Storage Performance In Entertainment And Content Creation

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
caedenv writes:
> ones didn't. In the end I found that it is all about a well balanced
> system, not about having 'the fastest' of any single part (unless you

Spot on, and at least for uncompressed video the old saying was it's all
about how much one can process, not how fast. Especially true for 2K,
4K and 8K film. Years ago it wasn't possible to get reliable high-speed
I/O on PCs, but that's changed, though consumer builds do of course lack
ECC protection which is a concern for pro users. Multi-socket XEON systems
tend to have this anyway so it's not a problem, but they're costly and
generally can't be overclocked. It's something I'm looking into atm, eg.
oc'd 990X or SB-E vs. dual-socket i7 XEON (main advantage of XEON systems
is the much higher RAM capacity - my Dell can take well over 100GB).


> lot of reading, and build around the specific software you plan on using
> because different ones use different technologies, and you don't want to

Top-end systems are often turnkey solutions supply ready to go, optimised
for the application, not really general purpose at all. Discreet systems
are like this.


> My own system is an i7 2600, 16GB of ram (need more but 8GB modules are
> too expensive lol), 3 single HDDs for system, scratch, and content drives

I have a range of systems, couple of Avid SGIs (O2, Octane2) and several
original Discreet systems (quad-1GHz Tezro with Flame/Smoke and aual-600
Octane2 with Flame, each with 24 x 146GB 15K SCSI; O2 Effect with 22 x
73GB 10K FC) all with HD/SD digvid options, breakout boxes, etc., a
couple of SGI for hardware-capture/playback (customised O2, Octane2 V12
with Cosmo2 board, both for reilable JPEG) and some other high-end
systems aswell (quad-500MHz Onyx2 IR4, 36-CPU Onyx3800 IR4, HD digvid,
72GB RAM, not really setup yet though). I have some Adobe Premiere
systems but don't really use them.

PC-wise I have several systems, some for benchmarking research: i7 875K
(overclock not yet sorted out), dual-XEON X5570 Dell T7500 with Quadro FX
5500 (might switch to a Quado 600 I've just obtained if it's better), an
Asrock X58 Extreme6 with another X5570 (should oc like crazy), and a 990X
not yet set up. A mix of FC, SCSI, SAS and SATA storage. All the other
systems are more for 3D/CPU benchmarking (two i7 870s, i3 550, i5 760
[though this will be away on loan for a while], i5 760, and a range of
AMD builds, mix of mbds, range of GPUs).

Combining SGI/PC hw in an optimised manner is my current focus, but I'm
also doing a lot of PC benchmarking.


> to use, so I went for it). For what I do (mostly 2-3 layers of video,
> with color correction and simple transitions in glorious 1080p) it is

I did test an SSD with Flame, no observable benefit for colour correction,
but the bottlenecks on a PC platform could be very different. Hard to say.

What you could do is run Process Explorer to see what I/O is occuring
while running your app.

Ian.

 
[citation][nom]palladin9479[/nom]I was more commenting on how the differences should be investigated. What exactly is QS doing differently then the "non QS" codec. Is there a level of settings that can have a software codec outputting the exact same data as the built-in Intel QS. I have no doubt that the Intel QS is going to be faster, on die instructions tend to be. I'm wanting to know why you got a ~700MB file with QS and a ~350MB without. If it was something on the order of 10~20MB then it's no big deal, but that big a size discrepancy warrants investigation.And again I have no issues with the bench's presented here as I highly doubt the above would make much different to a storage medium.[/citation]

Oh! Well, I believe for this specific story, I might have used a different bitrate setting for the QS encoding path. Honestly, I can't recall. The testing for this story was done back in November I believe, and we're kind of swamped trying to catchup with everything post-CES.
 
What about simultaneously recording and playing back 64 tracks of .wav files? Would like to see how it reacts in a big multi tracking environment .
 
Thought some here might find the following interesting:

http://www.sgidepot.co.uk/misc/LSI_SAS_3442E-R_3x_Seagate_600GB_15K7_hwRAID.gif

That's for 3 x Seagate 600GB 15K SAS, hw RAID on an LSI SAS3442E-R PCIe card (the array
just has 3 disk becasuse the 4th is being used for general data & backup in the system I built). The
3 disks cost me 225 UKP total, the SAS card cost 32 UKP.

This is just on a simple P55 system (Asrock P55 Deluxe) with an i5 760 @ 4GHz (loaned it to a
friend for use with After Effects).

Other data available on my site here:

http://www.sgidepot.co.uk/diskdata.html


Ian.

 
Umm, the fraps benchmark calls for 300+ MB/sec. Nope. That doesn't remotely pass a sanity check -- you ran this on a spinning drive that couldn't sustain that rate.

The actual data rate using your data panel numbers was 23 GB over 27 mins = about 14 MB/sec.
 
Status
Not open for further replies.