-
Notifications
You must be signed in to change notification settings - Fork 11
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.
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
Add reach only shuffle and aggregation phase. #1600
base: main
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 10 of 10 files at r1, all commit messages.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @kungfucraig, @ple13, and @renjiezh)
MODULE.bazel
line 328 at r1 (raw file):
) archive_override(
nit: Add a DO_NOT_SUBMIT comment referencing the any-sketch PR so it's
- Clear that this is intended to be changed before submitting/merging, and
- Clear which pending PR this code comes from.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 200 at r1 (raw file):
RETURN_IF_ERROR(VerifySketchParameters(request.sketch_params())); // Verify that the ring modulus is a prime.
nit: Extract an IsPrime
bare function. This way that function could be independently changed to e.g. use a more efficient implementation. It also makes the code self-documenting, making the comment unnecessary.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 477 at r1 (raw file):
// Count the non-zero registers. int64_t non_empty_register_count = 0; for (auto x : combined_sketch) {
nit: qualify auto
as much as possible. In this case, I believe it can be const auto
, since nothing will be mutating the copy of the integer.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 477 at r1 (raw file):
// Count the non-zero registers. int64_t non_empty_register_count = 0; for (auto x : combined_sketch) {
nit: prefer meaningful variable names. The common exception is loop indices, but this isn't an index.
Code quote:
x
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 10 of 10 files at r1, all commit messages.
Reviewable status: all files reviewed, 5 unresolved discussions (waiting on @ple13 and @renjiezh)
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.h
line 28 at r1 (raw file):
CompleteShufflePhaseResponse; absl::StatusOr<CompleteShufflePhaseResponse> CompleteShufflePhase(
nit: Can you add comments that Complete Shuffle Phase and CompleteAggregationPhase apply to Reach and Frequency measurements? You could consider renaming them in a follow-up PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewable status: 1 of 15 files reviewed, 2 unresolved discussions (waiting on @kungfucraig, @renjiezh, and @SanjayVas)
MODULE.bazel
line 328 at r1 (raw file):
Previously, SanjayVas (Sanjay Vasandani) wrote…
nit: Add a DO_NOT_SUBMIT comment referencing the any-sketch PR so it's
- Clear that this is intended to be changed before submitting/merging, and
- Clear which pending PR this code comes from.
Thanks. I've cut a release for any-sketch and it's ready now.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.h
line 28 at r1 (raw file):
Previously, kungfucraig (Craig Wright) wrote…
nit: Can you add comments that Complete Shuffle Phase and CompleteAggregationPhase apply to Reach and Frequency measurements? You could consider renaming them in a follow-up PR.
I've changed the name of the functions.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 200 at r1 (raw file):
Previously, SanjayVas (Sanjay Vasandani) wrote…
nit: Extract an
IsPrime
bare function. This way that function could be independently changed to e.g. use a more efficient implementation. It also makes the code self-documenting, making the comment unnecessary.
Done.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 477 at r1 (raw file):
Previously, SanjayVas (Sanjay Vasandani) wrote…
nit: qualify
auto
as much as possible. In this case, I believe it can beconst auto
, since nothing will be mutating the copy of the integer.
Done.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 477 at r1 (raw file):
Previously, SanjayVas (Sanjay Vasandani) wrote…
nit: prefer meaningful variable names. The common exception is loop indices, but this isn't an index.
Done.
@@ -72,9 +72,26 @@ absl::Status VerifySketchParameters(const ShareShuffleSketchParams& params) { | |||
return absl::OkStatus(); | |||
} | |||
|
|||
// Checks if val is a prime. | |||
absl::StatusOr<bool> IsPrime(int val) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is an inefficient way to check primality. It should be OK given how it is used, but we should be careful with this function (especially if we ever need a larger modulus); can you add a note about the O(sqrt(val)) running time to the comment above?
} | ||
for (int j = 0; j < sketch_size; j++) { | ||
// It's guaranteed that (combined_sketch[j] + share_vector[j]) is | ||
// not greater than 2^{32}-1. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think it is worth having a CHECK here.
// Transform share of non-zero registers to share of a non-zero random value. | ||
for (int j = 0; j < sketch_size; j++) { | ||
combined_sketch[j] = | ||
(combined_sketch[j] * r[j]) % request.sketch_params().ring_modulus(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe it is also worth having a CHECK here? As an alternative we can use 64 bit arithmetic here e.g. change the type of 'r' to 'std::vector<uint64_t>'.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 15 of 15 files at r2, all commit messages.
Reviewable status: all files reviewed, 5 unresolved discussions (waiting on @kungfucraig, @ple13, and @renjiezh)
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 77 at r2 (raw file):
// Checks if val is a prime. absl::StatusOr<bool> IsPrime(int val) { if (val < 0) {
Just return false in this case so the return type can be bool
. It's not incorrect to ask if a negative integer is prime, it's just that the answer is always "no".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewable status: 14 of 15 files reviewed, 5 unresolved discussions (waiting on @benjaminkreuter, @kungfucraig, @renjiezh, and @SanjayVas)
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 76 at r2 (raw file):
Previously, benjaminkreuter wrote…
This is an inefficient way to check primality. It should be OK given how it is used, but we should be careful with this function (especially if we ever need a larger modulus); can you add a note about the O(sqrt(val)) running time to the comment above?
Done.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 77 at r2 (raw file):
Previously, SanjayVas (Sanjay Vasandani) wrote…
Just return false in this case so the return type can be
bool
. It's not incorrect to ask if a negative integer is prime, it's just that the answer is always "no".
Done.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 246 at r2 (raw file):
Previously, benjaminkreuter wrote…
I think it is worth having a CHECK here.
I've replaced it with a util function that handles the overflow.
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 268 at r2 (raw file):
Previously, benjaminkreuter wrote…
Maybe it is also worth having a CHECK here? As an alternative we can use 64 bit arithmetic here e.g. change the type of 'r' to 'std::vector<uint64_t>'.
Thanks Ben for catching this! I've updated the code.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 14 of 15 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @benjaminkreuter, @renjiezh, and @SanjayVas)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 1 of 1 files at r3, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @benjaminkreuter and @renjiezh)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 14 of 15 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @benjaminkreuter and @renjiezh)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 2 of 2 files at r4, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @benjaminkreuter and @renjiezh)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewed 7 of 15 files at r2, 2 of 2 files at r4, all commit messages.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @benjaminkreuter and @ple13)
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 223 at r4 (raw file):
} const int sketch_size = request.sketch_params().register_count();
Do we need to check if sketch_size
is 0?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewable status: 13 of 15 files reviewed, 4 unresolved discussions (waiting on @benjaminkreuter, @kungfucraig, @renjiezh, @SanjayVas, and @stevenwarejones)
src/main/cc/wfa/measurement/internal/duchy/protocol/share_shuffle/honest_majority_share_shuffle_utility.cc
line 223 at r4 (raw file):
Previously, renjiezh wrote…
Do we need to check if
sketch_size
is 0?
Yeah, I've added the check.
No description provided.