Anandtech: Understanding SSDs and New Drives from OCZ

[Tzupy]

Yes, but when updating the firmware you lose all data. And if you get a bad 1199 IRC firmware when building a new PC, tough luck.
So IMO the update must be done on a different machine, not a new build.

[m^2]

Yes, but when updating the firmware you lose all data. And if you get a bad 1199 IRC firmware when building a new PC, tough luck.
So IMO the update must be done on a different machine, not a new build.

Just make a partition image and store it temporarily on another drive. And most people know sb. who has a working computer.

[jmke]

Hi,

I can tell you just one thing:

I'm using OCZ Vertex since almost 2 weeks and sold already my Samsung F1. I will post benchmarks tomorrow - after 9 day of using + benchmarks after 'reseting' an aligning but still with 'old' firmware. You can expect something interesting. I really don't care about benchmarks anymore - maybe Intel SSD is better with benchmarks but I do not notice ANY problems with my 30GB Vertex.
In real life usage (home user, WinXP SP3, FF cache on RamDisk) I don't see how higher random writes can speed up PC beyond this what I see now - simply speaking, everything is NOW. In my opinion there is no more problems with writes for SSD's - at least as a OS's drive. Only thing which I'm still concerned is longevity of those drives.

Sorry if I sound too enthusiastic but I'm using one of those Vertex (not just reading benchmarks) and before I was VERY skeptical about SSD. Now I'm just looking for some money to earn to bay more Vertex...

I needed to build a case (not a physical ATX case) to promote SSD vs HDD at work, showing benchmark numbers doesn't mean much for the people I have to present this for, best way to show the difference is just by doing that: showing.

http://www.madshrimps.be/gotoartik.php?articID=923

Installed two Dell D630 with identical OS (WinXP test 2, Win7 test 1) and compared conventional HDD to OCZ Vertex (30gb edition). The difference is massive to say the least

If you're wondering about upgrading to SSD, you have to check this out;

any and all feedback is most welcome!

[Eunos]

Thanks for the article jmke, I love the real world comparisons.

[|Romeo|]

I needed to build a case (not a physical ATX case) to promote SSD vs HDD at work, showing benchmark numbers doesn't mean much for the people I have to present this for, best way to show the difference is just by doing that: showing.

http://www.madshrimps.be/gotoartik.php?articID=923

any and all feedback is most welcome!

It'd be nice to see a cut together video as well - i.e. re syncing the video of the two laptops every time the windows boot phase changes (initial progress bar, time from password entry to desktop etc.). I like the video; but I'd like to be able to see which phases of the boot are going faster. I tried to follow it on the current video and got confused very quickly...

[jmke]

that will be a future project, got hold of some good software tools to measure performance of loading apps/opening files. won't be ready by tomorrow though

[jessekopelman]

Depends on the controller, doesn't it?

I thought you would use an ordinary on-board SATA controller, with a max bandwith of 300 MB/sec. About the "unnecessary", I am sorry, that was a bit harsh indeed.

That's "theoretical 300 MB/s per drive", not for all drives together.

No the 300MBps is for all the drives together. Remember, the "S" in SATA is for serial.

[shleepy]

I thought you would use an ordinary on-board SATA controller, with a max bandwith of 300 MB/sec. About the "unnecessary", I am sorry, that was a bit harsh indeed.

That's "theoretical 300 MB/s per drive", not for all drives together.

No the 300MBps is for all the drives together. Remember, the "S" in SATA is for serial.

I've seen someone do measurements on ICH10R (standard X58 southbridge) and his peak was around 600-650 MB/s (with 5 X25-M's, which means that there was barely any difference in theoretical peak performance between two X25-M's and five of them). But I've now gotten a RAID card that will do much more (and went for a 4th X25, heh). I can't wait to try this out!

[dhanson865]

I thought you would use an ordinary on-board SATA controller, with a max bandwith of 300 MB/sec. About the "unnecessary", I am sorry, that was a bit harsh indeed.

That's "theoretical 300 MB/s per drive", not for all drives together.

No the 300MBps is for all the drives together. Remember, the "S" in SATA is for serial.

http://en.wikipedia.org/wiki/Serial_ATA

The current SATA rev. 2.x specifications detail data transfer rates as high as 3.0 Gbit/s per device

That's 300 MB/s per SATA drive.

PATA was 133 MB/s per "channel" which could be one or two drives making the worst case scenario 66.x MB/s for a single drive.

The fact that one is parallel and the other is serial has no bearing on the fact that SATA of today is vastly faster than the PATA of yesterday.

A system with two hard drives on one PATA 133 channel vs a system with two hard drives on SATA 300 is looking at 133 vs 600. Add more drives and the SATA bandwidth for the system increases.

Is there an upper limit at which a host controller can't keep up with the SATA load? Sure, but that doesn't excuse trying to blame that on the word "serial".

Look at it the http://www.serialata.org/3g.asp way if it helps you to understand.

