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

[Poirot818]

so if the power savings through a 1.5 GBit interface are significant

Then why not do the same for the 17" MBP? Surely the 17" MBP needs all the power savings it can get to keep up battery life.

The 17" MBP keeps its 3 Gbit/s SATA connection and has an express card slot. The other MBPs have a 3 Gbit/s SATA connection that's capped at 1.5 Gbit/s and an sd slot instead of an express card slot. Draw your own conclusions from that.

 

[justit]

What if they crimped this to sell an upgrade SSD option at the apple store with the full 3 SATA speed using apple's pricing?

 

[Anuba]

FWIW, I stumbled upon a similar discussion regarding certain Lenovo models, and found this post...

I've investigated this issues at length through engineering and have received the following explaination of SATA data rates available on the Santa Rosa ('61) and current Montevina based systems (T400/T500, W500, W700, etc)...

"For Santa Rosa-based systems, the Intel ICH8 supports a SATA bus speed of up to 3.0 Gb/s. Lenovo made a design decision to prioritize maintaining compatibility with Ultrabay disk drives, which are connected via a SATA-to-PATA conversion chip which could not handle a 3.0 Gb/s SATA bus speed reliably. Therefore the system was standardized to 1.5 Gb/s.

In testing rotating media drives, our measurements show data throughput difference between 1.5 Gb/s and 3.0 Gb/s bus speed is less than 5% since the drive mechanics are the limiting throughput factor, rather than the SATA bus itself.

For those customers who choose to purchase an after-market SSD drive capable of SATA bus speeds up to 3.0 Gb/s, the system will interface with them at 1.5 Gb/s. Lenovo's official position is that the Santa Rosa systems are working as designed.

The Montevina based systems which began shipping last year have direct SATA interfaces for both drive bays and are enabled at a system level for SATA bus speeds of 3.0 Gb/s performance. Current Lenovo drives have firmware set to 1.5 Gb/s data rates.

Exchanging these drives for after-market drives which support SATA bus of 3.0 Gb/s should provide for the higher data rate at the overall system level. Again, it should be noted that our performance measurements show less than 5% performance improvement between 1.5 Gb/s and 3.0 Gb/s SATA bus speeds for rotating drives, since the drive mechanics are the limiting throughput factor, rather than the SATA bus itself.

After-market SSDs which support SATA bus speed of 3.0 Gb/s will operate at that bus speed. Depending on the data transfer test method used, your actual data throughput from a 3.0 Gb/s SATA bus speed should be 220-250 MB/s and about 90-120MB/s throughput when running on SATA bus of 1.5 Gb/s. This is due to the bus signaling used for the SATA bus, as well as overhead for error checking."

So it seems that the bus speed has an effect (albeit marginal) even on HDDs.

 

[314631]

Yes ... your new $2k Macbook Pro is worthless. Cancel your order and buy an Acer PC Laptop.

I just cancelled an order for a $3200 $17" MBP with extras because of this issue. Sorry I'm paying a premium price I expect a premium product. We were always led to believe PC manufacturers put in shoddy/cheap components to keep prices low. Apple is now doing the same and charging a huge premium. I'm going back to Dell's I think.

Very sad news.

 

[gnasher729]

Sure: http://www.techreport.com/articles.x/17010/7

See the red bar (longer is better)? That's Apple's stock 5400 RPM drive. The two green bars below it are the Seagate Momentus drives, one of them is Apple's stock 7200 rpm 2.5" drive.

The Momentus 7200.4 scores better in long sequential reads, but that's about it.

I am always very suspicious of benchmarks for hard drives, because often benchmarkers are not very clever and forget one very important thing about hard drives: Hard drives don't have one "speed". The read/write speed depends on which tracks of the hard drive you are reading. On the outside tracks, each rotation of the drive lets you access many more bytes than a track on the inside of the drive. So a drive could have a read speed "from 60 GB/sec to 30 GB/sec": When the drive is empty and you read the tracks on the outside, the same drive might read 60 GB/sec, and when the drive is full and you read tracks close to the centre of the driver, you only get 30 GB/sec.

Now if you take two brand new drives, one 320 GB, the other 500 GB, create two partitions on each drive, 250+70 on the first, 250+250 on the second, and you make speed tests on the second partition, then the second partition is close to the center and very slow on the first drive, but still quite far on the outside and a lot faster on the second drive.

When both drives are almost empty, the 320/7200 drive should be faster than the 500/5400 drive (if everything else is equal). As you copy one GB after the other to each drive, the 320GB fills faster and its speed will go down quicker. So if you intend to store 250-300 GB, the 500/5400 drive will end up faster.

 

[Glideslope]

Having worked with hardware developers who used every trick in the book to reduce power usage, there is a very simple rule: Power consumption grows with the square of speed. In the case of a drive interface, the power consumption would run over a shorter time, so at double the speed, the amount of energy used to transfer one Megabyte of data would grow linear with the speed of the interface.

