Page 1 of 1

Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 10:26 am
by meerkt

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 10:58 am
by Igor_Kavinski
I prefer his Meltdown article. It didn't make my head spin.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 11:00 am
by just brew it!
I'd expect performance anomalies like that from drive-managed SMR because of the way the zones are laid out, but for a non-SMR drive that's pretty odd. I wonder if any of the remaining HDD vendors still do that; that Samsung is a pretty old drive (from roughly a decade ago).

Edit: Oh wait, he also does a Toshiba X300 (current drive). Interesting.

Edit 2: Damn. Just finished skimming the whole article. Some interesting stuff. Need to go back and read it more carefully now. I had no idea different platters can have different densities.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 11:01 am
by just brew it!
Igor_Kavinski wrote:
I prefer his Meltdown article. It didn't make my head spin.

If your heads are spinning the drive is toast. Only the platters should spin.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 11:32 am
by Igor_Kavinski
just brew it! wrote:
Igor_Kavinski wrote:
I prefer his Meltdown article. It didn't make my head spin.

If your heads are spinning the drive is toast. Only the platters should spin.


I don't know where my platters are.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 11:58 am
by Captain Ned
Igor_Kavinski wrote:
I don't know where my platters are.

Mine are 12" in diameter, rotate at 33 1/3 RPM, and have no random access capability. Latency for the innermost track can be as much as 24 minutes.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 12:20 pm
by Wirko
Captain Ned wrote:
Mine are 12" in diameter, rotate at 33 1/3 RPM, and have no random access capability. Latency for the innermost track can be as much as 24 minutes.

Oh they do have random access capability. It's perfectly possible that you never need it, though.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 12:50 pm
by Igor_Kavinski
Running this microbenchmark on SSD and SSHD would be "interesting".

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 3:49 pm
by just brew it!
Captain Ned wrote:
Igor_Kavinski wrote:
I don't know where my platters are.

Mine are 12" in diameter, rotate at 33 1/3 RPM, and have no random access capability. Latency for the innermost track can be as much as 24 minutes.

https://www.businessinsider.sg/this-nex ... -playlist/

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 4:11 pm
by Captain Ned
just brew it! wrote:
https://www.businessinsider.sg/this-next-generation-turntable-works-like-a-cd-player-you-can-skip-tracks-and-program-your-own-playlist/

No, no, bad, sick, and wrong.

1. Tangential trackers, while perfect in theory, never actually work as advertised.

2. Tonearm to cartridge mountings MUST be tight and without possibility of vibration. That bouncy metal strip will cause resonances that will kill the music.

3. "Using a proprietary spring design, it is able to auto- balance any cartridge that is attached to it." The last thing one wants in a tonearm is a spring.

4. "Ortofon 2M Red Moving Magnet Cartdrige". This is a throw-in for $399 entry-level turntables at all of the major audio e-tailers.

5. Dollars to doughnuts the "Aluminium Black Anodised Linear Tracking Sensor Arm" never works as promised.

6. It's a KickStarter.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 4:15 pm
by just brew it!
I'm pretty sure I've seen other attempts at this over the years. No idea how well any of them worked, but my point is, people keep trying. :wink:

Linear tracking is definitely one of those "better in theory, but then there's practice" sort of things. With a pivoted tonearm, you're never going to completely eliminate tracking angle error; the best you can do is a compromise to minimize the error at the outermost and innermost tracks.

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Sat Dec 07, 2019 4:24 pm
by Captain Ned
just brew it! wrote:
I'm pretty sure I've seen other attempts at this over the years. No idea how well any of them worked, but my point is, people keep trying. :wink:

You can even spend $220K on the pursuit.

https://www.paragonsns.com/product/clea ... tement-v2/

Re: Discovering HDD geometry thru µ-benchmarking

Posted: Mon Dec 09, 2019 3:36 am
by Wirko
Captain Ned wrote:

That's a quantum computer. These are quantum computers. As they can do anything, they can also play music.