HD speeds

  • Thread starter Thread starter tumppiw
  • Start date Start date
T

tumppiw

Motherboard is Asus F1A75-M pro A6-3650 and 16GB RAM
Internal disks are connected on SATA ports 3 and 4 (BIOS says 6G mode)
(ports 5 and 6 are opticals (both DVD+-RW)
OS is Win7 64bit

Are these normal HD speeds or should I tweak something?

-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 x64 (C) 2007-2012 hiyohiyo
-----------------------------------------------------------------------
1st (system) partition on 1st internal disc Samsung HD103UJ 1000GB
Sequential Read : 101.077 MB/s Sequential Write : 95.066 MB/s
Random Read 512KB : 36.022 MB/s Random Write 512KB : 49.178 MB/s
Random Read 4KB (QD=1):0.466 MB/s [113.8 IOPS]
Random Write 4KB (QD=1):1.073 MB/s [261.9 IOPS]
Random Read 4KB (QD=32):0.714 MB/s [174.2 IOPS]
Random Write 4KB (QD=32):1.082 MB/s [264.3 IOPS]
----------------------------------------------------------------------
only partition on 2nd internal disc Samsung HD154UI 1500GB
Sequential Read: 101.734 MB/s Sequential Write: 96.616 MB/s
Random Read 512KB: 32.947 MB/s Random Write 512KB: 42.499 MB/s
Random Read 4KB (QD=1) :0.388 MB/s [94.8 IOPS]
Random Write 4KB (QD=1) :0.802 MB/s [195.9 IOPS]
Random Read 4KB (QD=32) :0.708 MB/s [173.0 IOPS]
Random Write 4KB (QD=32) :0.840 MB/s [205.0 IOPS]

----------------------------------------------------------------------
USB 2.0 connected 2TB disc WD Elements 2TB
Sequential Read : 32.732 MB/s Sequential Write :31.511 MB/s
Random Read 512KB : 19.346 MB/s Random Write 512KB : 32.474 MB/s
Random Read 4KB (QD=1) :0.345 MB/s [ 84.3 IOPS]
Random Write 4KB (QD=1) :0.729 MB/s [ 177.9 IOPS]
Random Read 4KB (QD=32) :0.449 MB/s [ 109.6 IOPS]
Random Write 4KB (QD=32) :0.632 MB/s [ 154.3 IOPS]
-----------------------------------------------------------------------
2.5” 160GB laptop drive in ext USB3.0 enclosure Toshiba MK1367GSX
Sequential Read : 42.955 MB/s Sequential Write : 41.779 MB/s
Random Read 512KB : 23.421 MB/s Random Write 512KB : 24.412 MB/s
Random Read 4KB (QD=1) :0.444 MB/s [108.4 IOPS]
Random Write 4KB (QD=1) :0.840 MB/s [205.0 IOPS]
Random Read 4KB (QD=32) :0.508 MB/s [124.0 IOPS]
Random Write 4KB (QD=32) :0.848 MB/s [207.1 IOPS]
-----------------------------------------------------------------------
WD6400AAKS in ext USB 3.0 docking station 640GB
Sequential Read : 115.457 MB/s Sequential Write : 114.711 MB/s
Random Read 512KB : 48.975 MB/s Random Write 512KB : 64.458 MB/s
Random Read 4KB (QD=1) :0.703 MB/s [171.6 IOPS]
Random Write 4KB (QD=1) :1.404 MB/s [342.7 IOPS]
Random Read 4KB (QD=32) :0.755 MB/s [184.3 IOPS]
Random Write 4KB (QD=32) :1.432 MB/s [349.6 IOPS]


--
 
tumppiw said:
Motherboard is Asus F1A75-M pro A6-3650 and 16GB RAM
Internal disks are connected on SATA ports 3 and 4 (BIOS says 6G mode)
(ports 5 and 6 are opticals (both DVD+-RW)
OS is Win7 64bit

Are these normal HD speeds or should I tweak something?

