Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Barret reduction for modexp #170

Draft
wants to merge 59 commits into
base: main
Choose a base branch
from
Draft

Barret reduction for modexp #170

wants to merge 59 commits into from

Conversation

fkrause98
Copy link
Contributor

@fkrause98 fkrause98 commented Oct 23, 2023 •
edited
Loading

Pending tasks:

  • Finish barret_reduction implementation.
  • Precalculate needed constants.
  • Do not hardcode addresses.
  • Add the necessary scratch buffers.
  • Integrate into modexp respecting the following:
    1. Make sure this reduction is done AFTER the first loop so the base is inside the [0, n**2) range.
    2. Make sure the reduction is not used if the modulo is a power of 2, which can done "easily" through
    ands.

Useful references:

ilitteri and others added 30 commits September 22, 2023 13:11
@ilitteri
Initial new implementation
7e59313
@ilitteri
Fix compilation error
9206d42
@IAvecilla
Implementation of add operation for big integers (#136)
9f11e74
@ilitteri
Implement Big UInt Left Shift (#139)
db34989 
* Add `bigUIntShl` implementation

* Add constants

* Fix compilation
@ilitteri
Implement Big UInt Right Shift (#137)
c954c0c 
* Add `bigUIntShr` implementation

* Add constants
@jpcenteno @ilitteri
Implement Big UInt Bitwise Or for modexp (#135)
16558cb 
* Implement bigUIntBitOr

* Fix bigUIntBitOr

* Fix bigUIntBitOr

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Fix missing closing brackets

---------

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
@jpcenteno @ilitteri
Implement big uint conditional select for modexp (#148)
b1645c0 
* Implement bigUIntCondSelect

* Fix missing curly braces

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Implement Big UInt Right Shift (#137)

* Add `bigUIntShr` implementation

* Add constants

* Implement Big UInt Bitwise Or for modexp (#135)

* Implement bigUIntBitOr

* Fix bigUIntBitOr

* Fix bigUIntBitOr

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Fix missing closing brackets

---------

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

---------

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
@IAvecilla
Implement mul operation for big UInts (#151)
02db93f 
* First implementation of mul operation for bigints

* Fix multiplication for big integers

* Fix some merge issues

* Improve comments and function docs

* Delete whitespaces
@fkrause98 @jpcenteno @ilitteri
Substraction with borrow (#149)
2ff02cc 
* First substraction draft

* Fix compile problems

* Working implementation

* Updated code

* Updated code

* Update subtract implementation

* Remove console_log

* Add docs for function

* Update function docs

* Remove tests from from ModExp.yul

* Fix typo

* Restore horrible whitespaces to avoid an ugly merge conflict

* Update precompiles/Modexp.yul

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Revert "Update precompiles/Modexp.yul"

This reverts commit 582bc41.

---------

Co-authored-by: Joaquín P. Centeno <jpcenteno@users.noreply.github.com>
Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
@ilitteri
Refactor modexp reimplementation (#156)
b86bed6 
* Make Big UInt API functions naming consistent

* Refactor `bigUIntAdd` variable names

* Refactor `bigUIntMul` variable names

* Refactor `subLimbsWithBorrow`

* Refactor `bigUintSubtractionWithBorrow`

* Refactor `bigUIntAdd`

* Fix `bigUIntSubWithBorrow`

* Format `storeLimbValueAtOffset`

* Refactor `bigUIntBitOr`

Made it consistent with the rest of the code convention and naming

* Refactor `bigUIntCondSelect`

Made it consistent with the rest of the code convention and naming

* Reorder `overflowingSubWithBorrow`

* Move comment to modexp API Docs section
@jpcenteno @fkrause98
@IAvecilla @ilitteri
Biguint division (#159)
4a6a723 
* Division draft

* Non working draft

* Fix compile errors

* Use proper pointers for quotient and remainder

* Add fix note, some more changes

* Add comment

* Implement `big_uint_bit_size`

* Increase pointer to prevent it from steping over console_log

* WIP divrem

* Fix loop and zero initializer

* Push test cases

* Add other test case

* Add docs and tests for `big_uint_inplace_or_1`

* Fix bug related to bit shifting

* Fix borrow return in big uint sub function

* Delete playground file used for debugging

* Fix sub with borrow function

* Add playground again to check more big integer division tests

* Remove playground used for testing

* Write documentation for new shift functions

* Improve naming and documentation for new helper functions

* Rename bigUIntOrWith1 to bigUintInPlaceOrWith1

* Add tmp buffer parameters to bigUIntDivRem. Improve docs.

