Samsung F3 1TB & 2TB HD Tune graph & CrystalDiskMark results

vbimport

#1

I recently bought these two hard disks, the 2TB to backup my internal WD RE-4GP 2TB drive and the 1TB as a fast drive for video editing.

Samsung F3 1TB 7200RPM 32MB Cache - HD Tune Pro 4 graph

Samsung F3 1TB 7200RPM 32MB Cache - CrystalDiskMark 2.2 results

Samsung F3 2TB 5400RPM 32MB Cache - HD Tune Pro 4 graph

Samsung F3 2TB 5400RPM 32MB Cache - CrystalDiskMark 2.2 results

Benchmarks of the Samsung F3 500GB carried out by Dee can be viewed here. It’s even a bit quicker than the 1TB! :smiley:


#2

Those are impressive spinners. :iagree:
118MB/s from a 5400RPM drive is pretty impressive as well.


#3

Well i dont like the access time, 15 ms and 15.8ms for the 5400rpm disk? I was expecting more from the 7200 rpm drive.

Here is my old and everyday abused samsung spingpoint F1 and access time is much better 13.7ms (6800 hours on according to crystal disk info)



#4

I like the negative CPU usage :bigsmile:
Add some more HDDs and you don’t need to overclock :smiley:


#5

I would doubt this program’s accuracy since it measures HDD performance when Windows performs reads and writes to the HDD in the background…
//Danne


#6

I’m surprised about the poor average seek time for the 1TB drive, as I remember 12ms to 13ms typical for my previous 7200RPM hard disks. The above HD Tune tests were also carried with the drives offline (set to Offline in disk management), however, the CrystalDiskMark results had to performed with the drives live as it creates temporary files on an existing partition rather than direct disk access.

As an OS drive, I would use just the first 60GB or 80GB as the OS partition to reduce seek time and create a 2nd partition for data that doesn’t require fast access times, such as photos, music, video, etc…

The following shows the average access times within the first 60GB of both drives:



#7

Yes but you will use the extra space, so you will end up again with the 15ms access time, that is the big benefit you get from an SSD drive, almost 0 access time (plus the zero mechanical parts).


#8

[QUOTE=vroom;2499305]Yes but you will use the extra space, so you will end up again with the 15ms access time, that is the big benefit you get from an SSD drive, almost 0 access time (plus the zero mechanical parts).[/QUOTE]

And not all flash drives have a near-zero access time. One of my flash memory cards has a whopping 22ms access time!


#9

A flash card is not the same as an SSD though.
If an SSD has 22ms access time then something is very wrong with it.
Typically, SSD has an access time of between 0.1ms and 0.3ms.
Regarding these F3’s. Access time on my 500 F3 is around 13ms, thats not a problem as it’s used for storing pictures, audio and video. The larger the file, the less important access time becomes.
As an OS drive, 13ms access time wouldn’t be so good. :slight_smile:


#10

[QUOTE=Dee;2499353]Regarding these F3’s. Access time on my 500 F3 is around 13ms, thats not a problem as it’s used for storing pictures, audio and video. The larger the file, the less important access time becomes.[/QUOTE]

It’s also possible that the 1TB and larger F3s have the acoustic seek management feature enabled by default. This muffles the seek noise but increases the access time.


#11

I don’t like the blue curve - fall is too steep meaning this drive, in the outer zone, is performing at only half speed coming close the the performance of a DVD drive.

This is unacceptable especially for a TB drive


#12

[QUOTE=Millennium12;2499422]I don’t like the blue curve - fall is too steep meaning this drive, in the outer zone, is performing at only half speed coming close the the performance of a DVD drive.

This is unacceptable especially for a TB drive[/QUOTE]I assume you mean the inner zone?
At 73MB/s at the inner zone, the F3 is still more than 10MB/s faster than a 300GB VRaptor, and more or less 3x faster than the max transfer rate of a DVD drive.


