I said two separate things: a) That Safari's performance hadn't been tuned for when caching was disabled. b) That, in general, it's more realistic to test a m

I said two separate things: a) That Safari's performance hadn't been tuned for when caching was disabled. b) That, in general, it's more realistic to test a mixed cached/uncached workload. Which bit was incorrect?