Improve SecureRandom benchmarks (especially in multi-threaded settings) #375
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
fabrice102
force-pushed
the
multithreaded-rand-bench
branch
from
09e68b8 to
b3c1eab
Compare
WillChilds-Klein
force-pushed
the
multithreaded-rand-bench
branch
from
b3c1eab to
9eb49d1
Compare
geedo0
previously approved these changes
Apr 5, 2024
amirhosv
suggested changes
Apr 5, 2024
added 2 commits
April 10, 2024 18:28
This is to later add more non-Drbg Random algorithms.
1. Add more benchmarked randomness generation algorithms/providers (SUN/NativePrng, java.util.random, and some baseline using copy instead of randomness generation). SUN/NativePrng is the default Java implementation of SecureRandom when ACCP is not installed.
2. Remove configuration line in build.gradle.kts forcing single thread in all benchmarks. This is because command line arguments override annotations, and we need to use multiple threads in some benchmarks.
3. Use average measurement (ns/op) in benchmarks of randomness generation instead of throughput (op/ns). This is because for multi-threaded benchmarks of randomness generation, we want to know the duration of a single randomness generation on a single thread (and we want to check if performance drops because of multi-threading).
4. Make the data variable (the target of randomness generation) thread-local to avoid L1 cache contention.
5. Add a few more benchmarks:
a. multiThreadedLocal where the SecureRandom instance is thread local, instead of global: this is useful for users of ACCP to know whether they should have SecureRandom be thread local or not.
b. singleThreadedNew creating a new instance of SecureRandom: this is useful for users of ACCP to know whether it is ok to instantiate a SecureRandom every time they need it, or to instead create a global instance.
c. A variant of the benchmarks generating 4 bytes of random data, instead of 1,024 bytes. For example, generating a random int requires 4 bytes of random data.
fabrice102
force-pushed
the
multithreaded-rand-bench
branch
from
b7f9771 to
0fdb6e4
Compare
|
I've rebased to split the commits in two:
This is to facilitate the comparison with diff in Github (using the diff between the second commit and the first commit). |
geedo0
approved these changes
Apr 16, 2024
amirhosv
approved these changes
Jun 20, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Issue #, if available: N/A
Description of changes:
a. multiThreadedLocal where the SecureRandom instance is thread local, instead of global: this is useful for users of ACCP to know whether they should have SecureRandom be thread local or not.
b. singleThreadedNew creating a new instance of SecureRandom: this is useful for users of ACCP to know whether it is ok to instantiate a SecureRandom every time they need it, or to instead create a global instance.
c. A variant of the benchmarks generating 4 bytes of random data, instead of 1,024 bytes. For example, generating a random int requires 4 bytes of random data.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.