topic Re: X25-E Concurrent Writes in Archive

X25-E Concurrent Writes

idata — Mon, 21 Feb 2011 23:30:57 GMT

Hi,

We have a short questions about the internal concurreny of writes:

It is said that Native Command Queuing enables up to 32 concurrent operations, but how many (internal) operations can actually write concurrently.

In a normal HD I have x Read/Write Heads so (in theory) I could perform x concurrent operations on a physical level.

What is the theorectical and physical equivalent to this in the intel x25 e SSD

Thanks

Re: X25-E Concurrent Writes

idata — Tue, 22 Feb 2011 07:44:13 GMT

The X25-E controller has 10 channels to 10 independent NAND chips so I would assume 10.

Re: X25-E Concurrent Writes

idata — Tue, 22 Feb 2011 09:02:36 GMT

Would it not depend on queue depth? Performance scaling stops after QD32.

Re: X25-E Concurrent Writes

idata — Tue, 22 Feb 2011 21:02:52 GMT

If we are talking about concurrent as in how many operations can be performed in a clock cycle... I would guess 10 possible operations possible. If using toggle or double data rate NAND, it would be 20 possible operations in a clock cycle. This is my guessing based on what is public about the drive and assuming the clock is synchronous.

I guess it depends on how the question is framed.... Like how many IOPs do SSDs achieve in sequential reads at different QDs? That would be a different answer...