Does it make a difference what kind of external hard drive I use if it spins at 7200 RPM? USB vs Thunderbolt make a dif here?

[Kevin Rag]

Thunderbolt drives do make a big difference. RAIDs make a big difference over a single drive too. Go for a Thunderbolt RAID. RAID 5 for both speed and redundancy.

Kannan Raghavan
The Big Toad Films Pte. Ltd.

[Noam Osband]

I thought the advantage of RAID was simply size. If everything I have is not that big (120 GB), is it fast on a RAID as opposed to an external drive alone?

[Claude Lyneis]

If you use your RAID disk as RAID 0, it is writing back and forth on two drives, it is much faster. No redundancy, but fast.

[Joe Marler]

[Noam Osband] “…render these films with a lot of effects on them. It takes a while….does it make a difference what kind of drive they are on for render times. I imagine the big thing is the RPM of the drive and speed of the computer processor. For rendering, does it make a difference if I’m using Thunderbolt 2.0 or USB 3.0?…

Technically, “rendering” means resolving all effects and edit directives in the timeline, not exporting to an output file. Exporting means writing a rendered, encoded timeline to an output file. If the timeline is not rendered before export, then the export phase will do that. Also it may require encoding based on the selected output codec. However we often loosely use the term “rendering” to mean exporting to a file, but two different logical phases are involved.

Normally rendering is either CPU or GPU bound, not I/O bound. You can examine that using various monitoring tools such as Activity Monitor or iStat Menus — the I/O load is generally not very high. Likewise exporting or encoding is usually CPU bound (if to H264 or similar long GOP formats).

A chain is only as strong as the weakest link, which for rendering or encoding is usually the CPU or GPU. So it usually does little good to further improve I/O for this phase. On a Mac, I/O queue depth can be examined using the command line Dtrace tool iopending. Always exercise caution when using any terminal commands.

Note the Dtrace utilities cannot be used starting with El Capitan unless System Integrity Protection is first disabled: https://apple.stackexchange.com/questions/208762/now-that-el-capitan-is-rootless-is-there-any-way-to-get-dtrace-working

If you are editing or exporting to ProRes or other low-compression codec, and especially if multicam, then I/O performance can be more limiting. However in your case I doubt whether the interface is Thunderbolt or USB 3.0 will make much difference for a single 7200 rpm drive during the rendering phase.

In a large library, FCPX does a lot of 8k random I/O for thumbnail and plist generation and maintenance, so in this case random I/O performance can be important. Unfortunately spinning RAID arrays are not very good at small random I/Os.

[Keith Mullin]

Thunderbolt and USB 3.0 will not really have any noticeable difference in performance for a single spinning hard drive, as both protocols are significantly faster than a spinning drive can read/write.

[Robin S. kurz]

[Keith Mullin] “Thunderbolt and USB 3.0 will not really have any noticeable difference in performance for a single spinning hard drive”

Exactly. If we’re talking single HD drive, then whether it’s USB3 or TBo is completely irrelevant. SSD of course being a whole different ball ‘o wax. But even a HD RAID under at least 4 disks (depending on the level even 5) won’t saturate either bus. Thinking that any one HD is faster just because it’s on TBo as opposed to USB3 is unfortunately a common fallacy.

And rpms are virtually irrelevant also with today’s disks. A few hundred rpms more or less will make zero perceptual difference in speed.

– RK

____________________________________________________
Deutsch? Hier gibt es ein umfassendes FCP X Training für dich!