Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Which connector is your new unibody Macbook pro

  • Sata I - 1.5Gbit

    Votes: 218 69.6%
  • Sata II - 3.0Gbit

    Votes: 95 30.4%

  • Total voters
    313
Am i missing something here?

how are you guys testing the speed of the hdd connection with connections that are slower than 3gb/s?

i mean fw 800, 400, usb, sd card, are all way slower than 3gb/s.

i would think the only true way to test this is through an esata connection.

i had a expresscard slot with an esata connection and it was definitely 3Gb/s, so that would work. i think this is why this card didn't work on my computer. it baffled the genius bar and apple ( they personally called me and took my computer in and replaced the logic board, and it didn't fix anything) as to why it wouldn't work. as my computer has a sata 1 connection and essentially i tried to connect a sata 2 connection to it through the express card.
 
Isn't the primary benefit of the SSD technology available on sequential reads and reads in general? I know the specs always show 2 very different numbers for both.

Your test is geared toward sequential writes which would be slower on the SSD.

Cheers,

The primary speed benefit of SSD is avoiding rotational latency problems. So substantially better on random I/O (especially random and/or concurrent access ) workloads. The read sequential I/O is coming out better for SSD deciveces because their native block size is closer to the large block sized used in several of these benchmarks. That large block size is what will hurt in the write speeds when blocks are smaller than the native size ( so again the bigger block writes better , but the smaller blocks are more slower on SSD). Unless a program is mucking with its own blocks most of the time the filesystems have 4K, 8K blocks.

If you are clever and lay the data onto a hard disk, you can get rid of rotational latentices. You need an mostly empty disk. And you laid down the data sequentially so that when you finish with one cylinder of data the next sequential data cylinder is right next to the one you just used. That minimizes seek time and if there is no concurrent workload request to make access more random you can go faster. Under real life usage, that is very rare.
 
Am i missing something here?

how are you guys testing the speed of the hdd connection with connections that are slower than 3gb/s?

i mean fw 800, 400, usb, sd card, are all way slower than 3gb/s.

i would think the only true way to test this is through an esata connection.

i had a expresscard slot with an esata connection and it was definitely 3Gb/s, so that would work. i think this is why this card didn't work on my computer. it baffled the genius bar and apple ( they personally called me and took my computer in and replaced the logic board, and it didn't fix anything) as to why it wouldn't work. as my computer has a sata 1 connection and essentially i tried to connect a sata 2 connection to it through the express card.


well if the firewire 800 will load osx almost as fast as the sata II interface...then sata I would not make a significant difference....

i am admitting i am wrong here, i apologize to the people who i argued with that sata II would make a significant difference...it simply does not.

the fact is im running my mbp on fw800 and i completely forgot until i looked over to still see it plugged in...its simpy as fast doing what i normally do....sorry if that disappoints you...
 
Am i missing something here?

how are you guys testing the speed of the hdd connection with connections that are slower than 3gb/s?

...

i would think the only true way to test this is through an esata connection.

No. The primary thing you need to test 3GB/s is a storage drive that is faster than 1.5GB/s in access.

Conceptually if dropped 3-4 hard drives on the same SATA bus you could do it also (e.g., use RAID 0 and striping to create a virtual device that matched the SSDs avoiding rotational problems) , but the test would have to account for other variables you'd introduce.

An ExpressCard SATA test would only tell you about the SATA interface on the card. Not the one inside the machine. That's is a on a different PCI connection. Since you can only put one hard drive on the internal bus, one, faster than SATA I, drive is what you need.
 
I was mostly comparing random reads with bigger chunks of data, partly because earlier someone posted results pointing to as little as 8KB being enough to saturate SATA 1.5 Gb/s connection, and applications usually contain many files bigger than that.

If you have 12-14 cores all making 8KB requests (and the OS kernel is multi-threaded) then it would be much easier to get SATA 1.5 Gb/s. The number of requests on same wire is a factor.

Likewise, could have 2-3 programs concurrently doing sequential reads, but effectively to the targeted drive that is closer to random impact than sequential (again if the request are interleaved by the OS and drivers. )
 
Can anyone with a SSD and new MBP post a video of opening up applications, boot time, etc? It would be nice to see how daily life is being affected.

Doubt my tests add or help anything to this discussion but thought I'd post anyways.

Brand new 2.26 13" Macbook Pro with 4GB Ram.

Tested with stock Hard drive and brand new 128GB Corsair SSD.

It took 36 minutes to install OSX on the Corsair SSD. Not sure if this is fast/slow/average. This SSD is not known for being the fastest on the market, but it is top rated as far as reliability. After using a SSD in my Dell Mini for the past few months, I can never ever ever go back to a hard drive in a laptop again.