-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 x64 (C) 2007-2012 hiyohiyo
-----------------------------------------------------------------------
1st (system) partition on 1st internal disc Samsung HD103UJ 1000GB
Sequential Read : 101.077 MB/s Sequential Write : 95.066 MB/s
Random Read 512KB : 36.022 MB/s Random Write 512KB : 49.178 MB/s
Random Read 4KB (QD=1):0.466 MB/s [113.8 IOPS]
Random Write 4KB (QD=1):1.073 MB/s [261.9 IOPS]
Random Read 4KB (QD=32):0.714 MB/s [174.2 IOPS]
Random Write 4KB (QD=32):1.082 MB/s [264.3 IOPS]
----------------------------------------------------------------------
only partition on 2nd internal disc Samsung HD154UI 1500GB
Sequential Read: 101.734 MB/s Sequential Write: 96.616 MB/s
Random Read 512KB: 32.947 MB/s Random Write 512KB: 42.499 MB/s
Random Read 4KB (QD=1) :0.388 MB/s [94.8 IOPS]
Random Write 4KB (QD=1) :0.802 MB/s [195.9 IOPS]
Random Read 4KB (QD=32) :0.708 MB/s [173.0 IOPS]
Random Write 4KB (QD=32) :0.840 MB/s [205.0 IOPS]

----------------------------------------------------------------------
USB 2.0 connected 2TB disc WD Elements 2TB
Sequential Read : 32.732 MB/s Sequential Write :31.511 MB/s
Random Read 512KB : 19.346 MB/s Random Write 512KB : 32.474 MB/s
Random Read 4KB (QD=1) :0.345 MB/s [ 84.3 IOPS]
Random Write 4KB (QD=1) :0.729 MB/s [ 177.9 IOPS]
Random Read 4KB (QD=32) :0.449 MB/s [ 109.6 IOPS]
Random Write 4KB (QD=32) :0.632 MB/s [ 154.3 IOPS]
-----------------------------------------------------------------------
2.5” 160GB laptop drive in ext USB3.0 enclosure Toshiba MK1367GSX
Sequential Read : 42.955 MB/s Sequential Write : 41.779 MB/s
Random Read 512KB : 23.421 MB/s Random Write 512KB : 24.412 MB/s
Random Read 4KB (QD=1) :0.444 MB/s [108.4 IOPS]
Random Write 4KB (QD=1) :0.840 MB/s [205.0 IOPS]
Random Read 4KB (QD=32) :0.508 MB/s [124.0 IOPS]
Random Write 4KB (QD=32) :0.848 MB/s [207.1 IOPS]
-----------------------------------------------------------------------
WD6400AAKS in ext USB 3.0 docking station 640GB
Sequential Read : 115.457 MB/s Sequential Write : 114.711 MB/s
Random Read 512KB : 48.975 MB/s Random Write 512KB : 64.458 MB/s
Random Read 4KB (QD=1) :0.703 MB/s [171.6 IOPS]
Random Write 4KB (QD=1) :1.404 MB/s [342.7 IOPS]
Random Read 4KB (QD=32) :0.755 MB/s [184.3 IOPS]
Random Write 4KB (QD=32) :1.432 MB/s [349.6 IOPS]
Click to expand...

As a "plumbing test", the results in the bottom graph show
an A75 chipset achieving 291MB/sec read and 190MB/sec write.
The read rate is high enough, to correspond to SATA III.
Notice though, that some motherboards in an equivalent test,
can do 512MB/sec. The processor on the fastest motherboard
is stronger, but such tests should be doing DMA transfer,
and the entire CPU should not be used/needed while benchmark
testing. So for some reason, the A75 didn't do as well here.

http://www.techspot.com/review/432-gigabyte-g1-sniper2/page10.html

*******

Here is another test, of A75 SATA III plus "good SSD".
~500MB/sec sustained. This result looks better.

http://techgage.com/article/battle_of_the_sata_30_controllers/

( http://techgage.com/articles/storage/sata_30_chipsets/kingston_hyperx_ssd_atto_better_pc.png )

So the SATA III port, really can run at SATA III rates.

