[clue-talk] Benefits of SSDs : was new processors

[chris fedde]

after slightly rewriting the command line I get roughly the same results:
On the ssd:

sync; time (dd if=/dev/zero of=/tmp/foo bs=4k count=100000; sync)
100000+0 records in
100000+0 records out
409600000 bytes (410 MB) copied, 4.68162 s, 87.5 MB/s

real    0m6.013s
user    0m0.024s
sys     0m2.516s

and on the sata II 7200rmp

sync; time (dd if=/dev/zero of=foo bs=4k count=100000; sync)
100000+0 records in
100000+0 records out
409600000 bytes (410 MB) copied, 8.10456 s, 50.5 MB/s

real    0m10.611s
user    0m0.052s
sys     0m2.580s

On Mon, Jan 4, 2010 at 4:56 PM, David Ahern <dsahern at gmail.com> wrote:
> What do you get for each case with the following:
>
> sync
> time -p 'dd if=/dev/zero of=/tmp/foo bs=4k count=100000; sync'
>
> David
>
>
> On 01/04/2010 04:49 PM, chris fedde wrote:
>> On my sata attached ssd boot disk I see sequential write rates of:
>>
>> dd if=/dev/zero of=/tmp/foo bs=4k count=100000
>> 100000+0 records in
>> 100000+0 records out
>> 409600000 bytes (410 MB) copied, 4.76564 s, 85.9 MB/s
>>
>> and read rates are:
>>
>> dd of=/dev/null if=/tmp/foo bs=4k count=100000
>> 100000+0 records in
>> 100000+0 records out
>> 409600000 bytes (410 MB) copied, 4.57398 s, 89.6 MB/s
>>
>> On my sata II based data disk I see write rates:
>>
>> dd if=/dev/zero of=foo bs=4k count=100000
>> 100000+0 records in
>> 100000+0 records out
>> 409600000 bytes (410 MB) copied, 8.84794 s, 46.3 MB/s
>>
>> and read rates:
>>
>> dd of=/dev/null if=foo bs=4k count=100000
>> 100000+0 records in
>> 100000+0 records out
>> 409600000 bytes (410 MB) copied, 7.73974 s, 52.9 MB/s
>>
>> So the ssd is almost twice as fast as the 7200rpm sata II
>> Doing this test with count <= 10000 was much faster since the whole
>> file could sit in io buffers.
>>
>>
>> On Mon, Jan 4, 2010 at 2:46 PM, YES NOPE9 <yes at nope9.com> wrote:
>>>>
>>>> On Jan 4, 2010, at 12:13 PM, Nate Duehr wrote:
>>>>
>>>> Quick note: Sorry, top-posted because I'm on a webmail interface right
>>>> now...
>>>>
>>>> The speed difference between disk and everything else
>>>> (processor/RAM/video/etc) is being closed with SSD drives.  Still not
>>>> "100% ready for prime time" yet, looking over some of the errata on the
>>>> early drives, but... with a good brand name (meaning: Intel right now),
>>>> I'd trust one.
>>>>
>>>> (Some of the others have had some lovely bugs in firmware.  Kinda makes
>>>> sense, Intel's a chip manufacturer and the hard drive manufacturers are
>>>> learning how to do that, in a way.)
>>>>
>>>> Having seen a couple of machines with SSD's in them doing various tasks,
>>>> I can unequivocally state that the spinning hard disk platter for
>>>> anything other than mass storage, is dead.  It just doesn't know it yet.
>>>>
>>>> My next personal laptop will have an SSD for the OS's and swap, for
>>>> sure... and the data for whatever e-mail and PIM type client software
>>>> I'm using will also go on it to speed those common applications up.
>>>>
>>>> It's too big of a performance increase to ignore, from what I've seen.
>>>>
>>>> --
>>>>  Nate Duehr
>>>>  nate at natetech.com
>>>
>>>
>>> What kind of performance gains have you seen ?
>>> If you provide anecdotal stories  that is fine .
>>> What kind of data rates and latency do you get out of SSDs ?
>>> Gus
>>> _______________________________________________
>>> clue-talk mailing list
>>> clue-talk at cluedenver.org
>>> http://www.cluedenver.org/mailman/listinfo/clue-talk
>>>
>> _______________________________________________
>> clue-talk mailing list
>> clue-talk at cluedenver.org
>> http://www.cluedenver.org/mailman/listinfo/clue-talk
>