JBrainy issueshttps://git.cs.lth.se/noricc/jbrainy/-/issues2021-06-30T08:47:11Zhttps://git.cs.lth.se/noricc/jbrainy/-/issues/17Is there an impact on performance when you use timed collections?2021-06-30T08:47:11ZNoric CoudercIs there an impact on performance when you use timed collections?Compare the running times of benchmarks with normal and timed collections.Compare the running times of benchmarks with normal and timed collections.https://git.cs.lth.se/noricc/jbrainy/-/issues/16Add iteration to the list of timed operations on timed collections?2021-06-30T09:19:50ZNoric CoudercAdd iteration to the list of timed operations on timed collections?We track how long some operations take (insertion, deletion, search), how about iterations (for example in the case of `hashCode` and `toString`) it might be useful.We track how long some operations take (insertion, deletion, search), how about iterations (for example in the case of `hashCode` and `toString`) it might be useful.Noric CoudercNoric Couderchttps://git.cs.lth.se/noricc/jbrainy/-/issues/15Use time() to set the seed.2021-08-03T14:48:49ZNoric CoudercUse time() to set the seed.Right now we use a constant seed for generating the benchmark. Brainy uses the current `time()`.Right now we use a constant seed for generating the benchmark. Brainy uses the current `time()`.Noric CoudercNoric Couderchttps://git.cs.lth.se/noricc/jbrainy/-/issues/14Select important features for classification2021-06-23T12:13:42ZNoric CoudercSelect important features for classificationBrainy only gathers a subset of the PAPI counters to do classification. They used genetic algorithms to pick the right subset.Brainy only gathers a subset of the PAPI counters to do classification. They used genetic algorithms to pick the right subset.Noric CoudercNoric Couderchttps://git.cs.lth.se/noricc/jbrainy/-/issues/11Compare diversity of benchmarks for different generation schemes2021-06-23T11:52:30ZNoric CoudercCompare diversity of benchmarks for different generation schemesWe implemented a new distribution for methods with the BRAINY mode. We would like to compare that with other benchmark generation methods (Uniform, Polya, Markov). What we want to see is the number of winners for each collection and benc...We implemented a new distribution for methods with the BRAINY mode. We would like to compare that with other benchmark generation methods (Uniform, Polya, Markov). What we want to see is the number of winners for each collection and benchmark methods.Noric CoudercNoric Couderchttps://git.cs.lth.se/noricc/jbrainy/-/issues/6Test readCSVUnsafe against reference implementation2021-02-18T14:23:31ZNoric CoudercTest readCSVUnsafe against reference implementationI wrote a function called `readCSV` (which works), and I am now trying a function called `readCSVUnsafe` which is optimized to take less memory while loading the data.
I have tests for `readCSV`, and I assume they could be easily adapt...I wrote a function called `readCSV` (which works), and I am now trying a function called `readCSVUnsafe` which is optimized to take less memory while loading the data.
I have tests for `readCSV`, and I assume they could be easily adapted to test `readCSVUnsafe` too.Noric CoudercNoric Couderchttps://git.cs.lth.se/noricc/jbrainy/-/issues/2There was a bug with Map synthetic benchmarks.2020-03-12T17:59:22ZNoric CoudercThere was a bug with Map synthetic benchmarks.The method `runPutAll` did not add any elements because the `argument` it uses was not filled when building the benchmark.The method `runPutAll` did not add any elements because the `argument` it uses was not filled when building the benchmark.https://git.cs.lth.se/noricc/jbrainy/-/issues/1The methods "get(int)" and "get(Object)" collided in the function table2020-03-11T18:34:17ZNoric CoudercThe methods "get(int)" and "get(Object)" collided in the function tableThis should not be a big problem, since the `get()` method has different arguments for different interfaces... I changed the name of the mapped `runGet` method to `runGet` for ints (for lists) and `runGetObject` for maps. That means, how...This should not be a big problem, since the `get()` method has different arguments for different interfaces... I changed the name of the mapped `runGet` method to `runGet` for ints (for lists) and `runGetObject` for maps. That means, however, that there will be a bit of processing on the resulting training data before training.