Don't attribute to malice what could be explained by stupidity, or possibly by someone taking into account things that you don't know about. Now 99% or more of Mac users are _not_ going to replace the drive in their MacBook Pro with an SSD drive that doesn't come from Apple, and those users won't notice a difference outside benchmarks, so if the power savings through a 1.5 GBit interface are significant, and I have no idea how much you energy would typically be used by the SATA interface, then reducing the interface speed is indeed the right decision.

It should have also been made known at WWDC.

 

[deconstruct60]

presumably, but if Apple is touting a five year lifespan for these notebooks, who wants to be limited by 1.5 Gbps with the SSDs that will exist then (let alone two of three years).

Five years for the batteries or the computer itself? Applecare only goes 3 years so as far as putting their money where their mouth is... it is more like the battery has about a 5 year life. So by the time the battery is noticeably dying off you'd probably buy a new computer ( can the battery can be recycled along with the computer you are "trading in".)


However, the more times you cycle the battery the shorter its lifetime is going to be. The more power you draw will increase the number of times you cycle the batter. If 3 GB/s is a higher power draw ( extremely likely) than have taken the hit on that 5 years.

On the other hand, replacing a rotational hard drive with a SSD one would help save power later on in your laptop's life. Similar to reason outlined above that would stretch out your battery lifetime.

 

[Chupa Chupa]

I just cancelled an order for a $3200 $17" MBP with extras because of this issue. Sorry I'm paying a premium price I expect a premium product. We were always led to believe PC manufacturers put in shoddy/cheap components to keep prices low. Apple is now doing the same and charging a huge premium. I'm going back to Dell's I think.

Very sad news.

But the 17" is unaffected. The 1.5Gb cap is only on the 13/15" MBPs.

 

[Anuba]

I just cancelled an order for a $3200 $17" MBP with extras because of this issue.

But this issue doesn't affect the 17" MBP. Why on earth would you cancel it then... out of sympathy for owners of the 13" and 15" models?

 

[TEBnewyork]

I just cancelled an order for a $3200 $17" MBP with extras because of this issue. Sorry I'm paying a premium price I expect a premium product. We were always led to believe PC manufacturers put in shoddy/cheap components to keep prices low. Apple is now doing the same and charging a huge premium. I'm going back to Dell's I think.

Very sad news.

Well, If you had read the information before you cancelled you would have found that the 17" MBP is not impacted by this only the 13" and 15" are at 1.5.

 

[charlituna]

does customer satisfaction mean NOTHING to them?

yes because the 0.001% of folks that would actually be effected by this change is THE key to Apple's success. they should always do what will benefit those folks.

perhaps Steve should just hand you the throne cause you are so much smarter than he ever was. would that make you happy

 

[harshw]

Unless you're doing high-performance scientific computing across a HUGE dataset (which you will not be doing on a laptop) it is extremely unlikely you'll ever see anything even remotely approaching 1.5Gbps, much less 3Gbps.

And you're the smart one eh ? Well I do just that - on a laptop. And please don't tell me ' Ohh ... this is not for laptops '. So what is the MacBook Pro for then ? Watching Quicktime trailers and Apple v/s PC ads ? Some of us use laptops for a living including running high performance scientific computing on them

 

[SamCpp]

Oh that's funny... 'cause your signature states that you're using a 5400rpm HD. Even if you plan on upgrading to SSD in the future, I'm going to go out on a limb and say you won't ever see a difference in performance.

Do you have any idea how fast 1.5Gb/s is? 3Gb/s interfaces on laptops are for spec-whores.

1.5Gb/s (gigbit per second) is approximately 150MB/s (megabyte per second). Most if not all hard drives nowadays burst at over 150MB/s. Most SSDs average over 150MB/s read.

i.e. not just spec-whores are getting affected here. Anyone who uses new hardware will.

 

[deconstruct60]

Then why not do the same for the 17" MBP? Surely the 17" MBP needs all the power savings it can get to keep up battery life.

Bigger battery means more power (and cells) to throw at problems and still eek out the 8 hours. If haven't noticed the 17" are bigger than the 15"-13" models. ;-) [ more cells also means distributing the cell cycles over more units. The 13 and 15 have less cells so each one has a higher duty cycle. ]


Same rational would cap 17" below the 3Gz CPU option too.

 

[8theapple]

I just cancelled an order for a $3200 $17" MBP with extras because of this issue. Sorry I'm paying a premium price I expect a premium product. We were always led to believe PC manufacturers put in shoddy/cheap components to keep prices low. Apple is now doing the same and charging a huge premium. I'm going back to Dell's I think.

Very sad news.

17" is not affected and runs SATA 3. Good job.

