* Use the same result language, output format, and collection method * Submit commands synchronously as fast as you can; this should give a reasonable indication of the synchronous performance of the uncontended write path. It's not perfect, but it's sure as heck better than nothing. * Documentation
Use the same result language, output format, and collection method
Submit commands synchronously as fast as you can; this should give a reasonable indication of the synchronous performance of the uncontended write path. It's not perfect, but it's sure as heck better than nothing.
Feedback from code review: this should be integrated with the existing benchmark tool, not something separate. We can have a new flag for it, something like "--synthetic-facts = medium" to enter purely synthetic fact mode instead of using the existing seeded simulation mode.