13" and 15" MacBook Pros Have a Slower SATA Interface

[Evangelion]

Are you guys daft??? Take a look at the data from the benchmark pics mentioned in the official Macrumors post:

You are quoting benchmarks. Benchmarks are not the same thing as real-life use. It's like complaining that your 3D-card gets fewer 3DMarks as some other 3D-card, even though in real-life apps there is no difference between the cards.

Find some real-life tests, as opposed to synthetic benchmarks.

Besides, the numbers you quoted for the new MBP's don't make any sense, if the reason is the SATA 1.5-connection. I mean, SATA 1.5 should achieve 150MB/sec, when you quoted 115MB/sec and 95MB/sec. So the reason might be elsewhere. If the SSD was limited by the SATA 1.5, it should be pegged at 150MB/sec, not 115 and 95MB/sec.

 

[cloud 9]

I got a 15" 2.66GHz uMBP in March '09. It has the 6MB L2 cache as well as the 3.0GB SATA interface. It is capable of 8GB memory.

When Apple announced these new laptops at WWDC, I was quite upset.

However, after learning of the downgrades, I am quite convinced that my machine is the best of the lot, perhaps even the best Apple has released in years.

Never felt better about a computer purchase at this point.

hahah, this one is hilarious....

yes sir, you were that intelligent to buy just that machine that's the best of Apple in years...

It's not by coincidence, it's because you were that smart and competent.
There aren't many people with a machine like yours because most weren't that smart buying it at that time, they are stupid, unlike you.



seriously, raise your hand if you know what it all means <_<

 

[harshw]

Here's an interesting article regarding OCZ's Vertex Series Mac Edition SATA II 2.5" SSDs:

OCZ had to slow down its SSDs because Mac OSX can't handle the speed


The article quotes Tobias Brinkmann, OCZ's Director of Marketing EMEA, as saying, "The Mac version has different read and write specs due to Mac OS limitations."

The article offers this (below) regarding the issue:

Please don't post FUD. That misunderstanding was due to a off the cuff remark by someone from OCZ. There is no difference between the Mac drives and the normal Vertexes except for the fact that the Mac drives are validated by testing with Apple machines. And this has been confirmed several times by OCZ.

 

[Cole Slaw]

Of course everyone (me included) would rather have SATA 3.0 than 1.5. No question about it. But fact is that the difference between the two is neglible, even when using fast SSD's.



If you want computer that are "the best they can be" you would have computer that cost 3-4 times as much as they do now. That's basically what happened with the ol' UNIX-workstations in the nineties. Sun, SGI, Digital etc. didn't cut corners anywhere, and the end-result was that the machines cost a lot of money. A lot. There comes a point when getting any meaningful improvements to the machine would cost a lot of money.

No, this feature (or lack of it) is probably not due to money, but some other things. But you are basically demanding that Apple must have a feature that would not really improve performance at all. It's like complaining that your expensive stereo-system didn't come with cable-elevators, regardless of the fact that cable-elevators do NOTHING for the sound-quality. Sure, in some audiophile-theories, separating the cables from the floor offers theoretical improvements to sound-quality, in reality there is no difference. Same thing here. SATA 3.0 has twice the bandwidth of SATA 1.5, but the actual real-life performance-benefits would be nonexistant (when using normal HD's) or minor (when using hi-end SSD's).

Sure, SATA 3.0 would be nice. But I wouldn't lose any sleep because the Mac has SATA 1.5.



So you demand features that offer no real performance-benefit in order to "commensurate the price they are charging"? By that logic you should also demand that the sheets of aluminium must be washed by unicorn-tears before machining. Hey, it offers no real benefit, but it would be nice to have, right? To "commensurate the price of the laptop"....

What bugs me is the gradual loss of capability I think we are seeing in the MBPs.
For example, for what they charge for a MBP, do you think it would be too much to ask to have say, 2 firewire ports (400 and 800), AND an Expresscard slot?
I mean, even my 20 month old pre-unibody MBP has that.
Why do they have to whittle that down to just a Firwire 800 port?
That's the kind of thing that's been pissing me off lately with Apple. They cut features like that to save a dollar or two.
The Macbook Pro is "Pro" in name only these days. Sadly, today it's basically a consumer grade notebook with a "Pro" name on it and an inflated Apple price tag.

 

[winterspan]

You are quoting benchmarks. Benchmarks are not the same thing as real-life use. It's like complaining that your 3D-card gets fewer 3DMarks as some other 3D-card, even though in real-life apps there is no difference between the cards.

