Ok, will do.

It was my understanding too that Insert should be blisteringly fast... and I need it to improve some speed issue I have with existing stored procedures.

Fair point... so I tried to read back as well
Now, the Insert into the InMemory table is sometimes a faster but the read is a lot slower!

I...

Thanks 🙂

I am aware of issues related to Scalar UDFs and I don't think we have any in this procedure

What seems the best way to evaluate the cost of recompiles?

Extended events on that session (if I can reproduce) looking at waitstats?

We use a fair amount of temp tables with sometimes significant amount of rows (up to 100,000s) and I could spot a significant amount of Statement Recompile.

However, could...

If it was blocking on some table, then it would not affect the start time of the inner procedure.
I suppose I need to look at Wait Stats indeed...

I only noticed now that the statement firing the inner procedure started 15 seconds BEFORE the inner procedure itself started!

Why would that be???

My question to this forum is not about removing the performance bottleneck itself but more about understanding how two trace events, supposedly reporting the same thing report two drastically different...

Just in case it's useful to anyone...

I managed to load my file using the following BCP parameters
bcp MyTable in MyFile -S MyServer -c -T -t 0x01 -r...

Had never heard of the RAW option.

I'll give that a shot (on Monday!)

Thanks

I should add that I get error "Unexpected EOF encountered in BCP data-file"