Eric N Hanson

Principal Product Manager, MemSQL

[Eric N Hanson @HansonOnData]

@JackEllis @SingleStoreDB It could be better or worse depending on what you're doing. Certain optimizations like AVX-2 vectorized query execution are disabled. Given the 7-year gap, performance on your 2018 Intel vs 2025 ARM Mac might be in the same ballpark.

[Eric N Hanson @HansonOnData]

@JackEllis @SingleStoreDB Hello Jack! We don't have any benchmarks on the difference between our performance on Intel vs. interpreted (based on QEMU) on ARM-based Macs, but our general experience is that it is around 10x slower in emulation compared with running on an Intel chip of the same generation.

[Eric N Hanson @HansonOnData]

#SingleStore had an open position for a database programmability/extensibility Product Manager in India! Reporting to me.
https://t.co/Kr8H79ZnyS
#DBMS #JSON #WASM #productmanagement #productmanager #Database #SQL #SingleStoreDB
Are you qualified? Please apply or refer people!

[Eric N Hanson @HansonOnData]

Amazing joint work between SingleStore and SAS! Such a Pleasure to see this delivered.

Consolidate Complex Data Workflows Into Fast, Impactful Business Insights https://t.co/X8pOUo1hd3 #singlestore #singlestore #product via @singlestoredb

[Sam Altman @sama]

seeing very early but really impressive demos of using language interfaces to control computers.

the leverage on human ability and creativity we are going to see in the next decade from these tools will be quite something to behold!

[Eric N Hanson @HansonOnData]

@smartjoe221434 @tobias_petry Or just keep data in a table with a datetime column and when needed, use a SQL DELETE to remove old data. Some modern DBs like SingleStore can delete data so fast, you don't need to bother with the complexity of range partitioning.

[Eric N Hanson @HansonOnData]

@markcallaghan @kena42 @memsql @tpcbenchmarks Ok, we will talk to TPC. We have published at VLDB and SIGMOD in recent years on QP, though nothing on transactions. Let me talk to our storage devs -- might be good to write a paper on the new replication/durability stuff in 7.0. It draws concepts from Aries and has group commit

[Eric N Hanson @HansonOnData]

@markcallaghan @narayanarjun We use SELECT FOR UPDATE (https://t.co/VFntpJAsS2) in our TPC-C benchmark kit to ensure serializability. @adamprout talks about his more on HN, https://t.co/9Sr6si30i5

[Eric N Hanson @HansonOnData]

@narayanarjun @memsql @CockroachDB @AdamProut Also, we put up individual query runtimes for TPC-DS in an appendix now. And we put in GEOMEAN results.

[Eric N Hanson @HansonOnData]

@narayanarjun @memsql @CockroachDB Ran't received :) You make a lot of good points -- we appreciate your feedback. To others following this, @adamprout addressed a bunch of your feedback with you on https://t.co/9Sr6si30i5

[Eric N Hanson @HansonOnData]

@narayanarjun @memsql @CockroachDB @AdamProut We pulled down the comparison to Cockroach due to the fact that we were not rate limiting per the TPC-C spec, so that comparison was apples to oranges. Our apologies. So really our numbers are only for showing that we scale well on C. Graph has been revised.

[Eric N Hanson @HansonOnData]

@narayanarjun @memsql @CockroachDB @AdamProut We may not be able to run with sync durability again soon at 50K warehouses because that requires access to a lot of HW on AWS.

[Eric N Hanson @HansonOnData]

@narayanarjun @memsql @CockroachDB @AdamProut I'm going to check with our devs who ran the benchmark and see if we can run it again with synchronous durability turned on and make those numbers available. As @adamprout said on HN, we were using sync replication so records are forced to a 2nd node before commit.

[Eric N Hanson @HansonOnData]

@frasergeorgew @narayanarjun @markcallaghan @awgupta @CockroachDB @memsql @jstnlvndski @tpcbenchmarks @aphyr @fivetran @frasergeorgew @fivetran We don't consider ourselves a pure DW but we do get used for some of these use cases, so yes, we absolutely are interested, please mail [email protected] and ask them to forward it to me. We'd love to participate.

[Eric N Hanson @HansonOnData]

Start out free. Know you'll never hit a scale wall. https://t.co/56trdFn9We

[Eric N Hanson @HansonOnData]

https://t.co/81oexPRdTc

[Eric N Hanson @HansonOnData]

https://t.co/dCrlGGsWv5

[Eric N Hanson @HansonOnData]

https://t.co/rAUGDLl4yl

[Eric N Hanson @HansonOnData]

https://t.co/6vAY5vCGes