Most databases are going to be bottlenecked on CPU well before they handle ~50GB/s. And it's gotten considerably more likely with the spectre/meltdown mitigations. Especially OLTP is pretty syscall heavy.
-
-
Show this thread
-
A lot of this perception was formed when there were no SSDs, and memory was so scarce that it'd never fit a meaningful percentage of ones workload. But especially the latter has been wrong for a looong time. A *lot* of databases are in ~100MB - ~10GB range.
Show this thread
End of conversation
New conversation -
-
-
Even more CPU used if DBMS does checksum and/or decompress after each storage read
-
True. It's impressive though what kind of decompression bandwidth e.g. lz4 can have. It's feasible to reach >2GB/s even for some database uses (where typically the number of blocks decompressed at once is small-ish).
- 8 more replies
New conversation -
-
-
Who has NVMe disks for their production workloads? In the cloud few, as they are mostly ephemeral. And since the cloud has an increasing market share, my answer to your Q is "definitely not soon". It could actually be increasing in the opposite direction with faster CPUs.
-
Even if you don't, you have access to high bandwidth, very parallel, but higher latency storage.
- 3 more replies
New conversation -
-
-
Who has this perception? Postgres often crawls hundreds of kb/s for me with 100% cpu load, toasting, detoasting and especially calling gist penalty in an almost-O(N^2) loop when trying to build it.
-
Most databases don't run postgis. I can saturate 6GB/s on my laptop with today's postgres for some workloads. We really should throw out pglz and replace it with lz4 by default. But I'm currently working on other 'really should's.
- 2 more replies
New conversation -
-
-
I read that and agreed. But then there are the systems in "enterprise installs" that are installed in a VMware machine with insufficient ram sitting on top of a massively overbooked SAN. And they are most definitely limited by IO and in my experience are in a big majority.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Everything is a CPU problem
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.