That analogy is just plain wrong. While a synthetic graphics benchmark like 3Dmark is different from running Crysis, the harddrive "benchmark" used in the data I posted shows raw data stream read and write. The performance will be very close to the actual read/write file transfer performance you will moving files around in Finder.

Besides, the numbers you quoted for the new MBP's don't make any sense, if the reason is the SATA 1.5-connection. I mean, SATA 1.5 should achieve 150MB/sec, when you quoted 115MB/sec and 95MB/sec. So the reason might be elsewhere. If the SSD was limited by the SATA 1.5, it should be pegged at 150MB/sec, not 115 and 95MB/sec.

Wrong. There are far more variables here than your short-sighted suggestion. Beyond different sources of overhead, SATA II is not just SATA I with twice the signaling rate. The differences in the controller are more complex than that, and obviously have a large difference with SSDs.

SSDs in general have different characteristics and sensitives than platter-based drives, and obviously the architecture/firmware/etc of the new SATA I drive controller in the new 13"/15" MB Pro is causing throughput to fall below the theoretical maximum of 150MB/sec.

I'm sure over the next 24-48 hours, there will be HUNDREDS of similar postings of benchmark and performance data that will confirm this.

 

[Cole Slaw]

I'm not defending Apple, and also think this slower sATA stinks, but I think some differentiation is in order here.
Hardware vendors need to make choices for their customers which features and hardware to include. If a feature is (almost) not used, then it needs to go for another feature that is highly requested and will boost sales. After all, Apple is in this for the money.

There are many people that have been screaming for a card reader for a long time and Expresscard is on its way out. This sATA matter is different though, as I cannot imagine that a faster sATA would take up much more space on the logic board, or would be massively more expensive..

I see what you are saying. I agree that perhaps most people will NOT take advantage of the Expresscard slot, but as a Macbook PRO, I think these machines SHOULD have the Expresscard slot over just a SD slot.
Isn't that part of the reason you are paying more for a Macbook Pro over a regular Macbook?
Being a "Pro" machine, to me, means it ought to be as versatile as possible. Apple should spend a few dollars more on it and give it a full complement of ports, for example.
As far as Expresscard being "on it's way out", didn't Apple say the same thing about Firewire only to end up putting it back on the 13 inch?
How about putting an SD card reader in the Macbook and leaving Expresscard in the MBP?

 

[winterspan]

(moved this post)

Take a look at the data from the benchmark pics mentioned in the official Macrumors post:

Test: Sequential File Read and Write Throughput
App: HD Tune Pro 3.50
Drive: OCZ Vertex 128GB (MLC)

SATA II: 13" MBA, 13" MB, 17" MBP (and old versions of 13"/15" MBP)

Sequential READ = 225 MB/sec
Sequential WRITE = 180 MB/sec

SATA I: New 13" MBP, 15" MBP

Sequential READ = 115 MB/sec
Sequential WRITE = 95 MB/sec

The speedup from the SSD's fast random access will still be felt, making the user interface and applications snappier.
However, this will cause an enormous performance degradation when copying large files or transferring large files to and from external drives --- on the order of losing >50% of your throughput!

 

[SydneyDev]

However, this will cause an enormous performance degradation when copying large files or transferring large files to and from external drives --- on the order of losing >50% of your throughput!

And that's just with today's SSDs. The powers that be have been rushing SATA3 out the door due to unexpected advances in SSDs, and here is Apple going backwards with their SATA.

And Snow Leopard too - that should have been optimized for SSDs like Windows 7 was, but instead they have been focussing on the CPU. That's great, but the disk is the bottleneck for most things.

 

[deconstruct60]

Are you guys daft???
....
However, this will cause an enormous performance degradation when copying large files or transferring large files to and from external drives --- on the order of losing >50% of your throughput!

What external drives connect to these specific machines that have > 200 Gbp/s interface?

I certainly see the point where there are multiple devices on a SATA network and copying from one to another. However, those are different contexts. There is one and only one SATA device here.

The only faster than SATA I "drive" you can connect to is a RAM disk. Or you can copy from one pace on the SDD to another. (read / transform / write ). However, read and write speeds are asymmetric. (Likewise if the external device is also asymmetric in read/write speeds depending on which way the transfer is going. )


It is easy to come up with synthetic benchmarks that fit just right into a sweet spot. I am quite sure the drive vendor can write/choose a benchmark that can max out the bandwidth if that is the primary purpose of the program.

