Home Forums Peripherals & Multimedia Peripherals how noisy are external hard drives?

Enjoy your new iMac and Yosemite. While you are waiting for it to be delivered it would be an opportune time to be sure all and I mean ALL of your applications, kexts, etc. are up to date and make a list of those you will need to update after you are running Yosemite. The most common glitch with Yosemite is out of date software.

I have been doing this for a few weeks. I have a list of updates that can be downloaded when I get Yosemite and I just ordered new calibration software from Xrite (the old software is PPC only). Most of my apps will work and, luckily, I rarely (if ever) use a lot of the ones that won't. I will be trashing the useless apps before migrating. There's no sense wasting time and disk space on them.

I just ordered a refurb iMac 5K 27" Retina display, 3.5 GHz, 1 TB fusion drive, 8 GB RAM. The price was very attractive ($2119.) so I pounced. I bought an extra 8 GB RAM from OWC so that will bump it up to 16 GB. For everyone's information, the salesperson told me that migrating data would be faster using FW - Thunderbolt (Target Disk Mode) than via Ethernet. This may be useful for someone else.

Here is an interesting additional consideration on Thunderbolt vs. USB 3, as posted by a reader on MacInTouch:

"Feb. 24, 2015
Thunderbolt

Michael Skinner
Dirty secret on Thunderbolt: When setting up a new email server (Kerio Connect), we spec'd a Drobo Mini with four 512GB SSDs installed for the mail database. Had the option of connecting the Drobo via USB 3.0 or Thunderbolt. Decided to test both connections, and discovered something unexpected. USB was ever so slightly slower than Thunderbolt (given these SATA SSDs that could not leverage Thunderbolt's full potential), but CPU usage was much higher using Thunderbolt than USB 3.0. When copying a 7GB file, Thunderbolt was a few seconds faster, but idle CPU went down to 83%. Copying the same file over USB, idle CPU stayed above 96%. This was wholly unexpected, because it is the opposite of conventional wisdom regarding FireWire vs. USB. Who's have thought Thunderbolt would be such a CPU hog? Based on this, I rolled out the server with the Drobo attached to USB. I have been happy with the decision, the server rarely peaks 10% CPU even during intensive operations.
[I wonder if UASP (USB Attached SCSI) was at play here? -Ric Ford]"

Very interesting, Ira. When I get the whole system set up, I'll play with USB 3 vs. TB and see what happens to CPU. Good to know...

that IS interesting.... firewire and thunderbolt are supposed to be DMA (direct memory accesss) - basically the computer and the peripheral negotiate a memory region for a buffer, and the peripheral fills the buffer directly, and then the computer does whatever with it (write it out to another disk etc) This is a bit of a security headache, as a peripheral with unrestricted DMA can snoop or write anywhere in memory. (and i've had a firewire peripheral hard trash my internal hdd on one occasion)

Thunderbolt was explained to me as requiring a number of special io chips in each end of the cable (which was to explain the cost of the cable) and that was supposed to remove the workload from the computer, normally present with USB, which requires the CPU to read in data and fill the buffer manually.

Keep us informed, I'd like to know. I'll do a little testing here, I have a lacie that does TB and USB3, with an SSD internal, and will test it with copying to/from a new retina's SSD.