-
# Overview
Some `cub::Device*` algorithms are/were documented to be run-to-run deterministic, but the implementations no longer fulfill that guarantee. This has been a major pain point for several …
-
Some of the suggested uses cases seem to require that sources of non-determinism (such as RNG seeds and the current time) are captured to ensure that, for example, saving and sharing a web page can oc…
-
**Describe the bug**
The [`Fft` generation here](https://github.com/zama-ai/tfhe-rs/blob/5c189d6bf3330370c54aba68a39129d2a01b96ea/tfhe/src/core_crypto/fft_impl/fft64/math/fft/mod.rs#L165) is based …
-
In the proto file of the liveupdate manifest, the field '**version**' in **ManifestFile** is marked as '**optional**', although nothing works without it. This is slightly confusing during development,…
-
-
Hi, for the MLTR and RandomSurvivalForest I get different estimates for survival probabilities on each run.
Is there any parameter to regulate training?
-
#### `build.boot` content (if applicable)
```clojure
(set-env!
:resource-paths #{"resources"})
(deftask aux []
(merge-env! :source-paths #{"src"}
:resource-pa…
-
```bash
cd coreir
git checkout determinism
make clean
make install -j
pip install delegator.py pytest
while pytest tests/binary/test_concat.py; do :; done # should eventually fail
```
Conte…
-
Comparing prod and prod2 as part of #3782, @achave11 found some insignificant differences in the `/index/project` and `/index/summary` endpoint responses for otherwise identical catalogs. I think it …
-
The syscall order isn't always consistent in the etcd example. For example, see the following two runs:
```diff
diff --git a/shadow.data/hosts/server1/server1.etcd.1000.strace b/shadow.data.2/host…