So how are you liking the Fusion drive?

[Yougotcarved]

Originally Posted by motrek

I suppose anything's possible but what would be the point of that design? Do you think Apple wants to prevent you from accessing your data too fast?

Of course they would if it makes me want to splurge $1300 on their overpriced SSD!

Anyway, I'm not necessarily saying this conspiracy theory that they're making it slow on purpose. Does Microsoft want their browser to be awful? Does Dell want to make crappier AIO's than Apple? Bad design doesn't mean they want to make it bad.

It's not a far out suggestion. It makes sense in a way, if I haven't opened a file in a few weeks it doesn't seem that unreasonable that Fusion might think of it as "unused data" and move it onto the HDD to give a bigger write buffer on the SSD. Does make perfect sense in a way when you think about it.

[motrek]

Originally Posted by Yougotcarved

Of course they would if it makes me want to splurge $1300 on their overpriced SSD!

Anyway, I'm not necessarily saying this conspiracy theory that they're making it slow on purpose. Does Microsoft want their browser to be awful? Does Dell want to make crappier AIO's than Apple? Bad design doesn't mean they want to make it bad.

It's not a far out suggestion. It makes sense in a way, if I haven't opened a file in a few weeks it doesn't seem that unreasonable that Fusion might think of it as "unused data" and move it onto the HDD to give a bigger write buffer on the SSD. Does make perfect sense in a way when you think about it.

I doubt they would go the extra mile to offload more data when there's no indication that you're going to overflow your write buffer in the first place... if anything, after you do a big write to the drive, it would increase the buffer at THAT point, not at some vague time later when you haven't been stressing the write buffer.

[atteligibility]

Originally Posted by motrek

I doubt they would go the extra mile to offload more data when there's no indication that you're going to overflow your write buffer in the first place... if anything, after you do a big write to the drive, it would increase the buffer at THAT point, not at some vague time later when you haven't been stressing the write buffer.

I think you're missing the point of what he's saying.
He's not asking what would be the best implementation of fusion technology or how it 'should' behave. But he asks how it does actually behave, and if the claims from arstechnica review have been verified or observed by other testers.
I am an iOS developer, and believe me, things from Apple are not always implemented the way you would expect them to be, or the way they should be. So let's not conclude that it's implemented a certain way, just because it would not make sense otherwise...

I was also impressed by arstechnica review, but I would also like it to see more detailed reviews and not rely on a single non official source as the source of truth.

[motrek]

Originally Posted by atteligibility

I think you're missing the point of what he's saying.
He's not asking what would be the best implementation of fusion technology or how it 'should' behave. But he asks how it does actually behave, and if the claims from arstechnica review have been verified or observed by other testers.
I am an iOS developer, and believe me, things from Apple are not always implemented the way you would expect them to be, or the way they should be. So let's not conclude that it's implemented a certain way, just because it would not make sense otherwise...

I was also impressed by arstechnica review, but I would also like it to see more detailed reviews and not rely on a single non official source as the source of truth.

I'm also an iOS developer and I'm coming at this from the point of view of Occam's razor. If we have a good working explanation of how Fusion drive works, why worry that it might be more complicated? Besides, there's no way we're ever going to know exactly how it works because Apple will never disclose that level of technical detail, so what's the point of all this hand-wringing?

Besides, I think we all agree that hard drive performance is *acceptable* although not *ideal*. So basically everybody is extremely worried about Fusion drive because it might "only" give you *acceptable* performance in some corner cases. This is really one of the most ridiculous debates I've ever seen.

Apple makes all sorts of design decisions all the time when designing its hardware and software products and I rarely see so much debate about how they do it. For example, why aren't there vigorous arguments about how Apple calculates a tap point from a contact patch on its trackpads, with people complaining bitterly that Apple won't let them change the algorithms, etc. but for some reason when Apple introduces what amounts to a big hard drive cache, everybody starts getting grey hairs from stress over how it's implemented.

[atteligibility]

Originally Posted by motrek

I'm also an iOS developer and I'm coming at this from the point of view of Occam's razor. If we have a good working explanation of how Fusion drive works, why worry that it might be more complicated? Besides, there's no way we're ever going to know exactly how it works because Apple will never disclose that level of technical detail, so what's the point of all this hand-wringing?

Besides, I think we all agree that hard drive performance is *acceptable* although not *ideal*. So basically everybody is extremely worried about Fusion drive because it might "only" give you *acceptable* performance in some corner cases. This is really one of the most ridiculous debates I've ever seen.

Apple makes all sorts of design decisions all the time when designing its hardware and software products and I rarely see so much debate about how they do it. For example, why aren't there vigorous arguments about how Apple calculates a tap point from a contact patch on its trackpads, with people complaining bitterly that Apple won't let them change the algorithms, etc. but for some reason when Apple introduces what amounts to a big hard drive cache, everybody starts getting grey hairs from stress over how it's implemented.

if you don't care at all on how this thing works, then great, not sure why you would want to stop the rest of us from trying to understand though.

Now your comparisons with the other technologies don't make as much sense. Although understanding the trackpad working details would be interesting for some of us, fact is it would not impact any decision we would make about it.

The fusion drive, for some configuration, requires an extra $450, I think it's fair for people to want to understand the technology before investing that much in it

[motrek]

Originally Posted by atteligibility

if you don't care at all on how this thing works, then great, not sure why you would want to stop the rest of us from trying to understand though.

Now your comparisons with the other technologies don't make as much sense. Although understanding the trackpad working details would be interesting for some of us, fact is it would not impact any decision we would make about it.

The fusion drive, for some configuration, requires an extra $450, I think it's fair for people to want to understand the technology before investing that much in it

You make a good point and I can see why people might be very interested in understanding exactly how their $250-$450 upgrade works.

Although I might point out that so far, not one single person on this thread has contributed one iota to that understanding. The work done by Anandtech and Ars Technica was interesting. All the posts to this thread amount to nervous, pointless hand-wringing. What's the point of *speculating* that Fusion drive *might* not work *100% exactly* like the Ars Technica implies? etc.

[harcosparky]

Originally Posted by Fishrrman

I prefer to partition my drives and keep system and data files "separated" (I have 8 drive icons on my desktop right now), so "fusion" isn't for me.

Indeed, it's "going in the opposite direction" from where I want to be.

If it works for others, fine.

I sense that within a year or two, we're going to start seeing posts that go, "my fusion drive won't boot -- what now?"

If they had two -separate- drives (SSD & HDD) inside, with a bootable OS on both, they'd be up running within two minutes....

I with you.

I'm old school about things.

My internal drive pretty much just has OSX on it, I have two external drives.

One for Apps, the other for data and each of those are mirrored onto two other drives.

I don't want to suffer too much should the internal drive ever fail and should the computer die I have everything outside to move over to another machine.

You only have to live through a catastrophic failure with massive data loss once to learn the hard way!


I also have OSX installed to a bootable USB Thumb Drive so I can boot from that if I need to.

[Yougotcarved]

Originally Posted by motrek

Although I might point out that so far, not one single person on this thread has contributed one iota to that understanding. The work done by Anandtech and Ars Technica was interesting. All the posts to this thread amount to nervous, pointless hand-wringing. What's the point of *speculating* that Fusion drive *might* not work *100% exactly* like the Ars Technica implies? etc.

Because earlier someone said the ars technica review proved 100% what was going on. And I said actually it didn't prove it 100%. That's all I said and I don't really see what can even be argued about that.

[APPLE-FANATIC]

For some reason, I cannot tell whats on the fusion drive and whats not. How are guys seeing this information?