OCZ Technology Reveals Indilinx Everest 2 Controller

[tecmo34]

OCZ Technology reveals its second generation Indilinx Everest 2 controller at CES 2012

OCZ Technology Reveals Indilinx Everest 2 Controller : Read more

 

[Pyree]

Sounds promising (550/500 MBps read/write and 90000 IOPS). Looks like what I will be getting to replace my Vertex 2.

 

[evang]

can't wait to see how much that 2tb costs

 

[joytech22]

[citation][nom]evang[/nom]can't wait to see how much that 2tb costs[/citation]
You and me both.

PLEASE be under $2000! HAHA Just kidding, One can dream though..

 

[jrharbort]

That is a MASSIVE performance increase over their second generation controller. Looks like SandForce will finally get a proper run for its money.

 

[alidan]

if anyone knows the the die size for how big a gigabyte is out whatever nanometer, please tell me I would love to do the math and figure out how much they up the price on SSD, because really right now I can't see them gouging us too much, but then again I don't know how big these chips are.

In my chips I don't mean the plastic case around the chips on their I mean actual chips themselves, the silicon.

about 21 TB per wafer
wafer costs $50,000
21 divided by 50,000
it comes to about $2380 per terabyte

if I knew the die size I could tell you exactly how much it would cost, the materials alone.

------------------

I used to just want them to lower the price of SSD's, but then I realized that they are made of silicon-based memory and silicon cost crapton to make a wafer, now I understand why to go for speed over cost

 

[back_by_demand]

The SSD market should have capitalised on the recent HDD flood debacle to have an industry wide price drop, ramp up production and uptake would have gone through the roof.

 

[stingstang]

[citation][nom]back_by_demand[/nom]The SSD market should have capitalised on the recent HDD flood debacle to have an industry wide price drop, ramp up production and uptake would have gone through the roof.[/citation]
That actually did happen. It was only just recently that some ssd's went at or below the 1gb/$ threshold.

 

[back_by_demand]

[citation][nom]stingstang[/nom]That actually did happen. It was only just recently that some ssd's went at or below the 1gb/$ threshold.[/citation]
It was close to that level already, just natural decrease to that level isn't good enough, a hard cut, maybe 75 cents per GB would have been enough to tip people over the edge.

 

[freggo]

It would be interesting to see a graph comparison of the thru put development of hard drives vs. SSDs.
I have a feeling that hard drives have reached a plateau where the sizes go up but the MB/sec are at their technical limit while SSDs seem to be getting still faster and faster.

There's got to be the point when PC makers start offering systems that come standard with an SSD boot drive and a HD for mass storage.

HP...Dell... anyone reading this ?

 

[CaedenV]

[citation][nom]freggo[/nom]It would be interesting to see a graph comparison of the thru put development of hard drives vs. SSDs.I have a feeling that hard drives have reached a plateau where the sizes go up but the MB/sec are at their technical limit while SSDs seem to be getting still faster and faster.There's got to be the point when PC makers start offering systems that come standard with an SSD boot drive and a HD for mass storage.HP...Dell... anyone reading this ?[/citation]
HDDs can be quite fast on sequential throughput, especially when you bring cost and size into consideration. You can easily have a larger and 'faster' RAID array for the same price as many high end SSDs. The problem is when you start doing anything non-linier, Random seek time for HDDs is somewhere arround 4.5-7ms, while a full seek time from center to edge is ~8.5-10ms, and those numbers largely have not changed from the turn of the century. It is in this random IO throughput and having a 0ms seek time that SSDs become so powerful, and no amount of RAIDing HDDs can catch up because it is a physical limit for each and every drive within the RAID. Anywho, my point is that SSDs are not the best tool for ever job, and it will be a really really long time before they are considered a good replacement for general bulk storage of things like video and audio which require a ton (relatively a ton) of space, but do not require much throughput at all.

As far as SSDs becoming standard, that is one of the requirements of Ultrabooks. Granted the SSDs they will be getting will be slower cheaper SSDs, not the screamers like these.

 

[willard]

[citation][nom]CaedenV[/nom]HDDs can be quite fast on sequential throughput, especially when you bring cost and size into consideration. You can easily have a larger and 'faster' RAID array for the same price as many high end SSDs. The problem is when you start doing anything non-linier, Random seek time for HDDs is somewhere arround 4.5-7ms, while a full seek time from center to edge is ~8.5-10ms, and those numbers largely have not changed from the turn of the century. It is in this random IO throughput and having a 0ms seek time that SSDs become so powerful, and no amount of RAIDing HDDs can catch up because it is a physical limit for each and every drive within the RAID. Anywho, my point is that SSDs are not the best tool for ever job, and it will be a really really long time before they are considered a good replacement for general bulk storage of things like video and audio which require a ton (relatively a ton) of space, but do not require much throughput at all.As far as SSDs becoming standard, that is one of the requirements of Ultrabooks. Granted the SSDs they will be getting will be slower cheaper SSDs, not the screamers like these.[/citation]
You'd need several 15k RPM drives in an array to get anywhere close to a good SSD's performance, and it's going to cost you nearly as much. The question is really, how much performance do you really need?

