That's a very interesting collaboration and I'm looking forward to see where it ends up. Bringing the "2 worlds" together is going to be fascinating.
By the way, a small note on this memory map 2 GB limitation. The Lucene search library (which powers Solr, ElasticSearch, OpenSearch) makes heavy usage of that one and does so for indexes that might even be hunderds of GBs in size. Looks like they had to do it in chunks of 2GB as well but in later versions they override the limitation by making use of Java 21+ 64-bit version of that through MemorySegment.
It might also be interesting to compare with a C version that solves the same "1B row challenge".
Am I the only one seeing "Video unavailable. Error code: 15" on both embedded videos?
Casey, could you, please, post the link to the interview separately as well? I can access the playlist link just fine... Thank you Substack
Update:
Oh, it works on desktop, but not in their app, lol. Here is the link to the interview for anyone else having trouble: https://www.youtube.com/watch?v=apREl0KmTdQ
Great :-) As an OpenJDK guy, I'm pretty excited by Casey entering our waters :-D
That's a very interesting collaboration and I'm looking forward to see where it ends up. Bringing the "2 worlds" together is going to be fascinating.
By the way, a small note on this memory map 2 GB limitation. The Lucene search library (which powers Solr, ElasticSearch, OpenSearch) makes heavy usage of that one and does so for indexes that might even be hunderds of GBs in size. Looks like they had to do it in chunks of 2GB as well but in later versions they override the limitation by making use of Java 21+ 64-bit version of that through MemorySegment.
It might also be interesting to compare with a C version that solves the same "1B row challenge".
"unannounced project"? uuuuuuhhh.. is that 1935 maybe!?