Interesting before and after results

[jimholcomb]

I ran SR 6.1 on a BarraCuda 1TB hard drive, and got these results. I suspect the before results are off but does anyone know what happened here?

(Problems uploading the screen shot initially).

Thanks.

 

[DanR]

The Before results suggest that the mid point and end areas of the drive had NEVER been written to. Hence they contained all zeros. The drive's firmware, recognizing this, apparently returned all zeros WITHOUT actually reading the drive. Hence the apparent SATA III max speeds.

The After results reflect that the firmware now knows that something has been written to the drive and must now read and return all zeros (in this case). The After results more reasonably reflect HDD performance, with the end being roughly half of the beginning, as it should be. The mid point is typicality roughly midway between the beginning and end speeds.

 

[jimholcomb]

The Before results suggest that the mid point and end areas of the drive had NEVER been written to. Hence they contained all zeros. The drive's firmware, recognizing this, apparently returned all zeros WITHOUT actually reading the drive. Hence the apparent SATA III max speeds.

The After results reflect that the firmware now knows that something has been written to the drive and must now read and return all zeros (in this case). The After results more reasonably reflect HDD performance, with the end being roughly half of the beginning, as it should be. The mid point is typicality roughly midway between the beginning and end speeds.

Thanks Dan!

 

[ShadowMeow]

That would be the translator layer.

 

[ColbyBouma]

Those results indicate that drive is probably SMR.

(Problems uploading the screen shot initially).

The file size limit on this forum is 500 kB. I REALLY wish it would say that somewhere instead of just saying "it broke".

 

[peterblaise]

Seagate Baracuda model number ... ?

So we can look it up.

 

[jimholcomb]

Seagate Baracuda model number ... ?

So we can look it up.

Seagate BarraCuda SMR 1TB

 

[peterblaise]

Seagate
Model Number ST1000LM049
Serial Number WM90W1B4
Firmware LXM4

From https://www.seagate.com/support/internal-hard-drives/laptop-hard-drives/barracuda-2-5/

2.5 Inch 1TB 7mm 128MB cache
5-year limited warranty

Data Sheet
https://www.seagate.com/content/dam.../pdfs/barracuda-2-5-DS1907-3-2005US-en_US.pdf

7,200 RPM, Data Transfer Rate up to 160 MB/s
- hahahahahaha!

Uses Seagate proprietary MTC Multi Tier
Caching technology
https://www.seagate.com/files/www-c...-tier-caching-technology-white-paper-2017.pdf
"... DRAM ... NAND ... magnetic storage medium ..."
DRAM is probably where it gets the 160 MB/s
potential.
But, yeah, it also uses the platter as cache, so
... wait for it ... it has to shuffle re-writing
whatever is in disk cache in order to write it
where it ultimately belongs - yeah, double
writing, just go away and get some coffee,
it'll be done soon ...

User Manual
https://www.seagate.com/content/dam...op-fam/barracuda_25/en-us/docs/100807728h.pdf

No firmware update as of 2025-04-22.

- - - - -

Yeah, first read performance was from most
areas being blank, it never actually looked at
the platter since it 'knows' there's no data
there, and responded to a read request
immediately with no delay.

The second read performance report is actual
performance when reading recorded data, as
expected.

It's a great ( inexpensive ) backup drive for
WORM Write Once Read Many static data,
inappropriate for a boot operating system or
RAID use.

I've never lost any data on SMR drives.

I've lost only inordinate amounts of time.

 

[jimholcomb]

It's a great ( inexpensive ) backup drive for
WORM Write Once Read Many static data,
inappropriate for a boot operating system or
RAID use.

I've never lost any data on SMR drives.

I've lost only inordinate amounts of time.

I have no idea where I got this drive, must have pulled it from some clunker. Interesting per your comments, I had written “laptop backup” on a sticker on it and apparently it was a drive I’d back my work laptop to. Left that job a few years ago so I formatted it and tested my new SR install on it.

Thanks to everyone for the info.

 

[peterblaise]

I wonder what happens when SpinRite reads
and writes 32,768 LBA logical block addresses
at a time - 16 MB - and the drive handles it
within 128 MB cache.

Forcing a re-read in between writes most
probably 'busts' any cache shenanigans.

Using free HDDScan, here's a data transfer
responsiveness graph of one of my Seagate
2.5" 2TB SMR drives under the Marshal name:

So I was getting peaks near 145 MB/s, yet
there's the example of peak at track 0 zero for
modern hard drives is often only met by 1/2
performance at track max at the end, about
62 MB/s - that's the effective benchmark
results from SpinRite 6.1 after fully rewriting
all data - peak at track 0, then 1/2 at the end.

We also see wear at the front, perhaps a full
SE secure erase would reset everything to
provide the full zig-zag step behavior.

Amazing technology.