v0.5.1

What's Changed

• docs: Finalized Airbender documentation by @Fitznik in #143
• feat: Switch verifier 6 to rolling hash by @EmilLuta in #149

Full Changelog: v0.5.0...v0.5.1

Dev release - 0.7.0

Dev release

Used for eth proofs, with keccak precompile and 4MB ROM size.

v0.5.0

What's Changed

• feat: adapt cli run to parse sequencer json input format by @mm-zk in #125
• feat(gpu_prover): setup trees offloading to and querying from host memory by @robik75 in #127
• feat!: removing final machine support by @mm-zk in #132
• fix: bigint fix & naming typos by @mm-zk in #139
• feat: Add capability to merge multiple fris by @EmilLuta in #141

New Contributors

• @EmilLuta made their first contribution in #141

Full Changelog: v0.4.5...v0.5.0

v0.4.5

What's Changed

• fix: Updating readme to match latest arguments and requirements by @mm-zk in #116
• feat(gpu_prover): coset and tree recomputations by @robik75 in #120
• fix(prover_examples): correct the argument in trace_execution_for_gpu call by @robik75 in #124

Full Changelog: v0.4.4...v0.4.5

v0.4.4

Allow GPU to work for both airbender & boojum in the same binary

What's Changed

• chore: moving code from cli to execution_utils by @olesHolem in #107
• chore(gpu_prover): refactor native code namespaces and exports to avoid naming conflicts during linking by @robik75 in #109

Full Changelog: v0.4.3...v0.4.4

v0.4.3

What's Changed

• fix: Apply effect of StacktraceCollectionResult to fix diagnostics by @antoniolocascio in #108

Full Changelog: v0.4.2...v0.4.3

v0.4.2

Main changes

• Fix for the bigint bug in risc_simulator.

What's Changed

• fix: fix u256 delegation by @AntonD3 in #105
• chore: moving functions from cli_utils to execution_utils by @mm-zk in #106

New Contributors

• @AntonD3 made their first contribution in #105

Full Changelog: v0.4.1...v0.4.2

v0.4.1

What's Changed

• perf(gpu_prover): reduce startup time, remove global host allocator and reduce pinned memory usage by @robik75 in #103

Full Changelog: v0.4.0...v0.4.1

v0.4.0

Main change

2^23 final circuit that allows proving on smaller GPUs.

What's Changed

• feat(gpu_prover): allow enabling no_cuda building via the ZKSYNC_USE_CUDA_STUBS environment variable by @robik75 in #101
• feat: New recursion modes using reduced 2^23 machine by @olesHolem in #102

New Contributors

• @olesHolem made their first contribution in #102

Full Changelog: v0.3.6...v0.4.0

v0.3.6

What's Changed

• feat: Make new simulator state adhere to sim interface by @aikixd in #67
• chore: update build script, CI and docker build script to CUDA 12.9.1/13.0 by @robik75 in #92
• feat(gpu_prover): implement and use async host allocator inside prove function by @robik75 in #94
• perf(gpu_prover): Reduce NTT register spilling caused by nvcc 13.0 by @mcarilli in #95
• feat(gpu_prover): new logic for rotation of trace buffers by @robik75 in #96
• docs(pru_prover): add README.md with a How to add a new circuit section by @robik75 in #99

New Contributors

• @aikixd made their first contribution in #67

Full Changelog: v0.3.5...v0.3.6