 

[harshw]

Even better analogy. If you buy a C7 and never go to a track where you can run it flat out. You also won't see any difference. Folks are out there pretending that their normal car activity is a F1 race car when in reality to do their normal jobs all they do is ride around on local highways and city streets. It isn't a "wrong size" problem. It is a "you pragmatically don't leverage that much bandwidth/horserpower" anyway to get your work done. Unless your job is to post xbench scores, most folks are not going to see a difference.





Or if you have a HUMMER H2-H3 and you never go off road, it doesn't really
matter that you don't have a real off road chasis underneath that "hummer-ish" body styling. It is a "talking-smack" feature that the car looks like it is a military all terrain vehicle.

And you wonder why the car companies are in trouble ?

 

[br0adband]

(moved this post)

Take a look at the data from the benchmark pics mentioned in the official Macrumors post:

Test: Sequential File Read and Write Throughput
App: HD Tune Pro 3.50
Drive: OCZ Vertex 128GB (MLC)

SATA II: 13" MBA, 13" MB, 17" MBP (and old versions of 13"/15" MBP)

Sequential READ = 225 MB/sec
Sequential WRITE = 180 MB/sec

SATA I: New 13" MBP, 15" MBP

Sequential READ = 115 MB/sec
Sequential WRITE = 95 MB/sec

The speedup from the SSD's fast random access will still be felt, making the user interface and applications snappier.
However, this will cause an enormous performance degradation when copying large files or transferring large files to and from external drives --- on the order of losing >50% of your throughput!

Ouch... bet that tire iron across the kneecaps hurts, don't it?

For those that consistently sit on that "you won't notice any real-world differences," you might not, but a ton of us will, with ease. That's greater than 50% speed hit when dealing with disk based operations... if you can't notice that sort of difference in your real-world use, perhaps you should stop spending money on such high perf hardware and just stick with some older hardware...

As I stated before, if I had grabbed a 13.3" MBP this weekend and got an SSD from Fry's here in Vegas, as soon as I got everything home I would have been doing my own benches, and if I noted that kind of performance hit and the sub-SATA I performance... man I'd return it and tell 'em it's defective (and stuff that 10% return fee in the process, bastards).

That's an INSANE hit in performance, by any measure... geezus.

 

[jgbr]

Ok Mac Mini (9400M):

MCP79AHCI = 3 GIGABIT

13MBP

MCP79AHCI = 1.5 GIGABIT

They are identical! Therefore a EFI firmware update should solve this?

SSD Benchmark = 197.28 Write MacMini
SSD Benchmark= 143.9 13MBP

 

[Evangelion]

What bugs me is the gradual loss of capability I think we are seeing in the MBPs.

Such as? Lack of ExpressCard? By that logic we should be complaining because Macs don't ship with floppy-drives anymore. Features get dropped, and new features take their place.

For example, for what they charge for a MBP, do you think it would be too much to ask to have say, 2 firewire ports (400 and 800), AND an Expresscard slot?

Sure, you might be able to get that for the money they are asking. And in return you might get slower CPU/GPU, less RAM and/or smaller HD. Or higher price.

I mean, even my 20 month old pre-unibody MBP has that.

And my ancient laptop from late-nineties has a floppy-drive and serial-port.

Why do they have to whittle that down to just a Firwire 800 port?

Because multitude of features and IO incresaes compelxity and price. Apple obviously feels that they rather invest that cost and complexity on some other things as opposed to (for example) FW400-port and ExpressCard.

you seem to think that Apple could easily pile lots of features in the machine (FW800 x2, Expresscard, maybe additional USB-port), and it would not have negative implications on other things (like price, weight, size, complexity etc.)....

That's the kind of thing that's been pissing me off lately with Apple. They cut features like that to save a dollar or two.

What exactly makes you think that it's just dollar or two that they will save?

The Macbook Pro is "Pro" in name only these days.

Oh cut the crap. Expresscard was hardly used at all. How exactly does removal of unused feature mean that the machine is no longer "pro"?

That analogy is just plain wrong. While a synthetic graphics benchmark like 3Dmark is different from running Crysis, the harddrive "benchmark" used in the data I posted shows raw data stream read and write.

But it's still not same as running actual apps. I would withhold judgement untill you see tests done with apps, as opposed to synthetic benchmarks.

SSDs in general have different characteristics and sensitives than platter-based drives, and obviously the architecture/firmware/etc of the new SATA I drive controller in the new 13"/15" MB Pro is causing throughput to fall below the theoretical maximum of 150MB/sec.

But that's not a problem of SATA 1.5, now is it? Instead of complaining about SATA 1.5, why not complain about buggy firmware instead? Hell, it might be that SATA 3.0 will become available with updated firmware.

