• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

seagate Barracuda is really slow

S

Shasoosh

n00b
Joined
Aug 1, 2011
Messages
37
I returned a faulty 3TB Barracude and got a new one recently (ST3000DM001)
I'm trying to figure out why this drive is so slow. (Read is really slow)
Here are my results:

t9FeVGf.png


From reading online I should be getting something close to this:

ATTO-Seagate-Barracuda-3TB-HDD-ST3000DM001.png


I've tried different sata cables, different sata ports, different power cables, etc'. The drive is on AHCI mode.

Should I try anything else or is that another fault drive?

Thanks,
 
Last edited:
advanced format / sector alignment? (did you align the partition to 4k sector boundaries)
 
Good point. The drive is dual booting with Windows 7 (NTFS) and Yosemite Mac (HFS+ ?)

Is there a chance that because there are two file systems on the drive it is slowing it down?

"did you align the partition to 4k sector boundaries" Nop. I do I do that?
 
i have no idea how an HDD could get 150MBps in 4K transfer so i have to say that test is false.
 
His test looks good, drive is performing exactly like it should.

That random one he found online, is completely false, someone really spoofed that result.
 
cyclone3d said:
The read on his test looks really low to me.
Click to expand...

his read is definitely not working right but if this is an OS drive or program is running and reading it can mess his results up. Did you run it several times? Its peculiar that read is worse than writes.

Also it looks like in that other ATTO test that someone took an SSD and put it in a SATA gen 1 port because those speed are false for sure or may even just made the whole thing up with false numbers everywhere.

EDIT: looking at it again. It appears to me that the drive has to be currently accessed during that test because he is getting inconsistent speeds as if something is accessing it during the rest. Or his drive is failing in some weird way and having inconsistent performance. Thats the best i can come up with :/ sorry
 
Thanks for taking the time to reply guys.
Although Windows suggested that there is no need for it I've run a defrag on the drive and run the tests again. It seems like CrystalDiskMark has improved significantly while ATTO stayed the same. I have no idea how this is possible. Which test should I trust?

H3KnYxl.png


Because CrystalDiskMark results seemed a bit off and inconsistent I've run that test again. Once on version 4 and once on version 3.
The results on version 3 are completely different:

f7d0NAf.png


I've no idea what is going on..
 
Last edited:
Something is not right... your reads should be faster than your writes in all tests.

Edit: just looked at a review for a motherboard where they use CrytalDisk.. and your reads just look weird and your writes should not be that much faster than reads in tests where the writes are faster... Do you have the SATA controller drivers installed on your motherboard and not just the default MS ones?
 
Last edited:
I'm using the default ones..never had a problem with them though.

There are only SATA Preinstall drivers for my board (H87-D3H).. I'm guessing that I'll need to reinstall windows to try them out?
 
Shasoosh said:
I'm using the default ones..never had a problem with them though.

There are only SATA Preinstall drivers for my board (H87-D3H).. I'm guessing that I'll need to reinstall windows to try them out?
Click to expand...

http://www.gigabyte.com/products/product-page.aspx?pid=4491#dl

make sure you got the most upto date drivers. It may not do anything but maybe it will...worth a shot :/ And thank you for giving us members updated tests to make sure it wasn't a one time fluc.

Errrr. I just noticed the HD Tune. Can you run it over the whole drive and not 40GB HD tune is reporting good speeds over the first 40GB of the drive.
 
SomeGuy133 said:
Errrr. I just noticed the HD Tune. Can you run it over the whole drive and not 40GB HD tune is reporting good speeds over the first 40GB of the drive.
Click to expand...

Great advice! I didn't noticed it too.

Here are 2 tests one after the other. If I'm reading this correctly the drive gets really slow on 3 places. Is that enough proof to get that drive replaced?

xQSh2xL.png
 
Last edited:
Shasoosh said:
Great advice! I didn't noticed it too.

Here are 2 tests one after the other. If I'm reading this correctly the drive gets really slow on 3 places. Is that enough proof to get that drive replaced?

xQSh2xL.png
Click to expand...

did those 3 spot show several times? I have had drives that are slow in a couple spots (just like that in the past) but not sure if its really a major issue. Does it act slow?

As far as HD tune shows your drive is golden minus the dips but not sure if thats an actual issue. I think those 2 other benchmarks are glitching. They maybe putting data across the while drive because its full so it runs a fragmented test, which show you why defragmented drive is good :) Download a free trial of disk keeper and do a complete defrag so everything is 100% moved. It'll take ~ 1 day to complete if drive is truly full and fragmented
 
Is your computer busy doing something else during those 3 dips? They all look about the same distance (time?) apart.
 
