feat(examples): Add a useful set of high quality pseudo-random number generators #2868
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
wyhaines
requested review from
jaekwon and
thehowl
and removed request for
a team
github-actions
bot
added
the
🧾 package/realm
Codecov ReportAll modified and coverable lines are covered by tests ✅ 📢 Thoughts on this report? Let us know! |
wyhaines
requested review from
deelawn,
mvertes,
moul,
piux2 and
zivkovicmilos
as code owners
wyhaines
force-pushed
the
kh.add-prngs
branch
from
7f3cc3f to
4510d34
Compare
I ported a number of my pseudo-random number generator implementations from Ruby to Gno, building them be compatible with the standard Rand() implementation, so that any of these can be used as a drop-in replacement for the default PCG algorithm. All of these are faster than PCG, while still having competitive-to-superior statistical properties and predictability resistance. Further, the ISAAC family of generators are cryptographically secure, and when properly seeded, still have no known practical attack vectors.
|
Hey @wyhaines, can you update the PR branch with the master branch? 🙏 |
Revise all of the output to use ufmt.
Tidy gno.mods; add a Sprintf %f test.
…intf float formats a bit more to be closer to Go's support.
wyhaines
force-pushed
the
kh.add-prngs
branch
from
f56bcea to
0fff418
Compare
n2p5
requested changes
Nov 12, 2024
There was a problem hiding this comment.
Very cool work.
It would be valuable to decouple your uint64 and float64 work from the math/random work. The changes you introduce there are generally useful and it would give a cleaner path for merging a PR.
This way the larger discussion on where the isaac and isaac64 work should live can be done in isolation from this. I'm still learning about our namespacing conventions myself and @thehowl brings up valid points that maybe work like this should live in a personal namespace and maybe be considered for "promotion" to official namespaces over time.
|
I'm a bot that assists the Gno Core team in maintaining this repository. My role is to ensure that contributors understand and follow our guidelines, helping to streamline the development process. The following requirements must be fulfilled before a pull request can be merged. These requirements are defined in this configuration file. Automated Checks🟢 Maintainers must be able to edit this pull request Manual Checks
Debug
|
Fix typo. Thanks @thehowl. Co-authored-by: Morgan <[email protected]>
Fix a typo. Thanks @thehowl. Co-authored-by: Morgan <[email protected]>
That's fine with me. This was work that I did as much to make available some interesting things that I have experience with already as it was to just better learn Gno and some of the gnovm internals. So, my intention is to just make sure that it remains a useful set of packages that is available to people. I'll revise this PR to put this under |
|
OK. I have moved all of the PRNGs to |
5 tasks
#3263 has been created. Without that PR, this one will need to retain it's changes to |
wyhaines
added a commit
that referenced
this pull request
Dec 4, 2024
This PR includes the changes to `ufmt.Sprintf()` from #2868 as well as another branch where I had added support for `%v`. These changes add support for a variety of float formatting options to Sprintf, as well as support for the %v flag to automatically chose a default representation for a given type. Tests were added for these additions. This PR is needed for the PRNG PR (#2868) to be fully functional, as the generators include some built in statistical examples/tests which will not function without `ufmt.Sprintf()` support for floats. <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [x] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [x] Updated the official documentation or not needed - [x] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs </details> --------- Co-authored-by: Nathan Toups <[email protected]>
n2p5
approved these changes
Dec 4, 2024
omarsy
pushed a commit
to TERITORI/gno
that referenced
this pull request
Dec 7, 2024
This PR includes the changes to `ufmt.Sprintf()` from gnolang#2868 as well as another branch where I had added support for `%v`. These changes add support for a variety of float formatting options to Sprintf, as well as support for the %v flag to automatically chose a default representation for a given type. Tests were added for these additions. This PR is needed for the PRNG PR (gnolang#2868) to be fully functional, as the generators include some built in statistical examples/tests which will not function without `ufmt.Sprintf()` support for floats. <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [x] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [x] Updated the official documentation or not needed - [x] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs </details> --------- Co-authored-by: Nathan Toups <[email protected]>
omarsy
pushed a commit
to TERITORI/gno
that referenced
this pull request
Dec 7, 2024