* Simplify subLimbsWithBorrow

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Remove `mul` call from `bigUIntInPlaceOrWith1`

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Remove multiplications from copyBigUint

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Optimize bigUIntBitSize loop

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Simplify zeroWithLimbSizeAt

---------

Co-authored-by: Francisco Krause Arnim <fkrausear@gmail.com>
Co-authored-by: IAvecilla <iavecilla@fi.uba.ar>
Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
@IAvecilla @fkrause98 @jpcenteno
Implement mul mod operation for big UInts (#161)
06df32c 
* Division draft

* Non working draft

* Fix compile errors

* Use proper pointers for quotient and remainder

* Add fix note, some more changes

* Add comment

* Implement `big_uint_bit_size`

* Increase pointer to prevent it from steping over console_log

* WIP divrem

* Fix loop and zero initializer

* Push test cases

* Add other test case

* Add docs and tests for `big_uint_inplace_or_1`

* Fix bug related to bit shifting

* Fix borrow return in big uint sub function

* Delete playground file used for debugging

* Fix sub with borrow function

* Add playground again to check more big integer division tests

* Remove playground used for testing

* Write documentation for new shift functions

* Improve naming and documentation for new helper functions

* Rename bigUIntOrWith1 to bigUintInPlaceOrWith1

* Add tmp buffer parameters to bigUIntDivRem. Improve docs.

* Add big uint mul mod skeleton

* Remove wrong comment

* Update algorithm comment

* Add limb size doubling and divide by two for mul mod operation

* Functions to duplicate and halve limb size work in place

* Use camelCase

* Remove console_log

* Add docs

* Update doc

---------

Co-authored-by: Francisco Krause Arnim <fkrausear@gmail.com>
Co-authored-by: Joaquín P. Centeno <jpcenteno@users.noreply.github.com>
@fkrause98
Add parseCallData function
1fc184d
@IAvecilla
Add function to left-pad big uints
174b160
@ilitteri
Merge branch 'main' of github.com:lambdaclass/zksync_era_precompiles …
92010d6 
…into modexp_reimplementation
@IAvecilla
Remove console log function
776dc1d
@IAvecilla
Change left padding functions for big uints to not work in place
3c0f00c
@ilitteri
Add parseCalldata function (#168)
1b8b2f6
@fkrause98
Remove redundant parse call data declaration
b775975
@fkrause98 @ilitteri
Free memory pointer (#169)
ebc92a8 
* Add free memory pointer function

* Update precompiles/Modexp.yul

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

* Update precompiles/Modexp.yul

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>

---------

Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
@IAvecilla
Start parsing the input calldata
55ace22
@fkrause98
Correctly parse call data
bfbcbae
@IAvecilla
Add left pad steps for modexp inputs
4fabd3c
@fkrause98
Add pad if needed function
1f51a9e
@jpcenteno @IAvecilla
@ilitteri @fkrause98
Modexp for big UInts skeleton (#164)
730931e 
* WIP: modexp skeleton

* Use of mul mod function for big integers

* imlement aux function to check if big uint is larger than 1

* minor fix

* Restore modexp from target branch

* Fix mul mod

* Finish modexp implementation

* Remove playground used for debugging

* Update modexp with final state of modular exponentiation function

* Fix merge issue

* Change all names to camel case

* fix typo

Co-authored-by: Francisco Krause Arnim <56402156+fkrause98@users.noreply.github.com>

---------

Co-authored-by: IAvecilla <iavecilla@fi.uba.ar>
Co-authored-by: Ivan Litteri <67517699+ilitteri@users.noreply.github.com>
Co-authored-by: Francisco Krause Arnim <56402156+fkrause98@users.noreply.github.com>
@fkrause98
Add simple integration
54fe792
@IAvecilla
Fix calldata buffer in zero check
2ecd8a3
@IAvecilla
Uncomment checks for base cases
027ccfe
@IAvecilla
Fix result length to match with mod length
01f69aa
@IAvecilla
Fix condition in parse call data
b7da051
@fkrause98 fkrause98 marked this pull request as draft October 23, 2023 21:21
Base automatically changed from modexp_reimplementation to main October 30, 2023 23:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@fkrause98 @ilitteri @IAvecilla @jpcenteno