*******

This is an HD103UJ. 99.3MB/sec sustained, near the start.

http://extreme.pcgameshardware.de/a...t-normal-hdtune_benchmark_samsung_hd103uj.png

Burst rate = 132.8MB/sec is supposed to be a way of
proving the bandwidth available in the cable. The
burst rate measurement is tricky to do correctly,
which is why for accuracy, the latest version
of the utility is required. When I tested my computer
here, and its current hard drive, I had to reboot
the computer and run the benchmark, before the burst
rate achieved a reasonable value. That rate
is suspiciously close to Ultra133 for some reason.

A burst rate of 132.8MB/sec is not conclusive. It's close
to the top of SATA I transfer rate.

When I do SATA I versus SATA II testing in the picture,
I get "clipping" at 130MB/sec in SATA I mode. The burst
result of 132.8MB/sec is rather close to that value.
Either the cable/interface in the example is in SATA I
mode, or the burst transfer is doing poorly for a SATA II
setup.

You can use the burst rate, as a "check of the plumbing",
as a means to verify whether the cable transfer mode
is an issue. You can get a free version of HDTune here,
suitable for read benchmarks.

http://www.hdtune.com/files/hdtune_255.exe

These are my results, comparing my current disk
using WinXP and Win2K operating systems, with a
disk and Intel Southbridge running in SATA II mode.
Burst transfer is 197MB/sec in WinXP and 211MB/sec
in Win2K. Those rates are sufficiently high, to be
in SATA II transfer range. But not high enough
to qualify for SATA III. But my chip is not
SATA III, so it's not reasonable to expect a higher
burst rate.

http://img685.imageshack.us/img685/9205/st500dm002burst.gif

The burst rate is a function of the Southbridge, cable,
hard drive controller and front-end RAM cache. If the RAM
cache does not have sufficient bandwidth to handle full
cable rate, then the burst transfer rate could suffer.

On older drives, a SATA to IDE chip inside the hard drive,
limits burst transfers to around 100MB/sec. The very first
SATA drives were "bridged" implementations, using an old
IDE controller schematic, with a SATA to IDE chip glued
to the front. The second generation of SATA drives,
used "native" chip designs, and those today reach
~500MB/sec in burst tests, on SATA III controllers.

But if you see a 100MB/sec burst measured value in HDTune,
do not despair. It means you need to reboot and
repeat the test after Windows has stopped accessing
the disk after the boot process is complete. I got
absurdly bad burst transfer numbers just now, and
had to carefully repeat the test to get "good" numbers.

You cannot even get the utilities to agree on the value,
the measurement is that difficult to do well.

Paul
 
You cannot even get the utilities to agree on the value,
the measurement is that difficult to do well.

Paul
Click to expand...

Sorry about the setting of my clock. It seems booting
into Win2K, changed the time by one hour, when I returned
to WinXP.

Paul
 
Motherboard is Asus F1A75-M pro A6-3650 and 16GB RAM
Internal disks are connected on SATA ports 3 and 4 (BIOS says 6G mode)
(ports 5 and 6 are opticals (both DVD+-RW)
OS is Win7 64bit

Are these normal HD speeds or should I tweak something?
Click to expand...

I'll usually get nowhere near an optimal suggested, but go by a
sequential operation (constant transfer rates over prolonged periods)
from a defragmented preparation. A lessor part might be location, or
partition, in the case of fewer but larger drives, or generally how
the computer is used in subjective, or some lesser sense perhaps tests
fully don't embody. If to say what I might however expect is
something along twice my current speeds if anytime soon I were to run
into a supportively ported MB. Sweet enough among hardware
advancements.
 
tumppiw said:
Motherboard is Asus F1A75-M pro A6-3650 and 16GB RAM
Internal disks are connected on SATA ports 3 and 4 (BIOS says 6G mode)
(ports 5 and 6 are opticals (both DVD+-RW)
OS is Win7 64bit

Are these normal HD speeds or should I tweak something?

