Solid State Drives better?

[msteffke]

I am spec'ing some workstations, Dell T3500. I have decided to go with SSD for the first time. My IT guy told me that they may not be as fast as a SATA 10K rpm drive because they run thru an emulation.
Has anyone tried these? Are they seeing good results?

Thanks a bunch!

[bgruman]

Mark,

Your IT person is correct. With my testing of 10K drives vs SSD's using
the OCUS benchmark, there is no difference in Pro/E performance. Because
10K drives are less expensive, I typically go with that option when
building a custom GC machine. However, another factor to consider is
availability. The Monsoon rains in Haiti (this last summer) have damaged
the factory where Western Digital makes their Velociraptor drives.
Therefore there is currently a worldwide shortage of 10K drives. (This is
also driving the price up on 10K drives...) As a Dell VAR, the last time
I tried to order a machine with a 10K drive I was told that there was a 6
WEEK waiting period.

If you've got the time to wait, I'd still say go with a 10K VR Drive. If
you don't mind spending a little extra money on SSD's you can get the
machine sooner and still have very good Pro/E performance.
Side note: DO NOT make the mistake of buying a standard 7200RPM drive. It
is tempting because they are cheep, but your Pro/E Performance will suffer
because of it.

Good Luck
Bernie

Bernie Gruman
Owner / Designer / Builder
www.GrumanCreations.com


On Fri, Dec 9, 2011 at 5:43 AM, Steffke, Mark J
<mark.steffke@manitowoc.com>wrote:

> I am spec'ing some workstations, Dell T3500. I have decided to go with
> SSD for the first time. My IT guy told me that they may not be as fast as
> a SATA 10K rpm drive because they run thru an emulation.
> Has anyone tried these? Are they seeing good results?
>
> Thanks a bunch!
>
>
>

[rrich]

Your IT guy is just un educated in this realm. Do the research you will
find SSD are much better than even 10K SATA drives. The closest race I had
ever seen was when a raid array of 4 striped 10K drives with huge
cache32Gig(basically a SSD) were pitted against one SSD but the SSD still
won out. The SSD was the PCI plug in type not the SATA type.



Here is some reading for you.


pick, runs circles around mechanical storage"



Regards,

Ron Rich

[bgruman]

Mark, All,

Sorry, I just re-read my previous e-mail and realized that I was not very
clear to the original question... (its still early... I need coffee!)

SSD's are not slower because of emulation. Instead they perform on par
with 10K drives using the OCUS Benchmark test. SSD's are faster with other
windows/office programs, but they are not faster or slower with Pro/E.

Sorry for any confusion.
Bernie

[rrich]

Bernie is correct you will not see much in the difference of running proe
other than it will boot faster. Once proe is running it is all in computer
memory and has very little hard drive involvement, until you reach your
Memory limit then it will do swaping. Your benefits will come from all
the other things you do on your computer. Firing up programs storing and
retrieving information and other mundane everyday tasks.



Remember back in the day when we though multicore or multiple processors did
not give you an advantage. True it didn't when running pro, because it used
only one processor, BUT remember how you were then able to open word, check
email, run a spreadsheet, or whatever while that was up. That is the kind
of performance gains that you will realize with a SSD.



Regards,

Ron Rich

[bgruman]

Ron, Great analogy. Spot on!

[DanielReid]

I'd like to add that there are some operations that most certainly do
read from disk "live", and for those disk read time does impact
performance. We found this because we have maintained network share
installs instead of local installs, and at some point our servers were
moved a bit farther from us (increasing latency). This was still
bearable in WF2. It is annoying in WF5, especially if you get a taste
of local performance.



For example, editing a table cell in a drawing took a very long time to
open (10-15s) the dialog on the network install. With a local install,
it opens in less than a second. In WF2, this was only a problem the
first time but in WF5 it was every time. That bit of code for that
dialog is read from disk each time (ridiculous). Other bits of code are
similarly read from disk when needed, or at least that is the only
explanation I have from the evidence of network traffic and disk
activity.



Like other "real world" activities, I doubt the benchmarks capture all
of these types of events. The users certainly do though. That said, we
see very little real difference in this respect between our 7200 drives
(IT swapped out from our originally spec'd and configuration tested 10K
drives without telling us) and our SSDs. That table edit goes from just
under a second to even less than that -- but at that point it doesn't
matter. However, launching WF5 and pulling objects into session from a
WS are significantly faster with SSD. With our large truck assemblies,
this means a lot.

[JOES]

Sorry for jumping in here so late on SSDs, but here is my take. I've got
two SSDs in RAID0 & yes, they are fast. Intralink 3.4 opens very quickly,
much faster than on my work machine (7200RPM). Assemblies open faster as
well. Another factor is my home machine is 64bit, work machine is not.

If I can ever get the time, I will benchmark my SSD setup but I wouldn't
trade it for any old fashioned spinner.

Regards,
Joe S.