Ran atto just as a comparison and your read speeds do seem a bit off
Capture_zpszawemym6.png
[/URL][/IMG]
mine gets dips as well in hd tune...i don't think it means anything myself
 
primetime said:
Ran atto just as a comparison and your read speeds do seem a bit off
Capture_zpszawemym6.png
[/URL][/IMG]
mine gets dips as well in hd tune...i don't think it means anything myself
Click to expand...

is your drive full or empty? Doesn't ATTO require creating a file to run test? I don't think HD Tune Pro requires such a thing. I think it takes files and tests because how would it work with a full drive? ATTO on his drive may have the test file scattered all over the splatter resulting in bad speeds. Yours may be in a single small section
 
HDTune reads the blocks of the filesystem directly. ATTO creates files on the filesystem.
 
drescherjm said:
HDTune reads the blocks of the filesystem directly. ATTO creates files on the filesystem.
Click to expand...

so am I right in my understanding? His issues with bad test results could be the file is scattered all over the platter because the drive is full and he is getting results based off fragmented test file?
 
The first CDM read benchmark is very low. I can't explain that. And the drive is only 45% full at that time. Surely there will be enough free space to allocate a contiguous file of a few GB.
 
drescherjm said:
The first CDM read benchmark is very low. I can't explain that. And the drive is only 45% full at that time. Surely there will be enough free space to allocate a contiguous file of a few GB.
Click to expand...

file system puts file where it pleases IIRC so it has nothing to do with ATTO or crystal disk mark, which is why i said download the trial of total fdisk or disk keeper...forget what its called.

here
http://www.condusiv.com/evaluation-software/default.aspx?p=home

Do the complete optimization where it moves every file to the outer part of the platter "towards the front" in terms of speed. Then re run the test. I bet your drive looks like data has been shot gunned all over it

EDIT: i bet its low because it recreates the file each test so each time his hard drive is slightly different so that one test was probably in a very fragmented spot. I am think this is the best theory until he scans and completely optimizes the drive in order to rule that theory out. I'll check back on this thread and ponder after i see a pic of his drives data lay out and the results of the optimization. Make sure you run it now because it will take fracking forever with a 3 TB drive half full,
 
SomeGuy133 said:
is your drive full or empty? Doesn't ATTO require creating a file to run test? I don't think HD Tune Pro requires such a thing. I think it takes files and tests because how would it work with a full drive? ATTO on his drive may have the test file scattered all over the splatter resulting in bad speeds. Yours may be in a single small section
Click to expand...

The partition i checked is mostly empty and is one of 3 the drive is divided into. Cant speak for his but mine stays pretty well defraged since this partition is only used for occasional windows backups
 
primetime said:
The partition i checked is mostly empty and is one of 3 the drive is divided into. Cant speak for his but mine stays pretty well defraged since this partition is only used for occasional windows backups
Click to expand...

yea but your fragments are limited to 1/3 of the drive...not the whole thing so the performance hit will be less. Imagine a 4 GB file spread across the entire platter or even half.
 
I formatted the drive. Here are the before and after tests.

E2XVEa8.png


zS8ZU8q.png


What does it teach me? I have no idea.
Getting the drive half full can degrade its speed so tremendously? Never happened to me before.
Should I still RMA it?

One thing I do know for sure is that the OS X partition wasn't the problem because it is still present on the drive.
 
To be honest I've given up worrying too much about mechanical drives speed results. As long as 1TB+ HDDs hit around 180-200MBps+ on the fist 100GB using HDTune Pro I'm happy.
 
crystal you can do while in windows but I have found HD tune is more precise in safe mode without programs accessing the drive.
That would be the drops in the graph.
 
as i thought the drive was good. It was getting accessed/fragmented as hell. You really need to use diskeeper or whatever. It takes forever but an optimization every so often is great. Also short stroking is great or partitioning to keep separate sections for different types of data. First 600-1000GB as a single partition followed by the rest or just have it split in halves or thirds. It allows less horrible fragmentation and prioritizing data.

It does perform odd though compared to my drives but i dont own many HDDs and i am not an expert to say whether that low 512k read (or is the 512k write high?) and high 4k write is normal

I always get around 1-1.5MBps 4K on my drives and read and write are always fairly close but i dont know if what you are now getting is uncommon or odd but its different to me but seems reasonable.
 
Last edited:
my crystal mark as a comparison....much different than your for sure
Capture_zpsomww40oc.png
[/URL][/IMG]
 
what version should i have used? nm i see it
Capture_zpsyknlcyda.png
[/URL][/IMG]
 
Last edited:
Do you mind running a transfer rate in HD Tune?
Bottom right picture from my last image.
 
to doubt check you did a slow (full) format? Might as well check for bad sectors and fix them
 
i am still totally lost on your 4k IOPS and some weird consistency. Canbad sectors cause those inconsistencies?
 
I copied my drive 800 gigs of data and run another test. I don't know if the drive is inconsistent or the tests are not that accurate.

4IUbors.png
 
Shasoosh said:
I copied my drive 800 gigs of data and run another test. I don't know if the drive is inconsistent or the tests are not that accurate.

4IUbors.png
Click to expand...

well you changed the variables of the test by adding data so the test should look different. I still would get diskkeeper trail and optimize it. If smart doesn't give any bad warning signs of bad sectors it should be fine but dont quote me
 
You must log in or register to reply here.
Back
Top