… generators (gnolang#2868) I ported a number of my pseudo-random number generator implementations from Ruby to gno while traveling to the retreat last weekend as an exercise in expanding my comfort level with gno code, and expanding my understanding of some of the code internals, while contributing code that others may find interesting or useful. I added two xorshift generators, xorshift64* and xorshiftr128+. These are both many times faster than the PCG generator that is the gno default, and produce high quality randomness with great statistical qualities. In addition to these, I added both the 32-bit ISAAC implementation (with an added function to return 64 bit values), and the 64-bit ISAAC implementation. ISAAC is a stellar pseudo-random number generator. Both implementations are significantly faster than PCG (though not near so fast as the xorshift algorithms), while producing extremely high quality, cryptographically secure randomness that can not be differentiated from real randomness. All of these were built to be compatible with the standard Rand() implementation. This means that any of these can be used as a drop-in replacement for the default PCG algorithm: ``` source = isaac.New() prng := rand.New(source) ``` All of these leverage the `gno.land/p/demo/entropy` package to assist with seeding if no seed is provided. In the case of the ISAAC algorithms, they require 256 uint values for their seed, so they leverage a combination of `entropy` and `xorshiftr128+` to generate any missing numbers in the provided seed. I also added a function to entropy to return uint64, to facilitate using it for seeding. I added tests to entropy, and wrote tests for the other generators, as well. There are a few other things that ended up in this PR. In order to make some fact based assertions about the performance of these generators, I included some code that can be ran via `gno run -expr`. i.e. `gno run -expr 'averageISAAC()' isaac.gno` that can be used to get some benchmarks and some very simple self-statistical-analysis on the results, and when I did so, I discovered that the current `ufmt.Sprintf` implementation didn't support any of the float output flags. I added float support to it's capabilities, which, in turn, required adding `FormatFloat` to the `strconv.gno/strconv.go` implementation in the standard library. I added a test to cover this. I also noticed that there is a test in `tm2/pkg/p2p` that is failing on both master and my branch. Specifically, there is a call to `sw.Logger.Error()` that passes a message and an error, but not `"err"` before the error. Adding that seemed to clear up the build failure. This, specifically, is line 222 of `switch.go`. Currently there is one failing test, which is the code coverage check on tm2, because it is non-obvious to me how to setup a test to properly exercise that one changed line. <details><summary>Contributors' checklist...</summary> - [X] Added new tests, or not needed, or not feasible - [X] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [X] Updated the official documentation or not needed - [X] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [X] Added references to related issues and PRs </details> --------- Co-authored-by: Morgan <[email protected]> Co-authored-by: Nathan Toups <[email protected]> Co-authored-by: Morgan <[email protected]>
r3v4s
pushed a commit
to gnoswap-labs/gno
that referenced
this pull request
Dec 10, 2024
This PR includes the changes to `ufmt.Sprintf()` from gnolang#2868 as well as another branch where I had added support for `%v`. These changes add support for a variety of float formatting options to Sprintf, as well as support for the %v flag to automatically chose a default representation for a given type. Tests were added for these additions. This PR is needed for the PRNG PR (gnolang#2868) to be fully functional, as the generators include some built in statistical examples/tests which will not function without `ufmt.Sprintf()` support for floats. <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [x] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [x] Updated the official documentation or not needed - [x] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs </details> --------- Co-authored-by: Nathan Toups <[email protected]>
r3v4s
pushed a commit
to gnoswap-labs/gno
that referenced
this pull request
Dec 10, 2024
… generators (gnolang#2868) I ported a number of my pseudo-random number generator implementations from Ruby to gno while traveling to the retreat last weekend as an exercise in expanding my comfort level with gno code, and expanding my understanding of some of the code internals, while contributing code that others may find interesting or useful. I added two xorshift generators, xorshift64* and xorshiftr128+. These are both many times faster than the PCG generator that is the gno default, and produce high quality randomness with great statistical qualities. In addition to these, I added both the 32-bit ISAAC implementation (with an added function to return 64 bit values), and the 64-bit ISAAC implementation. ISAAC is a stellar pseudo-random number generator. Both implementations are significantly faster than PCG (though not near so fast as the xorshift algorithms), while producing extremely high quality, cryptographically secure randomness that can not be differentiated from real randomness. All of these were built