 

[harshw]

NO... lol.. no one has the need for this speed.. in a laptop. They may "WANT" it.. they may even think they feel the new speed... but it is all subliminal. It does not translate into real world differences, except for the most demanding applications.. which (as has been said before) wouldn't be running on a laptop.

Thats right. We should just stop buying laptops every year and maybe buy every ten years. We just WANT gigahertz CPUs. We dont NEED them. People like you are the reason why Apple got away with under-performing PowerPC cpu's and mediocre hardware for so long

 

[Evangelion]

For those that consistently sit on that "you won't notice any real-world differences," you might not, but a ton of us will, with ease. That's greater than 50% speed hit when dealing with disk based operations... if you can't notice that sort of difference in your real-world use, perhaps you should stop spending money on such high perf hardware and just stick with some older hardware...

But do you see those transfer-rates in real-life apps? Benchmarks are not same thing as real-life apps. It's like 3D-benchmarks that report the pixel fill-rates of 3D-cards, and people then using that number to determine the performance of the card, as opposed to using actual apps to determine the performance.

Sure, benchmarks give a nice point of reference, but that doesn't change the fact that what matter is performance in real-life apps, not benchmarks. We do not buy computers in order to run benchmarks on them, we buy computers in order to run actual apps on them.

I would much rather see tests done with actual apps, as opposed to tests done with benchmarks.

 

[OllyW]

I am wonder if the SATA situation also applies to the high-end 15" MBP as it seems to have a different logic board akin to the 17"?

The MBP I looked at was the top model and it was SATA 1.5.

 

[deconstruct60]

As far as Expresscard being "on it's way out", didn't Apple say the same thing about Firewire only to end up putting it back on the 13 inch?

No they didn't.
Firewire -- rationale was that the large number of currently sold devices are drifting away from FW. (going USB, etc. For example Video cameras. ). That completely ignores the factor the currently deployed/owned devices that do have it. In short, it is in use or could be used right now without purchasing anything.

Express Card -- Apple counted the currently deployed devices. Their number was less than 1% of the current MPB owners were using it. "Use it or loose it". That's different from "I know you are using that but it is too old" or "just buy the newer stuff". Apple recognized several folks with higher end problems using the card in higher numbers. and decided that the 17" was a better fit for their needs and that once collected there would reflect more significant numbers/percentages; so kept it ( and moved the price of the 17" down).

The first ignored the number of folks who owned stuff. The second uses that as the factor in the decision. Mystified why folks keep throwing them into the same bucket of decision making process. [ in additional to chucking the battery compartment allow flexibility to move things around to make space on he side panel. ]

If the Express Card utilization had been higher (or Apple radically miscounted ) would have been in same boat. ( Apple would have had to make different set of trade-offs.)

 

[Icaras]

The MBP I looked at was the top model and it was SATA 1.5.

Well actually, the top model does have SATA 3.0. The 17".

 

[OllyW]

Express Card -- Apple counted the currently deployed devices. Their number was less than 1% of the current MPB owners were using it. "Use it or loose it".

They actually said single digit percent, so it could be anything between 1 and 9%.

 

[L0s7man]

That does it. I'm not buying the damn Apple. No chance.

The latest "update" is just simply downgrade. They were happy to say "We've cut the price", but they didn't mention how much hardware has gone out of the window with it.

I, for instance, was looking forward to getting Intel X25M ssd. Now, it seems like a bad idea.

And I would have a use for the ExpressCard (64GB ssd, for storing stuff like photos, music, etc; nicer than external or usb stick because it sits in EC slot all the time)

Not to mention the bloody mirror screen.

Or they've said that they've updated the MBP screen. Rubbish; it's probably about the Macbook that has name change to Macbook Pro. They've given it a better screen than the **** they used before and now they claim how the screen got better.

Apple can kiss my ass.

 

[OllyW]

Well actually, the top model does have SATA 3.0. The 17".

I was replying to a question about the top 15" model.

 

[turkay]

These downgrades are not special for Apple, all the brands do something like that. Because this is marketing. They will never give you a 100% of your dream machine. They will always give something better but not perfect, this is the point of marketing. I don't like the downgrades though.

 

[jgbr]

oh come on this is a EFI firmware update problem/soultion

 

[turkay]

oh come on this is a EFI firmware update problem/soultion

Well it is not, until Apple states otherwise.

 

[Evangelion]

They actually said single digit percent, so it could be anything between 1 and 9%.

Actually, they said "less than single digit percent"...