, with MPC B2A single bit conversion.
  • [Experimental] Div is calculated with direct goldschmdit method instead of multiplication of dividend and divisor reciprocal.
  • [API] change RuntimeConfig.enable_op_time_profile to enable_pphlo_profile.
  • [API] change RuntimeConfig.fxp_reciprocal_goldschmdit_iters to fxp_div_goldschmdit_iters.

20220308

  • [SPU] 0.0.4 release
  • [Feature] add silent ot support for various ot scenarios (chosen/correlated/random messages, chosen/correlated/random choices, 1o2/1oN)
  • [Feature] add non-linear computation protocols based on silent ot (comparison, truncation, b2a, triple, randbit, etc)
  • [Feature] add a 2PC protocol: Cheetah
  • [Improvement] concatenate is a lot faster
  • [API] add RuntimeConfig.enable_op_time_profile to ask SPU to collect timing profiling data
  • [Bugfix] fixed pphlo.ConstOp may lose signed bit after encoding

20220303

  • [spu] 0.0.3 release
  • [API] merge (config.proto, executable.proto, types.proto) into single spu.proto.
  • [API] change RuntimeConfig.enable_protocol_trace to enable_action_trace.
  • [API] change RuntimeConfig.fxp_recirptocal_goldschmdit_iters to fxp_reciprocal_goldschmdit_iters.
  • [API] add RuntimeConfig.reveal_secret_condition to allow reveal secret control flow condition.
  • [Bugfix] Fixed SEGV when reconstruct from an ABY3 scalar secret
  • [Feature] Left/right shift now properly supports non-scalar inputs

20220217

  • [spu] 0.0.2.3 release

20220211

  • [spu] 0.0.2.2 release
  • [Bugfix] Fix exception when kkrt psi input is too small

20220210

  • [spu] 0.0.2.1 release
  • [Bugfix] Fix exception when streaming psi output directory already exists

20220209

  • [spu] 0.0.2 release.
  • [Feature] Support multi-parties psi

20210930

  • [spu] Init release.