Speed limit jumper has no effect

[selwynpolit]

I've been testing a bunch of my old hard drives (and then running Spinrite level 4 to see how long it estimates the operation will take just for fun)

Curiously, I found one drive: a Seagate Barracuda ES ST3500630NS 500GB made in 2007 that has a jumper for limiting the speed to 1.5 Gbps. Readspeed reports 76 Mbps down to 38 Mbps whether I have the jumper on or not. Does that make sense?

It seems to me that halving the throughput of the drive would cause those numbers to change

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  500GB ST3500630NS                    76.6    74.3    66.4    55.4    38.9

                  Benchmarked: Saturday, 2021-01-09 at 04:21

Here are the results with the jumper on:

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  1.0TB WDC WD10EACS-00ZJB0            77.3    73.0    61.8    55.7    39.1
 82  500GB ST3500630NS                    76.6    74.3    66.4    55.4    38.9

                  Benchmarked: Wednesday, 2021-01-06 at 00:17

I also just tested a ST3750640AS drive which has the same jumper setup and got the same results.

Thoughts?

 

[ColbyBouma]

SATA I can do 150 MB/s, so I'm not surprised. My guess is that it's a compatibility setting for older motherboards.

 

[DanR]

If my math is correct, 1.5 Gbps calculates to about 192 MB/s.
Thus, with the drives operating at about 77 MB/s at 0% the jumper would make no difference.

The drive speed profiles shown look normal and healthy.

 

[PHolder]

limiting the speed to 1.5 Gbps.

Well my math says 1.5 * 1024 * 1024 * 1024 = 1,610,612,736 Now if we divide by 10 (as a rough approximation of bytes plus overhead) that gives us 161,061,273.6 or 161MBps, which is still far higher than 76MBps (note that the capital B is important, as I believe ReadSpeed reports in bytes not bits.)
Also, as others have noted, you're probably being limited by the max speed of your link which is lower anyway.

 

[DanR]

I interpreted Gbps as Giga bits per sec. Not Giga Bytes per sec. Was I wrong?

 

[PHolder]

Gbps as Giga bits per sec

No, that is the correct interpretation. But there seems to be some mixing of the two from the OP:

limiting the speed to 1.5 Gbps. Readspeed reports 76 Mbps down to 38 Mbps

For the second and third values (76 Mbps and 38 Mbps), he's quoting from his ReadSpeed results:

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  500GB ST3500630NS                    76.6    74.3    66.4    55.4    38.9

And as far as I know, these numbers are MBps (mega Bytes per second), not Mbps (mega bits per second).

 

[selwynpolit]

@PHolder When you say

you're probably being limited by the max speed of your link

what link are you referring to? That model drive has a 3 Gbps thoughput spec.

Note that another drive a Seagate ST1000DM003 had these much higher results in the same computer. This drive is specced as 6Gbps

  +--------------------------------------------------------------------------+
  | ReadSpeed: Hyper-accurate mass storage read-performance benchmark. rel 1 |
  |  Benchmarked values are in megabytes read per second at five locations.  |
  +--------------------------------------------------------------------------+

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  1.0TB WDC WD10EACS-00ZJB0            77.4    72.9    57.7    55.5    39.2
 82  1.0TB ST1000DM003-9YN162            171.6   193.3   174.1   143.5    92.3

                  Benchmarked: Wednesday, 2021-01-06 at 01:53

 

[bertwin]

A quick google told me these results are to be expected of your hdd. And these speeds are well below the maximum of what SATA 1 can handle. That jumper seems to change whether the interface is SATA 1 or SATA 2, but setting it to SATA 2 doesn't make the drive go any faster. As @Dagannoth said, it's probably there for backwards compatibility.

 

[PHolder]

what link are you referring to? That model drive has a 3 Gbps thoughput spec.

We need to keep our units consistent here. Drive speeds and drive link speeds are frequently measured in BYTES per second, not BITS. (This is the units ReadSpeed will use.) 3Gbps is 3 * 1024 * 1024 *1024 = 3,221,225,472 bits per second which if we strictly convert to bytes per second (by dividing by 8 and by ignoring any overhead) we would get 402 megabytes per second. I don't think the drive is rated to provide data at that speed, that is the theoretical maximum of a SATA II link ( https://en.wikipedia.org/wiki/Serial_ATA#SATA_revision_2.0_(3_Gbit/s,_300_MB/s,_Serial_ATA-300) ) In any case, I assumed it was on a SATA I link where it would have been limited to a lower speed of 150MBps ( https://en.wikipedia.org/wiki/Serial_ATA#SATA_revision_1.0_(1.5_Gbit/s,_150_MB/s,_Serial_ATA-150) ).

 

[Steve]

I interpreted Gbps as Giga bits per sec. Not Giga Bytes per sec. Was I wrong?

Dan (and everyone) the SATA link speed refers to the raw data bit rate. But the link is encoded with a 10-bit to 8-bit encoding to provide for self-clocking and other features. So, in addition to other overhead, the raw link rate also needs to be scaled down by 8/10 since ten physical bits are used to communicate 8 logical data bits —> "8b/10b_encoding"