OCZ is developing junk

[nanaki333]

ocz drives may be crappy but i like their technical support,unlike gskill they have a garbage ssd support forums.

yeah. ocz is very efficient in their rma department too. i sent a vertex 1 for the 2nd time and they got sick of messing with it and sent me a brand new vertex 3.

 

[PandaBear]

My biggest problem with 3LC memory is not that it only works for 1k cycles, because after wear leveling it would still be significantly more than what we need. The problem is that it would be FREAKING SLOW.

 

[nanaki333]

My biggest problem with 3LC memory is not that it only works for 1k cycles, because after wear leveling it would still be significantly more than what we need. The problem is that it would be FREAKING SLOW.

that's what everyone said about MLC when it was first beginning life. which, at the start, it was slow, but as controllers matured, they became faster and faster and were outperforming most SLC drives. now they make server MLC drives now.

if TLC comes out, it may be sluggish at first, but once controllers mature like before, it probably won't be any different to the end user between TLC and MLC. it's all about the controller 🙂

 

[AdamK47]

The majority of the performance benefits of SSDs are from the controller and the layout of the NAND. I can see 3LC performing good in the right configuration.

 

[groberts101]

I think a certain someone is actually enjoying going back and forth in this thread. This thread is going to last in perpetuity at this rate. Full of useless... whatever this is. I'd like to see an actual serious thread about what OCZ is developing.

Can't really deny that, I guess. Sorry for the pettiness and will try harder to stay on subject.

Is pretty easy to want to screw with some members around here as they seemingly deserve it at times. Two children certianly don't equal an adult though. lol

And I'd love to see a thread like that too.. but then these kinds of guys show up to protect us from ourselves yet again. Hard to have viable conversations involving OCZ it seems. Can't say I blame some for having bad feelings but it sure would be nice if they bit their tounges more often. Hard to act like adults when the kids show up, if you know what I mean.

 

[groberts101]

The majority of the performance benefits of SSDs are from the controller and the layout of the NAND. I can see 3LC performing good in the right configuration.

yep.. more channels of slower chip configs still equals higher throughput.

 

[LOL_Wut_Axel]

Pot....kettle.....black

There is a right way to express your "opinion" and a wrong way. Which path do you think you have taken in this thread, which has gone to sh!t now?

Granted I agree with your "opinion" of OCZ as stated in my post, but that again is just my opinion and I am not going to start personal attacks on people over it. Just say what you have to say and move on. You are not going to change other people's opinions using such tactics.

Well, it's not really opinion. Based on the facts we CAN conclude that SandForce has had more reliability issues than Marvell or Intel, or even Toshiba.

But I rest my case. Some people here are simply blind and because of that will continue to buying current SF/OCZ drives while saying they're just as likely to fail as any other controllers/drives and ignoring all the evidence to support otherwise. I personally call those people either "willfully ignorant" or in other cases "fanboys". If SF/OCZ make a reliable drive that's fast I'd consider it then.

 

[taltamir]

that's what everyone said about MLC when it was first beginning life. which, at the start, it was slow, but as controllers matured, they became faster and faster and were outperforming most SLC drives. now they make server MLC drives now.

Top of the line MLC drives are outperforming generations old SLC. However SLC is still a lot faster for the same generation and we still have SLC drives out there.

Also MLC is 100% increase in data while same transistor cost while TLC is a 50% increase in data coupled to a 20% extra transistor cost giving you a net gain of only 25% net gain in data/transistor aka a cost saving of a mere 20%.
Yet the penalties in speed are supposed to be even greater than the SLC to MLC transition..

20% cost reduction is barely worth the speed penalty (write limit/lifespan penalty is indeed meaningless for anything but a high volume server)