4 ms·
For server, in Neoverse-V1 this year. For client, stay tuned.
by my123 6y ago
For server, in Neoverse-V1 this year.
For client, stay tuned.
- The_rationalist 6y agoNice. OpenJDK has managed a breakthrough with their hardware agnostic Vector API (the first of its kind?) so that every SIMD algorithm coded with it will work equally well on ARM (and also can target at runtime the best vector width if available)
- iron2disulfide 6y agoTIL about JDK's vector API. That's awesome; I mostly work in the C and C++ space and have learned to embrace the gigantic ifdef's for various arch-specific vectorized instructions.
- nn3 6y agoOr rather work equally badly. Not sure if I would describe this as a breakthrough.
- The_rationalist 6y agoWhy equally badly? Did you miss this part can target at runtime the best vector width if available?
- pjmlp 6y ago.NET is already on its 2nd generation SIMD API, and Intel tried for several years to get one adopted for JavaScript. So not a first of its kind, and the current version is still too low level, with an higher level API expect for some later version.
- phire 6y agoApparently the new client cores that ARM are announcing for next year (Cortex-A79 and friends) will be Armv9. Which means you should be able to pick one up in some flagships phones early next year. Last year, the A78 was announced in July and we had first Snapdragon 888 phone launch on Jan 1st.