 

[turkay]

And you're the smart one eh ? Well I do just that - on a laptop. And please don't tell me ' Ohh ... this is not for laptops '. So what is the MacBook Pro for then ? Watching Quicktime trailers and Apple v/s PC ads ? Some of us use laptops for a living including running high performance scientific computing on them

BTW there's gonna be no "non pro" notebook in the near feature IMO.

 

[deconstruct60]

1.5Gb/s (gigbit per second) is approximately 150MB/s (megabyte per second). Most if not all hard drives nowadays burst at over 150MB/s. Most SSDs average over 150MB/s read.

i.e. not just spec-whores are getting affected here. Anyone who uses new hardware will.

Burst. You wonder why burst and SSD are about the same???? Could it be that you are reading out of he RAM cache of the HD when "bursting"?

How many times does burst happen under real world workload conditions?
Especially on the laptop with just one user. ( two programs access the same file often or rarely ? )

Got any real world benchmark simulations where these burst conditions are sustained for a real work sized files and file systems?

 

[farkasam]

Please,

Most of the "power" users posting in this forum aren't even going to notice any difference with 1.5g SATA. Don't be such narcissists, Apple makes products that have to straddle the line between "pro" and "PRO" users. If you're so hardcore, build your own and stop whining!

 

[harshw]

BTW there's gonna be no "non pro" notebook in the near feature IMO.

Quite right. And I was actually waiting for a 13" pro laptop - preferably one with a faster CPU. After putting in a Vertex on my 13" uMB I saw that tasks which were IO bound suddenly became CPU bound. The Vertex can achieve 200mb/sec - 240 mb/sec reads and 140mb/sec - 170 mb/sec writes. Having a faster CPU was fantastic but with the SSD speeds capped to 150 mb/sec, that faster CPU wouldn't matter anyways

It's sad as the 13" MBP is otherwise a beautiful machine. I find the 15" too big for my regular travel and usage style.

 

[frou]

Sure: http://www.techreport.com/articles.x/17010/7

Thanks for the link. Though, looking at the roundup, I wouldn't say it is cut and dry to call the Seagate slower.

And you're the smart one eh ? Well I do just that - on a laptop. And please don't tell me ' Ohh ... this is not for laptops '. So what is the MacBook Pro for then ? Watching Quicktime trailers and Apple v/s PC ads ? Some of us use laptops for a living including running high performance scientific computing on them

I think it's become obvious that Pro currently doesn't mean much beyond design and fit/finish (i.e. the actual components are often not best in class)

 

[Eidorian]

I can't imagine Apple not fixing this.

Well I've had a SATA 3.0 drive for ages and I'm on the latest firmware.

I don't think Apple is ever going to fix it unless we raise an uproar about it.

 

[harshw]

Burst. You wonder why burst and SSD are about the same???? Could it be that you are reading out of he RAM cache of the HD when "bursting"?

How many times does burst happen under real world workload conditions?
Especially on the laptop with just one user. ( two programs access the same file often or rarely ? )

Got any real world benchmark simulations where these burst conditions are sustained for a real work sized files and file systems?

Try running real world programs like SPSS, SAS, Mathematica etc that operate on large datasets and you'll see the difference. They are not 'simulations' - they're real and the Apple laptops are usually well liked in the scientific computing community ( nice GUI, BSD Unix etc ). This recent change in specs is not going to go well with that community

 

[turkay]

I don't think Apple is ever going to fix it unless we raise an uproar about it.

I don't think that's gonna make any difference. If some higher ranked Mac people, like Rob Griffiths moans about it then apple may take the issue serious or aware of it.

 

[Furrybeagle]

Bigger battery means more power (and cells) to throw at problems and still eek out the 8 hours. If haven't noticed the 17" are bigger than the 15"-13" models. ;-) [ more cells also means distributing the cell cycles over more units. The 13 and 15 have less cells so each one has a higher duty cycle. ]


Same rational would cap 17" below the 3Gz CPU option too.

Yes, but the 2009 13" and 15" uMBPs still have larger batteries than the 2009 13" and 15" uMBPs. Yet they still cut features? I'm confused... what's more important, 30 more minutes (at most), or faster interfaces?

Seriously, you think the Express Card slot is sucking up power when nothing is in it? Does the hard disk interface really take up so much power that it needs to be downgraded? What is this crap?

 

[Anuba]

Thanks for the link. Though, looking at the roundup, I wouldn't say it is cut and dry to call the Seagate slower.

No, it's faster for certain operations... but the bottom line is that the fastest 5400 RPM drive in the test (the WD Scorpio Blue) beat the crap out of the lot of them.. well, except the SSD. You'd think that the only 7200 rpm drive in the test would dance all over the 5400 drives, and this is probably the assumption most people make when they choose the 7200 rpm option.