Is the response time of a HDD pr. file it reads and writes, or just the time it takes to respond to a command?
I am wondering because the response time of an SSD is close to non-existant, while a HDD has about 30ms. Would i notice a difference in the response time if a HDD was as fast as an SSD?
my two cents worth here, i would say response time is negligable(i.e time it takes to think about the command) its the seek time ...dam.. got me thinking now...where is speed fan..ok that didnt have what I thought it would.....further to it looking at your question i would have to say that you must consider the following, HDD and SSD's must have a portion that deals with the request right...this part tends to be a chip that opperates at a given frequency....the response time for an SSD would we drive controller processing ==> waiting time for the NAND switches to be read and when it finally gets that data that is the comparative of seek time for a HDD...a HDD would be controller chip===> (angular velocity and head movement) seek time....I'd say that the controllers in the SSD/HDD processing the request is soooo tiny its diff to measure..
also from my understanding they are good at different things too," would I notice the difference in response time if a HDD was as fast as an SSD?"...no you wouldnt because if the were as fast as each other or as slow as each other, then which would be faster..lol..sorry re-phrase the q there i think?
"Response time" on a HDD is virtually all "Latency", which is the term for a few stages of finding the required information on the drive. Look at the overall sequence of operations at the hard drive when you request that it feed your RAM with a file from a known location of the disk.
1. OS uses the CPU and mobo chips to send to the HDD unit's onboard controller board a read request for the data located at a specified location, and continuing on for a certain length of data stream. The location is in the form of a digital Logical Block Address.
2. HDD's controller translates the LBA into its own internal co-ordinate system of Cylinders, Tracks, and Sectors. It sends signals to the head drive to position the head arms over the correct cylinder, then sets up to retrieve the data from the correct head and transfer it to its own buffer RAM. In fact, it will set up to read a whole bunch of Sectors in sequence and copy them all into buffer RAM, then send only the required parts of that back to the OS.
3. Wait. Until the head is positioned over the right place and stabilized. Now, start looking at the data coming back from the head, watching for the markers for the correct sector number to fly past the head.
4. Wait. Until the spinning disk brings the correct starting sector past the head, so the copying from heads to buffers can start.
5. Copy the data as fast as it flys past the head.
6. Select the correct data from the buffer RAM and send it back to the mobo, via the chipset and CPU, to be placed in main RAM for access.
7. Was that all that was required? Or, is there more data to be gathered for the original request? If more is needed, where is it to be found? If it's still on this same cylinder, you don't have to move the heads; if it does require a head movement, back to step 2. etc.
Now in all of this, BY FAR the longest processes are the mechanical operations of head positioning and platter rotation. The sum of those two waiting times is what generates almost all of the quoted average 30 ms Latency time. Virtually all the other steps are solid state electronics operations that happen in microseconds or less. In a SSD, none of that mechanical process happens, so virtually all delays in those devices are really electronic circuit response and settling times.
In essence, your question implies an impossible dream. Because the processes are fundamentally different, "if a HDD was as fast as an SSD" cannot happen ever.