4 ms·
SIMD call in SBCL is explicit, which I personally like. In NumPy, as a counter example, it just happens. It is just a method of talking to the processor to take
by medo-bear 3y ago
SIMD call in SBCL is explicit, which I personally like. In NumPy, as a counter example, it just happens. It is just a method of talking to the processor to take advantage of parallelization. Java will also have it soon: https://download.java.net/java/early_access/loom/docs/api/jdk.incubator.vector/jdk/incubator/vector/Vector.html https://download.java.net/java/early_access/loom/docs/api/jd...
- igouy 3y agoDoes that comparison show hand-coded SIMD in Fortran? (That comparison no longer seems to show Lisp?)
- medo-bear 3y agoI guess so. Most of the languages compilers use SIMD under the hood -ie implicitly. See https://stackoverflow.com/questions/1516622/what-does-vectorization-mean https://stackoverflow.com/questions/1516622/what-does-vector...