sascha-coenen commented on issue #6794: Query vectorization. URL: https://github.com/apache/incubator-druid/pull/6794#issuecomment-460818215 Great PR!! I do not want to go off topic, but with regard to: >> Benchmarks on the code in this PR indicate we could get a 1.3-3x speedup ..., I meant to mention that there is an Israeli startup which is specialising in FPGA accelerations. With AWS F1 instance types being a thing now, FPGA might finally be getting some more traction. This startup has customised Druid's Query Engine to delegate the aggregation to FPGA hardware and they claim a 35x speedup for mostly-in-memory use-cases. They might be an interesting technology partner in validating the design of a vectorisation layer and providing an FPGA based extension for it. I'm not affiliated with them in any way, so this is not me advertising or pitching them. I just found it interesting what they are doing and was curious as to whether you guys have heard of them and whether this sort of stuff would be a fit for what you are trying to achieve with this PR.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
