Wasmer 2.0, Its a big deal!
It's been almost 6-months since we released Wasmer 1.0, and the community support and reception were incredible. Since then, we've been hard at work on the next version of Wasmer.
Wasmer 2.0 is packed full of new features that add stability, security, and performance improvements. All public APIs remain the same, with a few exceptions on internal APIs that will not affect 99% of current users. However, passionate discussions within the engineering team about semantic versioning pushed us towards 2.0.
Our goals for Wasmer 2.0 are to provide the community with immediate value and set up the runtime to support future value. The main additions to the latest release include SIMD and Reference Types. Let's dive into the details.
SIMD stands for simple instruction, multiple data. In short, SIMD enables one instruction to operate on numerous pieces of data at the same time. The introduction of SIMD into Wasmer 2.0 empowers the community with higher performance data-level parallelism for many different use cases. Users creating machine-learning, video (transcoding, encoding, decoding...etc.), image processing, 3D physics, or graphics applications should see a significant boost in performance. Perhaps an underappreciated benefit of SIMD is its potential for reducing the energy signature of your applications.
Another prominent feature for Wasmer 2.0 is the inclusion of reference types. For the uninitiated, Reference Types allow Wasm modules to reference and share specific types of information with host environments or between multiple Wasm modules.
