- Back to homepage

Understanding SSDs and New Drives from OCZ

@ 2009/03/18

If you look at the SSD market today, you’d assume that it’s very different from what it was just six months ago when the X25-M launched. People are worried that the Intel drive has issues with degrading performance over time. Some vendors are now shipping “revised” JMicron drives with multiple controllers, supposedly fixing all of the problems I talked about last year.

I hate to break it to you guys. As different as the world may seem today, it’s all very much the same.

The Intel drive is still the best of the best. Yes, it, and other SSDs do get slower over time and later in this article I’ll explain why it happens and why it’s not as big of a deal as you’d think. The issues I complained about with the JMicron drives from last year are still alive and well today; they’re just somewhat occluded.

Comment from jmke @ 2009/03/20

interesting thread here about the Indilinx controller used in OCZ Vertex, it seems to create some issues when the SSD is full...

Quote:

my rig:
nforce4 ultra/Opteron170 with (all) 4 ports in use by SATA2-HDDs, no RAID, using WinXP-Pro with nforce-driver v6.99

1) got my OCZSSD2-1VTX30G delivered late last week
2) found out about new firmware v1199 (but not about alignment with diskpar yet), which the drive didn't have
3) flashed new firmware (on 2nd PC with ICH5-R, because nforce-sata with installed nforce-driver wouldn't work)
4) reattached vertex, cloned my OS-Partition (16GB in size) in Windows using OO-DiskImage to Vertex. 5GB remained free on C:
5) bootet from vertex, enjoyed results very much. Installed MFT, created one partition with all available space left on drive.
6) setup windows-swapfile to C: 64MB and MFT-Partition to 3GB. Enjoyed performance over the weekend. Didn't do much on both partitions. No major install (fill) of C:, did bench 1 time with atto on MFT-Partition, copied about 1GB of quake3-data to MFT-Partition and played for 15 Minutes, that's all. Nevertheless, i got pretty heavy swapfile usage, since my computer runs 24/7 while using firefox with 100+ tabs and doing stuff with graphics.
7) now it's getting interesting:
copied 4GB ISO-File from another HDD to MFT-Partition.
when done, copied that file from MFT to C: Partition, while at the same time, copied one 700MB avi-file from C: to MFT-Partiton (so 2 copies made at the same time on the same drive (since i'm using SuperCopier, i can see the actual transfer-speeds).
then, after maybe 20% of the 700MB file was copied my Computer froze, with HDD-LED permanent ON. Waited for about 5 minutes (in hope the smaller of the two transfers would finish and it might unfroze, but didn't).
pushed Reset.
Vertex couldn't be detected by BIOS anymore, just tried to detect drives for about a minute (HDD-led was still on permanently). disconnected both cables from drive, re-attached it, pushed reset again. All drive(s) were detected again.
WinXP wanted to check drive-integrity (chkdsk) on start. Indeed it had to "repair" file-security information on every single file which took a while. Then BSOD.
Reset again, switched boot-drive to old HDD (previously cloned drive) in BIOS. Went back into Windows. When accessing Vertex in Explorer everything seemed ok. Chkdsk repaired about 5MB data on C:, everything else seemed to work (backuped about 2GB userfolder from SSD to normal-HDD with no error).

Cleared vertex with gaijin's WipeDisk. Then found out about alignment which doesn't seem to work properly somehow, since the vertex wouldn't be detected by WinXP drive-managment, but it gets a drive-letter assigned and i can copy data to it. When cloning my OS to the aligned partition again, it wouldn't boot from it.

That's it for now. I'm planning on reflashing the Vertex tomorrow (with new firmware if possible), then align and format it again and do a fresh WinXP-install on that drive), which i'll backup regurarly, until all pending issues are fixed. Just to be safe.

7 pages long already: http://www.ocztechnologyforum.com/fo...ad.php?t=53290 near the end new firmware is tested by OCZ which does seem to improve/remove this issue

Comment from Kougar @ 2009/03/20

Yes, but I tend to view the performance hit/stuttering as the main issue. The TRIM and other commands will mostly fix this.

The uncesseary erase-rewrites are a secondary issue that won't be fixed any time soon due to how flash is designed... its bad for the "lifespan" of the cells, but is unavoidable regardless.

Comment from jmke @ 2009/03/20

it won't fix the issue though, will just average the delay result, so instead of having a drive performance at its best until it runs out of space, you'll have one which will be more consistent either full or empty

Comment from Kougar @ 2009/03/20

Quote:

Originally Posted by jmke

from what I can understand, the new Trim command (if that is what you are referring to) is a bit of egg vs chicken, HDD/SSD manufacturers wait for an OS to support it before implementing it; don't know if SATA 6gb/s address this; haven't read anything about it yet

They can't implement it until it becomes part of the official ATA specification. That is what they are waiting on.

Once the ATA spec is updated then as ya say they must wait for OS support and also update drives already on the market... no chicken/egg situation, but it could be the better part of a year before TRIM (and other ATA commands, not just trim) are added.

Given the two-part situation to how the SATA standard is developed (ATA organization for commands/specs, SATA-IO organization for serial connections/standards, etc) it very well might not be part of the new SATA 6GB/s standard.

Comment from npp @ 2009/03/19

The lowest price I found here in Germany was 129E for the 30GB Vertex, which is kind of a bargain! The guys already have it in stock, I wish I'd only got the money to buy it... By the way, 30GB isn't that tiny at all, all my important stuff right now (except movies and music of course) fits perfectly well in 20GB. Damn, this looks sweet!

Comment from jmke @ 2009/03/19

Comment from Kougar @ 2009/03/19

This was eactly what I suspected, and worried about:

Quote:

There was a strange phenomenon a few people noticed, something I unfortunately picked up on after the review went live; if you filled the X25-M up and re-benchmarked it, it got slower.

If there was any element of truth to it, he would have found it eventually.

I'm glad Anand posted this. Every time I mention the JMicron drives still seem to suffer from stuttering or point at the problematic random write performance, people call me out on it. Or say the test was an outlier or wasn't a legit test because the other (sequential) write performance tests were not showing the same effect. About time I had a single reputable source to refer back to.

So far this is the best explanation I have seen that covers all angles and makes sure to drill the key points into place. I'm definitely bookmarking this one.

--

Finally, someone covers the new SSD-centric ATA commands I've been trying to research.

These commands are set to become part of the ATA spec sometime later this year. I could not get confirmation if they are included in the SATA 6GB/s spec or not. Hopefully Windows 7 will be updated to support them quickly once they do arrive.

First OCZ Summit tests I have seen. So much for this drive being the new"flagship" OCZ SSD. Used Vertex results drop less and used Vertex write latency is less impacted than the Summit...

Comment from jmke @ 2009/03/19

€140~150 OCZ Vertex 30Gb: http://www.alternate.de/html/product...Disk&l3 =SATA
http://www.memoryc.com/storage/solid...tml?Brands=OCZ
ordered one, let's try it out

Comment from npp @ 2009/03/19

I read the whole 30 pages or so and it's really the most in-depth review of SSDs I've seen so far. The sad conclusion is that despite the huge amount of SSD hovering around, most of them are pretty much the same (and they're crap) - something that I didn't realise fully before.

Comment from jmke @ 2009/03/18

hugely interesting & a must read for anybody considering buying SSD in this year.

All information and graphics contained in Madshrimps are sole property of the Madshrimps crew and may not be reproduced or copied in any manner without written permission from us.