Each test I did 5 times and averaged all the results out.

Stock Harddrive in new 13" Macbook Pro with only the pre-loaded programs installed, and all updates applied. The only other thing I did was delete Safari and Mail and install Firefox and Thunderbird.

Boot Time: 34.3 seconds
Shutdown Time: 4.1 seconds

New Corsair 128GB SSD with everything exactly the same, same pre-loaded programs, Deleted Safari/Mail and installed Firefox/Thunderbird.

Boot Time: 22.6 seconds
Shutdown Time: 2.8 seconds

Where the SSD really shines is launching apps, instead of the 3-4 dock bounces I got on my iMac when launching programs like Excel and Photoshop, every single program I have tried has launched on the first bounce.
 
But I was wrapping up from WWDC and flying from West to East coast, so just haven't had much time to sit down.

arn

Did they make you stand on the plane? :p

Cheers for getting this story up, especially when a lot of other places have been ignoring the news. Hopefully we'll get an answer sooner and a fix as well.
 
Dueces, thank you so much for the testing results!

The Corsair SSD is in fact a great SSD (It think it's only 90 read, 70 write) Very slow compared to other, but it apparently handles daily activities to perfection and in known for very fast boot speeds.
 
Dueces, thank you so much for the testing results!

The Corsair SSD is in fact a great SSD (It think it's only 90 read, 70 write) Very slow compared to other, but it apparently handles daily activities to perfection and in known for very fast boot speeds.

actually there are 2 corsair drives

the p and s series

P128 - 220 read / 200 write
S128 - 90 read / 70 write
 
actually there are 2 corsair drives

the p and s series

P128 - 220 read / 200 write
S128 - 90 read / 70 write

Yeah, and the P128 just came out. The S128 is the one everyone (currently) raves about for being reliable and having fast boot/app times. I'm sure the P128 is even better.

But the S128 might actually be an excellent option if Apple doesn't fix the 1.5 gig issue.
 
Yeah, and the P128 just came out. The S128 is the one everyone (currently) raves about for being reliable and having fast boot/app times. I'm sure the P128 is even better.

But the S128 might actually be an excellent option if Apple doesn't fix the 1.5 gig issue.

i agree, i rocked the s128 for months on my classic mbp , and then on my unibody for a few months
 
The primary speed benefit of SSD is avoiding rotational latency problems. So substantially better on random I/O (especially random and/or concurrent access ) workloads. The read sequential I/O is coming out better for SSD deciveces because their native block size is closer to the large block sized used in several of these benchmarks. That large block size is what will hurt in the write speeds when blocks are smaller than the native size ( so again the bigger block writes better , but the smaller blocks are more slower on SSD). Unless a program is mucking with its own blocks most of the time the filesystems have 4K, 8K blocks.

If you are clever and lay the data onto a hard disk, you can get rid of rotational latentices. You need an mostly empty disk. And you laid down the data sequentially so that when you finish with one cylinder of data the next sequential data cylinder is right next to the one you just used. That minimizes seek time and if there is no concurrent workload request to make access more random you can go faster. Under real life usage, that is very rare.

Great explanation. In addition to these issues do SSD drives still have a finite lifespan whereby they just cease to function? Or was that just something related to the early drives?

Cheers,
 
Complete Speculation

I think you only need to scan through the last two or three pages to get the overview. Not much has changed in the last 8 hours. It seems that mostly we're just getting new persons who are either confirming what has already been posted/known or new contributers who are arguing the same points that have already been beaten to death.

Basically:

1.) It appears nearly certain that the new 13" and 15" MacBook Pros are all reporting a SATA interface running at 1.5Gb and not the faster 3.0Gb rate that has been in pretty common use for the last few years. These new models have the Secure Digital (SD) slot and also appear to have redesigned motherboards.

2.) Those who are using standard hard disk drives will probably see no difference in performance. If that is you, you can stop reading now.

3.) Benchmarks on FAST solid-state drives (SSDs) are showing a decrease in RAW disk i/o transfer rates on these same systems (in comparison to the previous generation MacBook Pros and MacBooks).

4.) The largest differences in the benchmark results seem to be in large, sequential disk READS (one of the traditional strengths with SSDs).

5.) To the best of my knowledge, no one has done any test with REAL-WORLD operations to show that the user experience (i.e. "performance") will be decreased with the 1.5Gb SATA interface. That is to say that thus far we've only seen benchmarks done with RAW disk i/o benchmarking tools.

