Badrish Chandramouli

Benchmark (Release mode): `Embedded.perftest --op-workload PING --op-percent 100 -t 1` ### Framework: net8.0 ### Branch: main (on Kevin's fork) Running with 1 threads... Avg. throughput: 135719.46 Kops/sec; 135719.46 Kops/sec/thread (Max:...

> I reproduce this regression on my latest too, I'll spend some time digging (now that all feedback is addressed) and update when I have an explanation/fix. * Is it...

> Doesn't appear to be allocations based on local profiling. Might be useful to add `[MemoryDiagnoser]` to that benchmark, which I haven't had a chance to poke at yet. I...

> Mkay, because I can't leave well enough alone despite it being after 11 PM here... > > One relatively simple approach would be to slap a `int[]` (trading space...

> Still need to look at Badrish's proposal, so that's up next. Adding a link to this thread for reference - https://github.com/microsoft/garnet/commit/f20925a381c47050ac02ce0b75504852d70707a4

> exploratory commit that remembers CanAuthenticate I don't think this will work with authenticators that might time out, e.g., the Aad authenticator.

Idea to avoid the interface call to authenticator in common case (not necessary for this PR): https://github.com/microsoft/garnet/commit/17402f340806980a378682b807f1e4bb142268da

> Thinking out loud a bit, it seems to me a remaining perf killer is that we have to call into the `IGarnetAuthenticator` on every command to validate our `User`...

Instead of modifying `main/GarnetServer`, can we perhaps add a new `main/GarnetHost` (open to other namings) so that the existing straightforward server is not affected or have to carry additional library...

> Instead of modifying `main/GarnetServer`, can we perhaps add a new `main/GarnetHost` (open to other namings) so that the existing straightforward server is not affected or have to carry additional...