to be compatible with the standard Rand() implementation. This means that any of these can be used as a drop-in replacement for the default PCG algorithm: ``` source = isaac.New() prng := rand.New(source) ``` All of these leverage the `gno.land/p/demo/entropy` package to assist with seeding if no seed is provided. In the case of the ISAAC algorithms, they require 256 uint values for their seed, so they leverage a combination of `entropy` and `xorshiftr128+` to generate any missing numbers in the provided seed. I also added a function to entropy to return uint64, to facilitate using it for seeding. I added tests to entropy, and wrote tests for the other generators, as well. There are a few other things that ended up in this PR. In order to make some fact based assertions about the performance of these generators, I included some code that can be ran via `gno run -expr`. i.e. `gno run -expr 'averageISAAC()' isaac.gno` that can be used to get some benchmarks and some very simple self-statistical-analysis on the results, and when I did so, I discovered that the current `ufmt.Sprintf` implementation didn't support any of the float output flags. I added float support to it's capabilities, which, in turn, required adding `FormatFloat` to the `strconv.gno/strconv.go` implementation in the standard library. I added a test to cover this. I also noticed that there is a test in `tm2/pkg/p2p` that is failing on both master and my branch. Specifically, there is a call to `sw.Logger.Error()` that passes a message and an error, but not `"err"` before the error. Adding that seemed to clear up the build failure. This, specifically, is line 222 of `switch.go`. Currently there is one failing test, which is the code coverage check on tm2, because it is non-obvious to me how to setup a test to properly exercise that one changed line. <details><summary>Contributors' checklist...</summary> - [X] Added new tests, or not needed, or not feasible - [X] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [X] Updated the official documentation or not needed - [X] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [X] Added references to related issues and PRs </details> --------- Co-authored-by: Morgan <[email protected]> Co-authored-by: Nathan Toups <[email protected]> Co-authored-by: Morgan <[email protected]>
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.
I ported a number of my pseudo-random number generator implementations from Ruby to gno while traveling to the retreat last weekend as an exercise in expanding my comfort level with gno code, and expanding my understanding of some of the code internals, while contributing code that others may find interesting or useful.
I added two xorshift generators, xorshift64* and xorshiftr128+. These are both many times faster than the PCG generator that is the gno default, and produce high quality randomness with great statistical qualities. In addition to these, I added both the 32-bit ISAAC implementation (with an added function to return 64 bit values), and the 64-bit ISAAC implementation. ISAAC is a stellar pseudo-random number generator. Both implementations are significantly faster than PCG (though not near so fast as the xorshift algorithms), while producing extremely high quality, cryptographically secure randomness that can not be differentiated from real randomness.
All of these were built to be compatible with the standard Rand() implementation. This means that any of these can be used as a drop-in replacement for the default PCG algorithm:
All of these leverage the
gno.land/p/demo/entropypackage to assist with seeding if no seed is provided. In the case of the ISAAC algorithms, they require 256 uint values for their seed, so they leverage a combination ofentropyandxorshiftr128+to generate any missing numbers in the provided seed.I also added a function to entropy to return uint64, to facilitate using it for seeding.
I added tests to entropy, and wrote tests for the other generators, as well.
There are a few other things that ended up in this PR. In order to make some fact based assertions about the performance of these generators, I included some code that can be ran via
gno run -expr. i.e.gno run -expr 'averageISAAC()' isaac.gnothat can be used to get some benchmarks and some very simple self-statistical-analysis on the results, and when I did so, I discovered that the currentufmt.Sprintfimplementation didn't support any of the float output flags.I added float support to it's capabilities, which, in turn, required adding
FormatFloatto thestrconv.gno/strconv.goimplementation in the standard library. I added a test to cover this.I also noticed that there is a test in
tm2/pkg/p2pthat is failing on both master and my branch. Specifically, there is a call tosw.Logger.Error()that passes a message and an error, but not"err"before the error. Adding that seemed to clear up the build failure. This, specifically, is line 222 ofswitch.go.Currently there is one failing test, which is the code coverage check on tm2, because it is non-obvious to me how to setup a test to properly exercise that one changed line.
Contributors' checklist...
BREAKING CHANGE: xxxmessage was included in the description