Hi George,

Thanks for your response.  For now, all I really want is a simple way to tell 
if I'm getting SATA III access to my internal hard drives or still SATA II (and 
if I'm getting USB-3 rather than USB-2).  It doesn't have to be perfect.

Gregg

On Dec 15, 2013, at 2:00 PM, George N. White III <[email protected]> wrote:

> Different workloads can have very different characteristics which makes I/O 
> benchmarks tricky.  At one extreme you can have lots of small I/O requests 
> updating a data base where each update writes a small amount of data but also 
> requires updates to tables, indices, etc.   Numerical simulations can write 
> large amounts of data as fast as the disks can take it, but will wait for the 
> I/O to finish before making new requests.   Video capture (which these days 
> may involve multiple "cameras" and high resolutions) will drop data if the 
> disks can't keep up.
> 
> The most useful comparison would be based on whatever elements of your real 
> workload are currently a bottleneck.   When you make major changes to your
> mass storage, there may be opportunities to improve the workflow, e.g., 
> running
> multiple tasks in parallel, so it can be helpful to have some benchmarks for 
> several variations on the old workflow.  At work, moving from SGI 
> multiprocessor systems to Mac Pro, we found that SGI I/O could handle only 2 
> of our standard workload tasks, which (because we didn't want to slow down 
> interactive users) meant we were limited to one batch process at a time.  The 
> Mac Pros start to bog down at around 8 tasks (using 3 disks in a stripped 
> configuration).
>   
> 
> On Sun, Dec 15, 2013 at 1:46 PM, Dinse, Gregg (NIH/NIEHS) [V] 
> <[email protected]> wrote:
> Hi,
> 
> Can someone please recommend some (free?) software that would allow me to 
> assess read and write speeds for various disks (e.g., hard drives, SSDs, 
> flash drives)?
> 
> Alternatively, are there some simple commands that I can give in the Terminal 
> app?  At one time, I believe someone gave me such a command, but 
> unfortunately I have long ago forgotten it.  I think the "write" part of the 
> test involved writing a large amount of stuff to /dev/null.
> 
> In case anyone is interested in the background, I am hoping to use a pair of 
> PCIe cards to upgrade a mid-2012 Mac Pro by adding some external USB-3 ports 
> and some internal SATA III ports.  I wanted to assess "before and after" 
> speeds.
> 
> Thanks,
> 
> Gregg
> 
> -- 
> George N. White III <[email protected]>
> Head of St. Margarets Bay, Nova Scotia

_______________________________________________
MacOSX-talk mailing list
[email protected]
http://www.omnigroup.com/mailman/listinfo/macosx-talk

Reply via email to