0:20 - 8:00 external memory model 8:00 - 22:35 basic results in model 22:35 - 29:45 cache oblivious model 29:45 - 36:15 basic result in model 36:15 - 50:30 static cache-oblivious search trees (nodes laid-out in van Emde Boas order) 50:30 - 1:22:47 dynamic cache-oblivious B tree (ordered file maintenance used implicitly) 54:30 first 3 numbers are prime? coincidence? illuminati confirmed. xD
The better way would be to think of cpu + L1 cache (d/i) and try to optimise for the limited space available, using the rest to feed the L1 cache and cpu.
Is Eric Damiane still there in MIT? I'm listening carefully to these 2 lectures, because I just believe there's a lot of ideas in them that address the problems of(and thus may help in) Merkle Trees & suggested solutions to storing/deleting/inserting UTXOs, retrieving& caching proofs,...etc I just didn't put my hand on them yet. A simple thought just pop while hearing why not Buffer Trees (the idea of flush all UTXO operations once after processing each block), ...etc. Originally came back for re-visiting the thoughts of Van Emde Boas Trees or cache oblivious..... . If Eric Damiane is in MIT & so is the UTREEXO team have u thought of working together or at least exchanging ideas?
0:20 - 8:00 external memory model
8:00 - 22:35 basic results in model
22:35 - 29:45 cache oblivious model
29:45 - 36:15 basic result in model
36:15 - 50:30 static cache-oblivious search trees (nodes laid-out in van Emde Boas order)
50:30 - 1:22:47 dynamic cache-oblivious B tree (ordered file maintenance used implicitly)
54:30 first 3 numbers are prime? coincidence? illuminati confirmed. xD
The better way would be to think of cpu + L1 cache (d/i) and try to optimise for the limited space available, using the rest to feed the L1 cache and cpu.
This one feels so scattered and all over the place
Is Eric Damiane still there in MIT?
I'm listening carefully to these 2 lectures, because I just believe there's a lot of ideas in them that address the problems of(and thus may help in) Merkle Trees & suggested solutions to storing/deleting/inserting UTXOs, retrieving& caching proofs,...etc
I just didn't put my hand on them yet.
A simple thought just pop while hearing why not Buffer Trees (the idea of flush all UTXO operations once after processing each block), ...etc.
Originally came back for re-visiting the thoughts of Van Emde Boas Trees or cache oblivious.....
.
If Eric Damiane is in MIT & so is the UTREEXO team have u thought of working together or at least exchanging ideas?