Google has announced that NVIDIA TensorRT, a library for optimizing deep learning models, is now being integrated with TensorFlow. TensorRT will bring a variety of FP16 and INT8 optimizations to TensorFlow. According to Google, TensorRT will automatically select “platform specific kernels to maximize throughput and minimizes latency during inference on GPUs.”
In tests, the company found that ResNet-50 was eight times faster under 7 ms latency when using the TensorFlow-TensorRT integration compared with just running TensorFlow. “We are excited about the new integrated workflow as it simplifies the path to use TensorRT from within TensorFlow with world-class performance,” the team wrote in a post.
TypeScript 2.8 is released
IntelliJ IDEA 2018.1 is released
JetBrains has announced the release of IntelliJ IDEA 2018.1, the company’s Java IDE for developers. It adds several new Java features, such as code completion being aware of type casting, the ability to define custom postfix Java templates, and data flow analysis that tracks relationships between variables such as “greater than” and “less than.”
It also adds updates to the JVM debugger, Java compiler, editor, project configuration, search and replace, Kotlin plugin, Groovy, Scala, Android, Spring Boot, and version control systems, among others. A complete list of new features can be found here.
MariaDB acquires MammothDB
MariaDB has acquired enterprise analytics database company MammothDB. According to the company the new acquisition expands its analytics capabilities and global presence. As part of the acquisition, MariaDB is also adding to its engineering and consulting team and opening a new office in Sofia, Bulgaria.
“The MammothDB team joins MariaDB at a critical point in our growth, bringing with them an impressive track record of delivering big data solutions,” said Michael Howard, CEO of MariaDB Corporation. “Over the past year, we’ve seen a major increase in demand for MariaDB AX as organizations seek to fill an open source analytics gap left by proprietary offerings such as Oracle and Teradata. The addition of MammothDB’s deep analytics expertise will be invaluable to helping MariaDB meet this growing need and continue to innovate our analytic products.”