-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 x64 (C) 2007-2012 hiyohiyo
-----------------------------------------------------------------------
1st (system) partition on 1st internal disc Samsung HD103UJ 1000GB
Sequential Read : 101.077 MB/s Sequential Write : 95.066 MB/s
Random Read 512KB : 36.022 MB/s Random Write 512KB : 49.178 MB/s
Random Read 4KB (QD=1):0.466 MB/s [113.8 IOPS]
Random Write 4KB (QD=1):1.073 MB/s [261.9 IOPS]
Random Read 4KB (QD=32):0.714 MB/s [174.2 IOPS]
Random Write 4KB (QD=32):1.082 MB/s [264.3 IOPS]
----------------------------------------------------------------------
only partition on 2nd internal disc Samsung HD154UI 1500GB
Sequential Read: 101.734 MB/s Sequential Write: 96.616 MB/s
Random Read 512KB: 32.947 MB/s Random Write 512KB: 42.499 MB/s
Random Read 4KB (QD=1) :0.388 MB/s [94.8 IOPS]
Random Write 4KB (QD=1) :0.802 MB/s [195.9 IOPS]
Random Read 4KB (QD=32) :0.708 MB/s [173.0 IOPS]
Random Write 4KB (QD=32) :0.840 MB/s [205.0 IOPS]

----------------------------------------------------------------------
USB 2.0 connected 2TB disc WD Elements 2TB
Sequential Read : 32.732 MB/s Sequential Write :31.511 MB/s
Random Read 512KB : 19.346 MB/s Random Write 512KB : 32.474 MB/s
Random Read 4KB (QD=1) :0.345 MB/s [ 84.3 IOPS]
Random Write 4KB (QD=1) :0.729 MB/s [ 177.9 IOPS]
Random Read 4KB (QD=32) :0.449 MB/s [ 109.6 IOPS]
Random Write 4KB (QD=32) :0.632 MB/s [ 154.3 IOPS]
-----------------------------------------------------------------------
2.5” 160GB laptop drive in ext USB3.0 enclosure Toshiba MK1367GSX
Sequential Read : 42.955 MB/s Sequential Write : 41.779 MB/s
Random Read 512KB : 23.421 MB/s Random Write 512KB : 24.412 MB/s
Random Read 4KB (QD=1) :0.444 MB/s [108.4 IOPS]
Random Write 4KB (QD=1) :0.840 MB/s [205.0 IOPS]
Random Read 4KB (QD=32) :0.508 MB/s [124.0 IOPS]
Random Write 4KB (QD=32) :0.848 MB/s [207.1 IOPS]
-----------------------------------------------------------------------
WD6400AAKS in ext USB 3.0 docking station 640GB
Sequential Read : 115.457 MB/s Sequential Write : 114.711 MB/s
Random Read 512KB : 48.975 MB/s Random Write 512KB : 64.458 MB/s
Random Read 4KB (QD=1) :0.703 MB/s [171.6 IOPS]
Random Write 4KB (QD=1) :1.404 MB/s [342.7 IOPS]
Random Read 4KB (QD=32) :0.755 MB/s [184.3 IOPS]
Random Write 4KB (QD=32) :1.432 MB/s [349.6 IOPS]
Click to expand...

As a "plumbing test", the results in the bottom graph show
an A75 chipset achieving 291MB/sec read and 190MB/sec write.
The read rate is high enough, to correspond to SATA III.
Notice though, that some motherboards in an equivalent test,
can do 512MB/sec. The processor on the fastest motherboard
is stronger, but such tests should be doing DMA transfer,
and the entire CPU should not be used/needed while benchmark
testing. So for some reason, the A75 didn't do as well here.

http://www.techspot.com/review/432-gigabyte-g1-sniper2/page10.html

*******

Here is another test, of A75 SATA III plus "good SSD".
~500MB/sec sustained. This result looks better.

http://techgage.com/article/battle_of_the_sata_30_controllers/

(
http://techgage.com/articles/storage/sata_30_chipsets/kingston_hyperx_ssd_atto_better_pc.png
)

So the SATA III port, really can run at SATA III rates.

*******

This is an HD103UJ. 99.3MB/sec sustained, near the start.

http://extreme.pcgameshardware.de/a...t-normal-hdtune_benchmark_samsung_hd103uj.png


Burst rate = 132.8MB/sec is supposed to be a way of
proving the bandwidth available in the cable. The
burst rate measurement is tricky to do correctly,
which is why for accuracy, the latest version
of the utility is required. When I tested my computer
here, and its current hard drive, I had to reboot
the computer and run the benchmark, before the burst
rate achieved a reasonable value. That rate
is suspiciously close to Ultra133 for some reason.

