Just when I thought I had a handle on SSDs

[Phage]

OCZ do this.
http://www.anandtech.com/storage/sho...px?i=3702&p=10

Bl**dy hell.

[Infinite]

Whats your point? Newer technologies realised all the time. Considering the old Vertexs were out around a year ago. Its about time tbh.

Intels are out for a refresh mid 2010 and G3 out end of the year.

[CAT-THE-FIFTH]

Still too expensive though!!

[Phage]

Yes new tech, it's the new controller that's interesting. This isn't an OCZ innovation but a company called Sandforce. Their Durawrite method is very interesting as an alternative to Intel, despite it's drawbacks. What we're seeing is a fork.

And yes - they're all too expensive.

[CAT-THE-FIFTH]

It is good to see more competition in the market though!!

[Phage]

Aye - Like the man said "Competition is a good thing. We need companies to not only keep Intel aggressive on price, but competitive on features as well. Indilinx did the former and it looks like SandForce is going to do the latter"

[Funkstar]

I find the obsession about which controller and memory chips are being used in any given SSD quite interesting.

I wonder if this was the same at any point with hard dives. Surely it doesn't matter what controller or read/write technology is used in them, as long as the price or performance is there? You don't need to know what goes on inside a regular hard drive, as long as it works as you expect it to.

[Phage]

There's the rub. This is a new method of writing to the NAND pages. Sandforce use data compression and other super-sekret stuff to write to less blocks. However, this will only work if your data is uncompressed. If you already deal in compressed data, you will not see the benefit, and could therefore save a lot of cash.
Similarly, you would benefit from knowing that the Jmicron controller was much worse than the Indilinx and that using the first Intel TRIM update was a risky business.
I rather thought this sort of information was the point of an enthusiast site. You could say the same thing about any of the components in a modern home computer, or the whole system. Just buy any box, anything over £750 will do more than you need, why worry about the components ?

[Terbinator]

I'd already posted this link in the other SSD thread, performance is quite epic.

[Funkstar]

There's the rub. This is a new method of writing to the NAND pages. Sandforce use data compression and other super-sekret stuff to write to less blocks. However, this will only work if your data is uncompressed. If you already deal in compressed data, you will not see the benefit, and could therefore save a lot of cash.
Similarly, you would benefit from knowing that the Jmicron controller was much worse than the Indilinx and that using the first Intel TRIM update was a risky business.
I rather thought this sort of information was the point of an enthusiast site. You could say the same thing about any of the components in a modern home computer, or the whole system. Just buy any box, anything over £750 will do more than you need, why worry about the components ?

I'm not saying we shouldn't discuss these things, I just find it interesting that there is so much discussion about the different controllers and write methods

Definitely not having a go btw

[Phage]

Thanks.
It's interesting I think because it is the bleeding edge and a fork would mean competition. Competition means better value for money. Better value for money means...<slap>.
Ok, I'm back now. Sorry about that.

[kalniel]

I find the obsession about which controller and memory chips are being used in any given SSD quite interesting.

I wonder if this was the same at any point with hard dives. Surely it doesn't matter what controller or read/write technology is used in them, as long as the price or performance is there? You don't need to know what goes on inside a regular hard drive, as long as it works as you expect it to.

It wasn't really this way with hard drives, as the performance characteristics between different methods remained the same. It's precisely the large effect on performance characteristics that's caused controllers to be thrust into the spotlight for SSDs - at some point it becomes clearer to talk about an SSD having an Intel, Indilinx or JMicron controller than it does to describe the performance across a whole suite of tests that show up the differences like seq. read/write, random read/write, large cluster, small cluster, performance over time etc.