HD benchmark results from a defective drive

D

Daggah

2[H]4U
Joined
Mar 26, 2005
Messages
3,197
OK, so here's a little backstory. Like many gamers, I decided about 8 months ago to upgrade the hard drive on my PS3. I did the research, and chose a 5400 RPM drive because, at the time, the general accepted truth of it is that 7200 RPM drives might adversely affect performance. I picked a Western Digital Blue 250 GB 5400 RPM drive.

Ever since then (until last night, actually) I've had to deal with extremely slow hard drive performance on my PS3. Larger game demos took half an hour to install. The whole interface was a lot slower to load thumbnail images for games, movies, etc. Even game performance suffered in some games, if the game needed to load content from the hard drive during gameplay. Heavy Rain would pause and turn into a slideshow during more intense sequences, and a lot of the chapters would load with sound cutting in and out at first.

Some informal comparisons:

Bomberman Ultra demo from the PSN Store...a 73 mb demo...install time

On the 40 GB stock PS3 drive: 10 seconds
On the 250 GB WD drive: 1 minute
New 250 GB Seagate drive (also 5400 RPM): 8 seconds

Copying a 165 mb video from my media computer over my wireless connection (which makes this benchmark a bit suspect, I know)

On the 40 GB stock PS3 drive: 1 minute, 36 seconds
On the 250 GB WD drive: 3 minutes

So I did some more formal benchmarks, using "ATTO Disk Benchmark" found on the web.

Both of these tests are using a Thermaltake external hard drive dock connected via eSata.

WDBlue250GBBenchmark.jpg


vs

Seagate40GBBenchmark.jpg


I just thought this was kind of interesting. Yes, I could do more thorough testing...

I plan on RMAing the WD hard drive. These speeds are beyond merely "slow" - the hard drive, in my opinion, is defective.
 
And on a side note, I wonder if WD would be upset with me if I ship the drive back to them for RMAing in a Seagate box :)
 
They will not be upset. However I do not think your drive is defective. It could be but I am not convinced by just a benchmark. I believe it maybe your "Thermaltake external hard drive dock"

The seagate# indicate that for some reason the dock is limited to 40MB/s

Does the drive make strange noises?

Can you run CrystalMark to look at the smart data?
 
It's not the dock, I noticed horrible performance on this drive waaaay before I even thought about picking up the dock.
 
OK, I did some more benchmarks, with CrystalDiskMark. I haven't finished them yet, but the short story is, comparing the performance of my 40 GB PS3 drive, the performance is identical on and off the thermaltake dock.

Also, there's this:

WDBlue250GBInfo.jpg
 
Huh, that's weird...now the drive is performing OK (by OK, I mean on par with the PS3 drive)

WDBlue250GB_FAT32Benchmark02.jpg
 
Huh, that's weird...now the drive is performing OK (by OK, I mean on par with the PS3 drive)
Click to expand...

The drive probably did not reallocate a sector when you did the benchmark this time. This could mean that the drive has replaced all bad sectors or at least the ones that are involved in the test.

However with that many reallocated sectors 840 (0348 hex) I would send it back.
 
Last edited:
The problem is not the Reallocated Sector Count; those are fixed sectors that no longer could cause any problems.

The real problem is unfixed damage; bad sectors that have not yet been replaced by a reserve one. Those can be identified with the "Current Pending Sector" SMART-variable, and it tells you you have 5 bad sectors that have not been fixed.

To fix the damage, you need to write to those locations that have not been fixed. The easiest way to do that is a zero-write or full format under Windows7. That would lose you your data of course, but if completed it would allow the harddrive to fix its own damage; the Current Pending Sector count should be zero after this procedure and the number of Reallocated Sectors should rise.

So the most important SMART variable to look at is the Current Pending Sector count; as any non-zero value here means the drive should be fixed before continuing to be used.
 
The problem is not the Reallocated Sector Count; those are fixed sectors that no longer could cause any problems.
Click to expand...

I consider a drive with a large amount of reallocated sectors suspect at best. There is some physical problem causing that. On the 100+ drives I have checked at work most have 0 reallocated sectors even after 5 years of usage. A few have 15 or less. 1 that is still in use has 50 reallocated sectors but that is in a raid with a spare so I am not highly concerned. If it was under warranty I would definitely send it back if it had more than 20 or so reallocated sectors.
 
Well typically, what you should look for is how fast the reallocated sector count rises. If it is a high number but stays the same even after a month; that means that the drive is in no immediate danger. When its rising rapidly, each scan revealing new bad sectors; then you should replace the device and not use it anymore.

But regardless of reliability; the reallocated sectors themselves cannot cause any problems. It is the UN-reallocated sector count ("Current Pending Sector count") that causes problems. Those are bad sectors that are NOT HIDDEN from the operating system; so when the OS tries to reads them it will timeout and cause bad things to happen.

The bad sectors should be fixed before attempting to use the drive; or even recovering your data before sending it to the warranty service. Spinrite is good at fixing surface errors; else a zero-write on the affected sectors would fix the damage but lose any opportunity to recover the data those sectors were holding.
 
You must log in or register to reply here.
Back
Top