6.) No one really knows why this has been done and no one knows whether it can be fixed with a software/firmware update (it may or may not be able to be fixed).

Nice summary. Thanks for the information. Another reason not to buy or recommend this generation of MBPs.

I am going to speculate that this was done intentionally by Apple because they are planning on releasing a new MBP line with SSDs in the next revision and this would give them more marketing hype with regard to speed increases for the next revision. If Apple reintroduces the SATA 3.0 connection in the next revision with SSDs, then the increase in sequential writing will be a selling/upgrading point for those who want/use SSDs in their MBPs.

This is purely speculation, but why else would Apple redesign and reduce the I/O for MBP hard disks? Surely cost reduction can't be great enough to justify this change alone.
 
2009 MBP 2.93 Shows 3 Gigabit

Currently at the Apple just looked at all models and only the MBA, whitebook and 17"MBP have 3.0Gb...everything else is 1.5

1. 15" 2.8
2. 13" 2.53
3. MBA 2.13
4. 15" 2.53
5. MBA 1.86

This is of course not the newest one, it still has the Express slot. But it is showing NVidia MCP79 AHCI with my Intel SSD listed under that. It shows 3 Gigabit.
 
I am going to speculate that this was done intentionally by Apple because they are planning on releasing a new MBP line with SSDs in the next revision ...

Agreed, I posted earlier that it could be a future, though quite, upgrade option on the apple store with apple's own SSD pricing and SATA 3/Gbps
 
Just talked to Apple Tech Support

I just got the phone call from the Tech Support here in Japan.

I've CTO my macbook pro 15" with

3.06ghz, 500gb 7200rpm.

I asked if the one I ordered was only 1.5gbps and would be changeable to 3.0gbps if I buy SSD in the future on my own.

and another question was if I order 256gb SSD with an apple, would it be 3.0gbps.


The answers were as followed:

1. Right now it is being shipped as 1.5gbps, and they are not sure whether it would be software/firmware upgradable to 3.0gbps in the future as this is not confirmed as of yet.

2. If you order 128gb or 256gb SSD as a CTO, you will be sure to get the 3.0gbps speed.


So I just changed my order from 500gb 7200rpm to 256gb SSD.

I also asked whether it was only Toshiba for the SSD, but the answer was "since we deal with a lot of different companies, we cannot confirm the brand of the SSD"


I hope this is helpful and be posted up somewhere as an info.
 
I just got the phone call from the Tech Support here in Japan.
The answers were as followed:

1. Right now it is being shipped as 1.5gbps, and we are not sure whether it would be software/firmware upgradable to 3.0gbps in the future as this is not confirmed as of yet.

2. If you order 128gb or 256gb SSD as a CTO, you will be sure to get the 3.0gbps speed.

Helpful - very
Even more annoying - YES (if the only way to get 3.0 is return the laptop and buy a BTO with SSD even though I just bought one)
 
Does anyone know if it's a DIP switch or something, because it seems to me that they wouldn't install a special OS for each CTO item? Or maybe.
 
I just got the phone call from the Tech Support here in Japan.

I've CTO my macbook pro 15" with

3.06ghz, 500gb 7200rpm.

I asked if the one I ordered was only 1.5gbps and would be changeable to 3.0gbps if I buy SSD in the future on my own.

and another question was if I order 256gb SSD with an apple, would it be 3.0gbps.


The answers were as followed:

1. Right now it is being shipped as 1.5gbps, and they are not sure whether it would be software/firmware upgradable to 3.0gbps in the future as this is not confirmed as of yet.

2. If you order 128gb or 256gb SSD as a CTO, you will be sure to get the 3.0gbps speed.


So I just changed my order from 500gb 7200rpm to 256gb SSD.

I also asked whether it was only Toshiba for the SSD, but the answer was "since we deal with a lot of different companies, we cannot confirm the brand of the SSD"


I hope this is helpful and be posted up somewhere as an info.

I believe what you say, but I don't believe the person you talked to knows what they are talking about. This would mean different hardware / firmware for each SSD CTO. I just don't believe this at all. There's no way they would increase their costs by having two different part numbers just to support CTO SSDs.

I sure hope someone gets one soon so they can report what it says.
 
2. If you order 128gb or 256gb SSD as a CTO, you will be sure to get the 3.0gbps speed.

I hope this is helpful and be posted up somewhere as an info.

man, this is very lame

still, thx for ur support and info.
it certainly confirmed some of our speculations

much appreciated.
 
Lots of speculation over the weekend here - does anyone think that Apple will comment or communicate on this tomorrow?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.