The 16-bit wide parallel Ultra ATA bus is capable of transmitting two bytes of data per clock. Though Serial ATA transmits only a single bit per clock, the serial bus may be run at a much higher speed to compensate for the loss of parallelism. Serial ATA was introduced with a bandwidth of 1500Mbits/sec, or 1.5Gbits/sec. Because data is encoded using 8b/10b encoding (an 80% efficient encoding used with digital differential signaling to maintain a constant average “DCâ€

[jessekopelman]

dhanson865, I think you misunderstand what I am saying.

That's 300 MB/s per SATA drive.

Sure a single drive can be 300MBps, but the channel can also only handle 300MBps. If your single device is using all 300MBps, there will be virtually nothing left for other devices on the same channel.

This has nothing to do with SATA being faster than PATA. I never disputed that. It has to do with parallel vs. serial buses. A serial bus can only talk to one device at a time. If the bus has a bandwidth of 3Gbps, that applies not only to any individual device but to the aggregate of all devices being used simultaneously on the same channel. I guess the confusion here is that it is not either or. 300MBps is both the speed limit for an individual device AND for the bus. Meanwhile, with PATA the speed limit per device and per bus is different, because the bus is parallel (ie can talk to more than one device at a time).

But a drive connected by a single cable to a single port straight to the SATA controller will not share it's 3GB/s with another drive plugged in to the next port over.

Assuming that each port represents a dedicated channel, agreed. I do not think this is always the case, though. I'm pretty sure board makers have been known to build port multipliers right into their MB and give you more ports than the on board controller actually has channels for. Same deal for some of the cheaper add-in SATA controller cards.

[Tobias]

I obviously can't tell for each and all manufacturers, but I looked at the first best ASUS board I could find and the specifications say both 6 SATA ports as well as 6 connectors...
http://www.asus.com/product.aspx?P_ID=Q ... templete=2

As for looking at it from the board level here is a description from Anand on SB750:

The SB700/750 features six SATA 3.0Gb/s ports

http://www.anandtech.com/cpuchipsets/sh ... i=3360&p=2

I think it is safe to say that each port represents its own channel, otherwise we would see boards with 12-18 connectors, which we simply don't.

[dhanson865]

dhanson865, I think you misunderstand what I am saying.

That's 300 MB/s per SATA drive.

Sure a single drive can be 300MBps, but the channel can also only handle 300MBps. If your single device is using all 300MBps, there will be virtually nothing left for other devices on the same channel.

I'm about to start quoting the princess bride here.

You keep using the word channel the in such a way that makes me think you don't realize were the lines are drawn.

PATA channel = 1 cable with 3 connectors. 1 to the controller and 1 to the master and 1 to the slave.

SATA channel = 1 cable with 2 connectors. 1 to the controller and 1 to the drive.

There is no way for more than one SATA drive to be on the same channel short of a something being between the drive and the SATA controller. So I'd rewrite that as:

Sure a single drive can be 300MBps, but the channel can also only handle one drive. If your single device is using all 300MBps, any other drive will be on another channel. Short of port replication for eSATA enclosures you shouldn't be seeing that sort of bottleneck.

Two SATA drives in a standard ATX case going straight to a SB600 or SB750 chipset I'd expect those drives to get 600MB/s. I'd expect the southbridge to be able to handle 1200MB/s if you plugged in four drives on 4 cables to four ports on the motherboard.

Plug those same two or four drives in an eSATA enclosure/port replicator and I'd expect them to be stuck doing 300MB/s because the whole thing goes down to one cable between the PC and the external enclosure.

In short eSATA isn't going to be a good way to do RAID with SSDs as the bandwidth back to the PC isn't there.

But show me how you think the southbridge on a modern PC won't keep up with the 300MB/s per channel of the ports physically present on the motherboard. At this point I don't see it.

[dhanson865]

http://www.hardwaresecrets.com/article/27/4 shows you old school external SATA with bandwidth numbers.

SATA bandwidth will improve this year or next with SATA 6 GB/s replacing the 3GB/s standard we know and love.

But I fully expect SSDs to outstrip even that doubling of bandwidth in the next few years and multiple disk enclosures won't be as interesting to me down the road if they don't offer multiple data cables.

[jessekopelman]

But show me how you think the southbridge on a modern PC won't keep up with the 300MB/s per channel of the ports physically present on the motherboard. At this point I don't see it.

I don't think that.

[m^2]

But I fully expect SSDs to outstrip even that doubling of bandwidth in the next few years and multiple disk enclosures won't be as interesting to me down the road if they don't offer multiple data cables.

A few months is more like it. If not less. There's a lot of space for improving bandwidth.

[dhanson865]

[quote]Three year warranty now standard on all Vertex and Summit drives, even the ones already sold to customers

http://www.dailytech.com/article.aspx?newsid=15629

OCZ is also extending the warranty of all their premium level SSDs to three years. This includes all Vertex and Summit drives including the Mac Edition and Vertex EX, but excludes the Agility series due to their low cost NAND flash. Customers who previously bought Vertex and Summit drives will also have their warranties extended from two years to three. The company cited consumer demand, differentiation from competitors, and confidence in their products as key factors in lengthening the warranty period.

“Extending the warranty makes the OCZ total solution even more robust for consumers, and OCZ believes that product quality should always come first,â€