Michael,
it's an interesting hypothesis. Maybe you can describe the following
results. Recently I made a test program that just reads a file sequentially
with different block sizes and results are still confusing. This was the
same SSD drive and the functions were basically from api:
Those numbers make sense to me.
Since count is doing a linear walk throughcorrect?
#1 SSDs more than likely don't do "read ahead" like a disk drive. So what
you're seeing is what read-ahead does for you.
#2 Count is doing a linear scan of the table...Probability of 2K containing the
next
2 matches
Mail list logo