A burst rate of 132.8MB/sec is not conclusive. It's close
to the top of SATA I transfer rate.

When I do SATA I versus SATA II testing in the picture,
I get "clipping" at 130MB/sec in SATA I mode. The burst
result of 132.8MB/sec is rather close to that value.
Either the cable/interface in the example is in SATA I
mode, or the burst transfer is doing poorly for a SATA II
setup.

You can use the burst rate, as a "check of the plumbing",
as a means to verify whether the cable transfer mode
is an issue. You can get a free version of HDTune here,
suitable for read benchmarks.

http://www.hdtune.com/files/hdtune_255.exe

These are my results, comparing my current disk
using WinXP and Win2K operating systems, with a
disk and Intel Southbridge running in SATA II mode.
Burst transfer is 197MB/sec in WinXP and 211MB/sec
in Win2K. Those rates are sufficiently high, to be
in SATA II transfer range. But not high enough
to qualify for SATA III. But my chip is not
SATA III, so it's not reasonable to expect a higher
burst rate.

http://img685.imageshack.us/img685/9205/st500dm002burst.gif

The burst rate is a function of the Southbridge, cable,
hard drive controller and front-end RAM cache. If the RAM
cache does not have sufficient bandwidth to handle full
cable rate, then the burst transfer rate could suffer.

On older drives, a SATA to IDE chip inside the hard drive,
limits burst transfers to around 100MB/sec. The very first
SATA drives were "bridged" implementations, using an old
IDE controller schematic, with a SATA to IDE chip glued
to the front. The second generation of SATA drives,
used "native" chip designs, and those today reach
~500MB/sec in burst tests, on SATA III controllers.

But if you see a 100MB/sec burst measured value in HDTune,
do not despair. It means you need to reboot and
repeat the test after Windows has stopped accessing
the disk after the boot process is complete. I got
absurdly bad burst transfer numbers just now, and
had to carefully repeat the test to get "good" numbers.

You cannot even get the utilities to agree on the value,
the measurement is that difficult to do well.

Paul
Click to expand...

Link to HDtune 2.55 result for the boot disk (Samsung HD103UJ)
(the machine's been on for ~6hours)

--
 
tumppiw said:
Forgot.. here it is

https://skydrive.live.com/#cid=FDBD79FA40B204A0&id=FDBD79FA40B204A0!442
Click to expand...

OK, you have a nice smooth curve. And the info I could find earlier,
suggests 114.7MB/sec is not abnormal.

The "burst rate" of 123.2MB/sec does bother me.

Try rebooting the computer, then do the HDTune benchmark again
about two minutes after the computer has settled down.

I had the same problem as you, abnormally low burst, until
I rebooted. And then I could get 197MB/sec burst rate,
on a SATA II interface.

The 123.2MB/sec number is low enough, that the interface could
be in SATA I mode. It's possible to take a SATA II drive,
and artificially force the interface to run at SATA I rate.
That can be done with an ATA command, or it can be done
with a Force150 jumper. Not all SATA drives have a jumper
block, so on some brands, only the ATA command set works
to set the cable transfer rate.

*******

According to this, your "burst rate" should be 175MB/sec max.

http://www.myce.com/review/Samsung-Spinpoint-F1-HD103UJ-review-15169/

And look at this. When connected to an Intel SATA II Southbridge,
their HDTune curve and numbers, look exactly like yours.
Burst rate is only 121MB/sec.

http://www.myce.com/review/Samsung-...view-15169/Reading-and-Writing-Performance-2/

When they used the HDTach benchmark, the burst rate reported
there is 191MB/sec, consistent with SATA II performance. I would
be pleased if my result looked like this. (Even though the
Samsung specification predicts a burst rate of 175!)

http://static.ultrafastcdn.com/images_reviews/reviews/samsung_hd103uj/image015.png

You can get a copy of HDTach 3.0.4.0 here, if you need a copy.
The Simplisoftware site seems to have disappeared, so I used
the archive.org site instead. Click the download button.

http://web.archive.org/web/20110209...isoftware.com/Public/index.php?request=HdTach

Paul
 
Back
Top