82 Commits

Author SHA1 Message Date
Daniel Lubarov
3bc34c59d8
Refactor GMiMC code (#224)
* Refactor GMiMC code

Adds a sub-trait of `Field` called `GMiMCInterface`, which is similar to `PoseidonInterface`.

This lets us have different fields with different GMiMC constants in a type-safe way.

* Remove `Interface`

* Const generic for width
2021-09-07 18:28:28 -07:00
Nicholas Ward
c07f99ac81 merge 2021-09-06 21:38:52 -07:00
Daniel Lubarov
236a143abf
Move some Field members to a Field64 subtrait (#213)
* Move some Field members to a Field64 subtrait

I.e. move anything specific to 64-bit fields.

Also, relatedly,
- Tweak a bunch of prover code to require `Field64`, since 64-bit stuff is used in a couple places, like the FRI proof-of-work
- Remove `bits()`, which was unused and assumed a 64-bit field
- Rename a couple methods to reflect that they're u64 variants

There are no functional changes.

* Field64 -> PrimeField

* Remove `exp_u32`, `kth_root_u32`

* PrimeField: PrimeField

* Move `to_canonical_biguint` as well

* Add back from_noncanonical_u128
2021-09-05 10:27:11 -07:00
Daniel Lubarov
f89f49249a wip 2021-09-04 16:27:40 -07:00
wborgeaud
2fcfa230a6 Pr feedback 2021-08-22 10:36:44 +02:00
wborgeaud
1d368782f2 Fix tests 2021-08-20 12:15:15 +02:00
wborgeaud
e81001b9ba Clippy 2021-08-20 11:13:40 +02:00
wborgeaud
c6cf5cf130 Move PartitionWitness 2021-08-20 09:55:49 +02:00
wborgeaud
74c2be5090 First pass 2021-08-19 14:54:11 +02:00
Daniel Lubarov
69193a8dc2
Remove *_three methods (#182)
* Remove *_three methods

Since there's no longer a performance reason for them, and I think the *_many methods are about as short etc.

* PR feedback
2021-08-17 00:38:41 -07:00
wborgeaud
5a9c5b295c Minor 2021-08-16 10:41:12 +02:00
wborgeaud
b366482866 The mother of all arithmetic optimizations 2021-08-16 10:18:10 +02:00
wborgeaud
4433fd8048 Merge branch 'main' into optimize_arithmetic_ops 2021-08-10 09:11:08 +02:00
wborgeaud
b15e36d29c PR feedback 2021-08-10 09:07:01 +02:00
wborgeaud
c3a0b7cc82 PR feedback 2021-08-09 18:35:41 +02:00
wborgeaud
4b44578ffa More optimizations 2021-08-09 12:39:37 +02:00
wborgeaud
dd076e5c73 Auto resize partial witness 2021-08-09 09:58:09 +02:00
wborgeaud
213de80a3a Merge branch 'main' into partial_witness_vec
# Conflicts:
#	src/iop/challenger.rs
2021-08-06 18:17:31 +02:00
wborgeaud
e97b7b0737 PR feedback 2021-08-06 18:06:02 +02:00
wborgeaud
db236e4824 Fix tests 2021-08-06 15:14:38 +02:00
wborgeaud
7d11d0f8a1 Change PartialWitness to use Vecs 2021-08-06 14:58:39 +02:00
wborgeaud
a6490de7ad routed_wires -> wires 2021-08-06 11:30:56 +02:00
wborgeaud
8531cf042a Replace HashMap by Vec in generate_partial_witness 2021-08-06 10:52:11 +02:00
Daniel Lubarov
7b20f342df
More eval_unfiltered_recursively tweaks (#156) 2021-08-05 08:03:49 -07:00
Daniel Lubarov
8b8e4d223d
Optimize witness generation a bit (#153)
Mainly storing pending generators in a Vec rather than a HashMap.  Requires an extra check to make sure we don't run one twice after adding it to the Vec twice.
2021-08-04 09:55:11 -07:00
Nicholas Ward
fa5a5c5ebf cargo fmt 2021-08-03 08:59:25 -07:00
Daniel Lubarov
5c96e7b366 Trivial fixes from "cargo fix" 2021-08-02 10:58:03 -07:00
Daniel Lubarov
d6211b8ab8
Reuse a buffer of generated values (#142)
* Reuse a buffer of generated values

To avoid allocating `GeneratedValues` all the time. Saves ~60ms or so.