#13

[QUOTE=RJL65;2499418]It’s also possible that the 1TB and larger F3s have the acoustic seek management feature enabled by default. This muffles the seek noise but increases the access time.[/QUOTE]
You seem to have found the culprit :iagree: I checked the acoustic seek management and it was set at “Disabled”. The choices were Disabled, Fast, Middle and Quiet, so I set it to “Fast” and got the following seek result, which is about the average seek rate vroom got:

[QUOTE=Millennium12;2499422]I don’t like the blue curve - fall is too steep meaning this drive, in the outer zone, is performing at only half speed coming close the the performance of a DVD drive.

This is unacceptable especially for a TB drive[/QUOTE]
As Dee mentioned, the very end of the 1TB F3 curve was 73.5MB/s, which is faster than what many HDDs read on average. Note that the axis on the right is for the seek-times, not MB/s. The transfer rate axis is shown up the left of the graph.

I’ve also yet to see [U]any[/U] other 1TB+ hard disk finish off at above 70MB/s! The 1TB F3 drive also has an average of 119.4MB/s, which is quicker than what most HDDs start at.



#14

[QUOTE=Dee;2499428]I assume you mean the inner zone?
At 73MB/s at the inner zone, the F3 is still more than 10MB/s faster than a 300GB VRaptor, and more or less 3x faster than the max transfer rate of a DVD drive.[/QUOTE]

looking here http://en.wikipedia.org/wiki/Hard_disk_platter

sector 0 is next to the spindle (inner zone) with highest transfer rate while the last sector is at the outer end of the platter with the lowest transfer due to a longer seek timing. The graph clearly shows highest transfer rate at 0 GB. Therefore 73MB position is at the outer end of the platter (1GB position), not inner.

you think me wrong?

btw, my DVD has a burst rate of 75 MB/s


#15

[QUOTE=Millennium12;2499437]looking here http://en.wikipedia.org/wiki/Hard_disk_platter

sector 0 is next to the spindle (inner zone) with highest transfer rate while the last sector is at the outer end of the platter with the lowest transfer due to a longer seek timing. The graph clearly shows highest transfer rate at 0 GB. Therefore 73MB position is at the outer end of the platter (1GB position), not inner.

you think me wrong?

btw, my DVD has a burst rate of 75 MB/s[/QUOTE]It’s simple physics. :slight_smile:
The data density is equal per sector throughout the platter.
In the case of the Samsung F3, its rotational spindle speed is 7200RPM. Regardless if this is the inner or outer region of the platter, the platter rotates 7200 times every minute.
The read/write head does not rotate. However, the velocity of the platter in relation to the read/write head increases, the further out into the platter you go.

You can easy check this yourself.
Take an old DVD media.
Using a pen, mark dots with the pen approx 1cm apart right around the inner edge of the recording area, and lets call each dot a sector. you’ll end up something like 14 dots.
Now do the same at the outer edge of the disc, you’ll end up with around 45 dots, each one representing a sector.
Now draw a straight line from the inner edge to the outer edge.
Now rotate the disc slowly for one complete rotation.
For one rotation, the inner edge will read 14 sectors, while the outer edge will have read 45 sectors.
So it’s simple physics, the outer edge of the platter has a higher velocity, so data transfer at the outer edge is faster than at the inner edge.

BTW: Burst rate is only a measurement of the interface, not what the drive is physically able to read or write at.


#16

[QUOTE=Dee;2499471]… it’s simple physics, the outer edge of the platter has a higher velocity, so data transfer at the outer edge is faster than at the inner edge…[/QUOTE]

while you are physically correct the transfer rate is slower on the outer edge. Why? coz the heads have to travel [U]a longer way[/U] to read a certain amount of data compared to the inner zone - this takes more time!

you can trust me on that - I repaired hdds on bit level for 10 years…okay it’s a long time ago but hdd physics did not change much since then except everything go smaller :flower:


#17

[QUOTE=Millennium12;2499474]while you are physically correct the transfer rate is slower on the outer edge. Why? coz the heads have to travel [U]a longer way[/U] to read a certain amount of data compared to the inner zone - this takes more time!

you can trust me on that - I repaired hdds on bit level for 10 years…okay it’s a long time ago but hdd physics did not change much since then except everything go smaller :flower:[/QUOTE]Quite simply, you are wrong. :flower:
While seek times are higher on the outer edge of the platter, once the head is there transfer rate is higher and access times are lower, because there are more sectors to read or write without having to move the head a large distance. That is seen in the graphs of these tests, the beginning of the tests is the outer edge of the platter.

