Show HN: TensorBase: 5x~10000x Faster Drop-In/Accelerator for ClickHouse in Rust (tensorbase.io) 19 points by jinmingjian 5y ago ↗ HN
[–] cnmjbm 5y ago ↗ Before details of the benchmark is disclosed, whoever believes in this claim, it is not me.
[–] pachico 5y ago ↗ Unless this means "sure, the result was already cached in memory" I don't see how you can be 10000x faster than Clickhouse. This claim is too big. [–] zX41ZdbW 5y ago ↗ Actually in their example, no data is processed.The query compiled to a loop over natural numbers that was optimized by the compiler:uint64_t x = 0; for (uint64_t i = 0; i < 1000000000000; ++i) x += i;print x;This loop is completely eliminated by compiler optimizations.But it does not make much sense to run these queries in SQL :)
[–] zX41ZdbW 5y ago ↗ Actually in their example, no data is processed.The query compiled to a loop over natural numbers that was optimized by the compiler:uint64_t x = 0; for (uint64_t i = 0; i < 1000000000000; ++i) x += i;print x;This loop is completely eliminated by compiler optimizations.But it does not make much sense to run these queries in SQL :)
3 comments
[ 3.6 ms ] story [ 20.6 ms ] threadThe query compiled to a loop over natural numbers that was optimized by the compiler:
uint64_t x = 0; for (uint64_t i = 0; i < 1000000000000; ++i) x += i;
print x;
This loop is completely eliminated by compiler optimizations.
But it does not make much sense to run these queries in SQL :)