* PR feedback
2021-08-02 10:55:10 -07:00
wborgeaud
2bbcf17699 Test eval_unfiltered_* functions 2021-08-02 14:17:42 +02:00
Daniel Lubarov
018fb005f8
Move stuff around (#135)
No functional changes here. The biggest change was moving certain files into new directories like `plonk` and `iop` (for things like `Challenger` that could be used in STARKs or other IOPs). I also split a few files, renames, etc, but again nothing functional, so I don't think a careful review is necessary (just a sanity check).
2021-07-29 22:00:29 -07:00
Daniel Lubarov
e382decc9f Import fixes 2021-07-29 11:45:58 -07:00
Daniel Lubarov
bcf524bed0
Have add_gate take a generic type instead of GateRef (#125)
* Have add_gate take a generic type instead of GateRef

There are a couple advantages
- Users writing their own gates won't need to know about the `GateRef` wrapper; it's more of an internal thing now.
- Easier access to gate methods requiring `self` -- for example, `split_le_base` can just call `gate_type.limbs()` now.

* Update comment

* Always insert
2021-07-22 23:48:03 -07:00
wborgeaud
a70e97befc Fix merge issues 2021-07-23 08:21:55 +02:00
wborgeaud
9a1c289f8e Merge branch 'main' into remove_acc_in_gmimc
# Conflicts:
#	src/gadgets/arithmetic.rs
2021-07-23 08:16:23 +02:00
wborgeaud
ca3a2fcfc8 Clippy 2021-07-22 16:09:54 +02:00
wborgeaud
15a64017dc We need only 126 wires now 2021-07-22 14:42:42 +02:00
wborgeaud
467485c3f0 Remove accumulator wires in GMiMCGate 2021-07-22 14:33:02 +02:00
wborgeaud
3a24e8f4c1 Manually implement eval_unfiltered_base for all gates 2021-07-22 14:00:55 +02:00
Daniel Lubarov
eb18c7ea33
Faster witness generation (#116)
Saves ~300ms in the test. The main change is to have generators return fixed-size `Vec`s instead of `HashMap`s, which have more overhead.
2021-07-21 08:26:34 -07:00
Daniel Lubarov
b8ce1d1967
Public inputs (#113)
With this approach, we don't need `Target::PublicInput`; any routable `Target` can be marked as a public input via `register_public_input`.  The circuit itself hashes these targets, and routes the hash output to the first four wires of a `PublicInputGate`, which is placed at an arbitrary location in the circuit.

All gates have direct access to the purported hash of public inputs. We could think of them as accessing `PI_hash_i(x)` (as in Plonk), but these are now (four) constant functions, so they effectively have direct access to the hash itself.

`PublicInputGate` checks that its first four wires match this purported public input hash. The other gates ignore the hash.

Resolves #64.
2021-07-21 08:26:19 -07:00
Daniel Lubarov
35c8643681 Tiny cleanup 2021-07-18 23:31:21 -07:00
Daniel Lubarov
0a5d46bfa9
Have prove return Result (#100)
* Have `prove` return `Result`

To address that TODO.

* PR feedback
2021-07-18 23:14:48 -07:00
wborgeaud
097413479e PR feedback 2021-07-18 10:35:42 +02:00
wborgeaud
fe05da6720 Clipp 2021-07-15 10:39:57 +02:00
wborgeaud
139430c549 Fixed GMiMC 2021-07-13 09:15:16 +02:00
wborgeaud
fc0f8a78ce First try 2021-07-02 14:13:57 +02:00
wborgeaud
3bc27c65ef Rollback to previous semantics 2021-06-23 13:46:19 +02:00
wborgeaud
ea6a724560 Enforce copy constraints in partial witness generation. 2021-06-14 17:02:52 +02:00
Daniel Lubarov
cb7f8c8b8c
Draw challenge points from the extension field (#51)
* Draw challenge points from the extension field

* Now building

* Misc

* Default eval_unfiltered_base

* fmt

* A few field settings

* Add to Sage

* Display tweak

* eval_filtered_base

* Quartic in bench

* Missing methods

* Fix tests

* PR feedback
2021-05-30 13:25:53 -07:00
Daniel Lubarov
741dc4c94c Tweaks to bench_ldes 2021-05-22 09:38:07 -07:00