The good news is that for most desktop workloads you don’t really benefit from being able to execute more than 20K IOPS, at least in today’s usage models.
To add to what Bauss said, Anand himself in the Vertex 3 preview said that more than 20K IOPS was fast enough for consumer workloads in today's usage models.
From the last paragraph on page 1.
Since I mostly surf the web, fold and play games, SEQ speeds are more important to me than random speeds.
Yep. Games are something like 70 to 80% SEQ Reads, since your basically just streaming data(usually textures) the entire time. Loading and saving games might be random, but that is usually so fast on a PC anyway, I don't think it matters. I'm not sure about downloading big files, but what I plan to do is download things like installers and such to the SDD, then archive them to the HDD incase I need them again later.Since I'm building a new system mainly for gaming, and that If I ever do download or save big files in my Spinpoint 1 TB drive, than the SSD that I'd be using to install Windows and my primary games should have better Seq Speeds than Random Reads?
As I mentioned in another thread, On a P67 chipset MB 6Gb/s Intel port running Vantage HDD suite (which is much more closer to real PC usage than other benchmarks like ATTO/ASSSD/CDM) C400's get in the 60,000 range while C300's on the same setup get in the 45,000 range at best.
I think I should put my 2 cents in before people lose their collective minds.
This is the same stuff we saw when the Intel 510 benches started coming out. We see a drive that has OK random performance, while making huge strides in sequential performance and we all run for the 4k performance as if it's the only key to a fast drive. I think it's time for a re-evaluation.
Frankly, I think Intel and Crucial know something we don't. I feel like they realized you don't need gobs of random performance on a consumer SSD, in most cases. More often than not, sequential speed is just as (if not more) important.
Now I know this pretty much goes against everything we think we know, but consider this--Why would both Intel and Crucial take a step back in 4k performance and invest more in larger, more sequential performance? Because we don't need it. Applications don't use it. I bet they saw that:
a) Most computer applications have been programmed for DECADES to mask HDD latency by caching tons of crap in memory and using the drive itself for as little as possible.
b) Most computers that have an SSD in them will (by looking at who buys SSDs) have so much RAM in them as well (4GB at least) that our drives never become a bottleneck whilst IN an application.
The first generation of SSDs came with some pretty bad tradeoffs. For one, 4k writes were just as bad, or even worse, than HDDs. It was out of frustration with that generation that people started asking for random 4k performance more fitting for an SSD.
Remember this?
http://www.anandtech.com/show/2614/8
Since then, we've seen iterations on SSD tech by Intel, Samsung, SandForce, Marvell, etc. and for some time now, 4k performance has not only gotten better, but WAY better. It's now 100x better than it originally was in the early days of consumer SSDs.
But people still haven't got the memo that 4k performance now 'good enough' for our needs. I bet that Intel and Crucial found that investing in more 4k performance will only end up hurting larger sequential performance and stopped tuning firmware for it. I don't doubt that's why we're seeing dips in 4k performance, and huge gains in sequential performance. They're, essentially, finding the right balance.
Overall, if you're a consumer, and you're not putting this drive in a server, CHILL OUT. You'll be fine. Just because this drive does 70MB/s on 4k write and not the 200MB/s+ random 4k writes you think you need, doesn't mean it's gonna be as bad as the 0.40MB/s random 4k writes of the first-gen JMicron SSDs. There's a HUGE and TANGIBLE difference between those two numbers.
One last note, more often than not, higher sequential performance and decent random 4k performance will win the day in certain scenarios that consumers should be looking at. I mean look at any Intel 510 review. Look how it beats the C300 and the old Intel G2 9 times out of 10 in file copy and app load tests--The very things that will make our SSD feel VERY quick.
Anyways, I had to get that out, because I'm tired of seeing people rethink good SSD choices because of some false expectation. I'd take a 0.01s longer save time on my MS Word files over a 10s longer movie file transfer any day.
My biggest complaints about the 510 actually aren't about Intel's use of a 3rd party controller, instead they are about the drive's lackluster random read performance. In a horrible bout of irony Intel fixed its sequential performance and moved backwards in the random department. Random read performance, as it turns out, has a pretty major impact in the real world.
I think I should put my 2 cents in before people lose their collective minds.
This is the same stuff we saw when the Intel 510 benches started coming out. We see a drive that has OK random performance, while making huge strides in sequential performance and we all run for the 4k performance as if it's the only key to a fast drive. I think it's time for a re-evaluation.
Frankly, I think Intel and Crucial know something we don't. I feel like they realized you don't need gobs of random performance on a consumer SSD, in most cases. More often than not, sequential speed is just as (if not more) important.
<snip>
I'd take a 0.01s longer save time on my MS Word files over a 10s longer movie file transfer any day.
I’d argue that if you’re a desktop user and you’re using an SSD as a boot/application drive, what will matter most is latency. After you’ve got your machine setup the way you want it, the majority of accesses are going to be sequential reads and random reads/writes of very small file sizes. Things like updating file tables, scanning individual files for viruses, writing your web browser cache. What influences these tasks is latency, not bandwidth.
If you were constantly moving large multi-gigabyte files to and from your disk then total bandwidth would be more important. SSDs are still fairly limited in size and I don’t think you’ll be backing up many Blu-ray discs to them given their high cost per GB. It’s latency that matters here.
I am still going with the C300. It has 34nm NAND and a mature, issue free firmware.
There's too many people talking about "real world". In my real world, I run a couple of VMs for testing. I need the 4k random I/O as this lets me do testing quicker. Real world for the average office user is completely different and a first gen SF drive will probably suffice.
I knew that Intel will kick it down a gear once they get some good rep and market share. However, I did expected something different from Micron. I think I will have to bite the bullet and go for SF 2xxx and try not to fall for their firmware games (I'm not going with OCZ because of their deceptive selling tactics)
I'm seconding that emotion.
The 256GB drives are on Ebay for 400.00.
I'm gonna try and hold out for 350.00....but ya just never know.