Right now, we're still in the realm of scaling very well with increased disk performance. I've used good 7200 RPM drives in a RAID 0 array, and my mid-range SSD that I bought nearly a year ago is so much faster it's not even funny.

I'm not talking benchmarks either. There is a very real, very large, very tangible performance increase for the entire system by using an SSD. They're expensive, but you get what you pay for. The best mechanical drives on the planet can pull off ~150 MB/s on sequential reads on the edge of the platter. An average SSD these days easily triples that, and some come close to quadrupling it.

 

[jgutz2006]

This is great performance as their current chips drastically dip in performance as the capacities lower. So now they are competitive but i do not expect OCZ to replace sandforce and use this as the Vertex 4 just yet because it is simply just on par with its current Vertex3 drives. So they catch up to sandforce how many months after the current champ, sandforce chipset? What do you all think that sandforce is not working on the next gen controller?

I'm predicting a Vertex 4 to include the next Sandforce chip when released (And a competing Octane drive with in-house controller) and if OCZ can continue its development as rapidly as they have been, they could maybe drop the sandforce for indillinix on the Vertex 5 performance line unless they simply drop vertex naming and continue with the Octane naming.
I've got my money on that scenario unless the next sandforce chipset takes another year or so to come to market!

 

[scook9]

The vertex was an Indilinx Barefoot controller.....

http://www.ocztechnology.com/ocz-vertex-series-sata-ii-2-5-ssd.html

Right off OCZs own page.....

 

[jonny_76]

[citation][nom]alidan[/nom]if anyone knows the the die size for how big a gigabyte is out whatever nanometer, please tell me I would love to do the math and figure out how much they up the price on SSD, because really right now I can't see them gouging us too much, but then again I don't know how big these chips are.In my chips I don't mean the plastic case around the chips on their I mean actual chips themselves, the silicon.about 21 TB per waferwafer costs $50,00021 divided by 50,000it comes to about $2380 per terabyteif I knew the die size I could tell you exactly how much it would cost, the materials alone.------------------I used to just want them to lower the price of SSD's, but then I realized that they are made of silicon-based memory and silicon cost crapton to make a wafer, now I understand why to go for speed over cost[/citation]

Why frustrate yourself thinking about how cheap the materials are ? How much do you think is the raw die price of the latest multi core Intel CPU, is it really thousands of dollars/die ? How much is the raw materials cost of your car ? How much is the raw materials cost of the human body ? My point is that almost all items have minimal raw material costs and the huge markups are due to fab amortization (>$3B for 2xnm), supply and demand, convenience, profits, R&D and lastly what is customer willing to pay (HP tablets found that out recently).

 

[alidan]

[citation][nom]jonny_76[/nom]Why frustrate yourself thinking about how cheap the materials are ? How much do you think is the raw die price of the latest multi core Intel CPU, is it really thousands of dollars/die ? How much is the raw materials cost of your car ? How much is the raw materials cost of the human body ? My point is that almost all items have minimal raw material costs and the huge markups are due to fab amortization (>$3B for 2xnm), supply and demand, convenience, profits, R&D and lastly what is customer willing to pay (HP tablets found that out recently).[/citation]

actually i did the math on those cpus.

intel doesn't rebrand a cpu with disabled areas correct? so anything that isn't 100% working they throw out, now, the cheapest of those expensive ones are about 500$ correct, assuming some yield problems, and not enough cpus to sell to really make up all the r&d, more of it gets dumped onto them than a more mainstream models.

i believe the figure i came up with was 350$ as a base cost of the cpu, when you take into account yield problems, and extra r&d costs being dumped onto a lower number of chips instead of being spread out over a larger amount, intel isnt making anywhere near what you think they are.

on a side note, bulldozer is only pulling in at most 40$ per chip.

my point being, if i know how big the chips are now, and whatever nm they are being made at, i can reliably tell you how much XXXgb will cost you at XXXnm process. my current numbers are based off of a new ssd, that isn't priced to move. it will take till the 10nm process before they can realistically over take hhds in mainstream computers, and 5nm till they compete space wise with current hdds on $ per cost.

the reason i want to know is because i want to know how much the base chips cost apposed to how much they sell it for, is there really much of a premium they attack to the drives, are we being gouged because we can be.... i want answers for those questions and i will find them out if i can get the chip sizes and what nm they are on.