This time i’ll quote from wiki (see link in my previous post).

Data transfer rate
As of 2008, a typical 7200rpm desktop hard drive has a sustained “disk-to-buffer” data transfer rate of about 70 megabytes per second.[44] [B]This rate depends on the track location, so it will be highest for data on the outer tracks (where there are more data sectors) and lower toward the inner tracks (where there are fewer data sectors); [/B]and is generally somewhat higher for 10,000rpm drives. A current widely-used standard for the “buffer-to-computer” interface is 3.0 Gbit/s SATA, which can send about 300 megabyte/s from the buffer to the computer, and thus is still comfortably ahead of today’s disk-to-buffer transfer rates. Data transfer rate (read/write) can be measured by writing a large file to disk using special file generator tools, then reading back the file. Transfer rate can be influenced by file system fragmentation and the layout of the files.


#18

[QUOTE=Millennium12;2499474]while you are physically correct the transfer rate is slower on the outer edge. Why? coz the heads have to travel [U]a longer way[/U] to read a certain amount of data compared to the inner zone - this takes more time![/QUOTE]

While what you claimed was at one point true during the really early years of hard drives (whose platters had extremely low areal density by today’s standards) when the largest hard drive that was available held only 20MB worth of data (and even this capacity required a whopping 10 platters to achive this capacity point at the time, resulting in a desnity of only 2MB per double-sided platter), today’s hard drives actually start reading from the outer edge of the platters rather than the inner tracks. Why? Because if the heads remain on the data portions of the platter(s) all of the time that they are physically spinning, they become far more prone to severe and permanent damage. Plus, the heads in modern hard drives are parked on the innermost track only when the drive is completely powered down; once they are powered on, the heads actually relocate to the outermost zone. Finally, today’s heads sit so physically close to the platters that there is barely enough air space in between surfaces.


#19

[QUOTE=Millennium12;2499422]I don’t like the blue curve - fall is too steep meaning this drive, in the outer zone, is performing at only half speed coming close the the performance of a DVD drive.[/QUOTE]

[QUOTE=Millennium12;2499437]looking here http://en.wikipedia.org/wiki/Hard_disk_platter

sector 0 is next to the spindle (inner zone) with highest transfer rate while the last sector is at the outer end of the platter with the lowest transfer due to a longer seek timing. The graph clearly shows highest transfer rate at 0 GB. Therefore 73MB position is at the outer end of the platter (1GB position), not inner.

you think me wrong?

btw, my DVD has a burst rate of 75 MB/s[/QUOTE]

Actually, the “75 MB/s” burst speed from your DVD drive is only from its SATA interface connection (between the drive’s cache buffer and the motherboard’s controller) - in most cases, that would be the speed of the drive’s cache memory buffer itself. Physically, no current DVD drive can read or write any faster than 24x (or just slightly over 30 MB/s) due to rotational speed issues (the DVD drive requires at least 25,000 RPM - more than fast enough to cause any optical media to physically shatter into trillions of pieces, since even the most robust optical media can only withstand rotational speeds up to just over 10,000 RPM - in order to physically read or write at 75 MB/s even on the outer tracks). The blue curve in HD Tune refers to the physical sequential transfer speed of the tested hard drive. That’s the big difference between the two.

And as I stated, modern hard drives actually read from the outer edge inwards - not from the inner hub outwards. And today’s hard drive read/write technology is very different from what was the case back in the late 1980s and early 1990s: Those early drives wrote the exact same amount of data per track regardless of the linear speed of each track. But today’s drives write the same amount of data per linear inch of a track, which results in the outer tracks holding much more data than the inner tracks per rotational circumference.


#20

[QUOTE=RJL65;2499540]While what you claimed was at one point true during the really early years of hard drives…[/QUOTE]

Thank you, I realise technology has changed a lot and I am grateful for [B]all [/B]feedback here :flower: