MBA SSD facts and thoughts

[Dammit Cubs]

I think the assumption that 128GB SSD will be 2x faster than 64GB is completely ridiculous and most likely not true. It's based on how the controller manages each flash component. if you take smaller Flash memory and put then in parallel (having a controller manage the throughput), you could create dramatic performances. (as shown from several demos from past years)

That said, it depends out the SSD's are places on the board, how many flash chips are there for the 64, 128, 256. Also in each flash unit, how many flash die are there? How did they stack it? It all matters in figuring out how much throughput each unit has and then more importantly, how much throughput the controller can handle.

 

[daleski75]

I think this could be true if they did the same as on the new Vaio Z with 2 x 64gb SSD's striped together for performance.

Otherwise I think they will be the same speed.

 

[omenatarhuri]

Is the 128 SSD really faster than 64 or just a myth? This will change my buying decision if true.

I don't think it's true. In HDDs that is true because the disc needs to spin to correct place and on more dense platters the physical distance is shorter. Also large platters lessen the likehood of needing to change platters to get to the data.

These facts obviously don't apply to flash storage and the access time is usually < 2ms.

 

[diddl14]

I just realized that although the Toshiba T6UG1XBG SDD Controller is reported to support TRIM, the Apple flash-card reports itself as SATA device without TRIM support. So even after installing Windows 7 as alternative OSX, the drive would likely still not support TRIM.

This might be just some customized reporting of drive parameters (drive type, size, features) or TRIM could have been removed from the T6UG1XBG firmware. Either way, this would make it very unlikely that TRIM will be supported in a future OSX release. Doesn't make sense (even for Apple) to cripple the hardware if TRIM is planned for a future OSX release...

SATA Config of 11.6" MBA with 128GB flash storage:

(sorry for the German, taken in Austrian retail store..)

 

[Pressure]

Performance is in some degree still depending on the memory controller and the amount of NAND devices used.

Less NAND flash on board means lower sequential write speeds as there are less lanes (e.g. NAND flash) to write to.

This also explains why the Intel V-25 40GB is slower than the Intel X-25 80GB because the V-25 only has 5 NAND flash chips onboard vs the 10 of the X-25.

A higher channel implementation with more NAND flash chips is going to have faster speeds.

 

[hachre]

I just realized that although the Toshiba T6UG1XBG SDD Controller is reported to support TRIM, the Apple flash-card reports itself as SATA device without TRIM support. So even after installing Windows 7 as alternative OSX, the drive would likely still not support TRIM.

This might be just some customized reporting of drive parameters (drive type, size, features) or TRIM could have been removed from the T6UG1XBG firmware. Either way, this would make it very unlikely that TRIM will be supported in a future OSX release. Doesn't make sense (even for Apple) to cripple the hardware if TRIM is planned for a future OSX release...

[HUGE PICTURE]
(sorry for the German, taken in Austrian retail store..)

It always says Trim: No there because Mac OS doesn't support trim yet. It does this also on my Intel SSD X-25 Postville which definitely supports trim...

INTEL SSDSA2M160G2GC:

Capacity: 160,04 GB (160.041.885.696 bytes)
Model: INTEL SSDSA2M160G2GC
Revision: 2CV102HA
Serial Number: CVPO00430159160AGN
Native Command Queuing: Yes
Queue Depth: 32
Removable Media: No
Detachable Drive: No
BSD Name: disk0
Medium Type: Solid State
TRIM Support: No
Partition Map Type: GPT (GUID Partition Table)

 

[diddl14]

It always says Trim: No there because Mac OS doesn't support trim yet. It does this also on my Intel SSD X-25 Postville which definitely supports trim...

Great to learn that the OSX System Profiler obviously doesn't profile the hardware (or at least doesn't report it correctly)

For anyone else interested in this topic, here some additional background from AppleInsider.

 

[hachre]

Great to learn that the OSX System Profiler obviously doesn't profile the hardware (or at least doesn't report it correctly)

For anyone else interested in this topic, here some additional background from AppleInsider.

Yea I think it's been a placeholder there for a while but it always shows no until Mac OS gets trim support...

 

[ImperialX]

Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_1 like Mac OS X; en-us) AppleWebKit/532.9 (KHTML, like Gecko) Version/4.0.5 Mobile/8B117 Safari/6531.22.7)

Just install Windows 7 onto a separate partition, and every few months, format the Mac partition with NTFS, boot into Windows 7, run the TRIM command, then reformat it back to HFS+ and install OS X. Problem solved.

 

[JodyK]

Actually, that might have been the case in the early stages of SSD drives. But if you buy a 100GB or 200GB drive of the same sort today, there won't be any noticeable difference.

I completely disagree http://www.crucial.com/store/listmodule/SSD/~2.5-inch Solid State Drive~/list.html look at the write speeds directly from Crucial.

 

[Pressure]

I completely disagree http://www.crucial.com/store/listmodule/SSD/~2.5-inch Solid State Drive~/list.html look at the write speeds directly from Crucial.

Read this single post.

 

[Transporteur]

I completely disagree http://www.crucial.com/store/listmodule/SSD/~2.5-inch Solid State Drive~/list.html look at the write speeds directly from Crucial.

That's because the smaller drives use less chips, hence less channels. If you chose drives with the same amount of chips, but lower capacity, all channels can be used, which is why the speeds won't differ considerably.
In case of the Toshiba SSD Apple uses in the MBA, this seems to be the case (4 chips for all sizes).

As I already mentioned, most SSD manufactures chose smaller capacity chips, rather than just reducing them, to maintain the performance.

 

[methodlala]

Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_1 like Mac OS X; en-us) AppleWebKit/532.9 (KHTML, like Gecko) Version/4.0.5 Mobile/8B117 Safari/6531.22.7)

Just install Windows 7 onto a separate partition, and every few months, format the Mac partition with NTFS, boot into Windows 7, run the TRIM command, then reformat it back to HFS+ and install OS X. Problem solved.

Can someone with a MBA running windows 7 support this statement by running this command:

fsutil behavior query DisableDeleteNotify

If it spits back 0, it means trim is supported and that win 7 will indeed go to work on the drive when its idle. Thanks!

 

[KPOM]

Can someone with a MBA running windows 7 support this statement by running this command:

fsutil behavior query DisableDeleteNotify

If it spits back 0, it means trim is supported and that win 7 will indeed go to work on the drive when its idle. Thanks!

If I saw this 2 hours ago I'd have had my MacBook Air in front of me. I'll boot into Boot Camp tonight when I get home and see if it works.

Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_1 like Mac OS X; en-us) AppleWebKit/532.9 (KHTML, like Gecko) Version/4.0.5 Mobile/8B117 Safari/6531.22.7)

Just install Windows 7 onto a separate partition, and every few months, format the Mac partition with NTFS, boot into Windows 7, run the TRIM command, then reformat it back to HFS+ and install OS X. Problem solved.

Would running the recondition routine in DigiLloyd's DiskTester application have a similar effect" I was considering purchasing it for my Rev B shortly before the new Air came out (which rendered the issue moot for another few months or so since obviously I got a brand new drive with the Rev D).

 

[KPOM]

Can someone with a MBA running windows 7 support this statement by running this command:

fsutil behavior query DisableDeleteNotify

If it spits back 0, it means trim is supported and that win 7 will indeed go to work on the drive when its idle. Thanks!

I had to go into safe mode to get the command to work, but it did and I'm happy to report that it came back with 0. So it seems like we'll get TRIM support as soon as OS X supports it.

 

[methodlala]

I had to go into safe mode to get the command to work, but it did and I'm happy to report that it came back with 0. So it seems like we'll get TRIM support as soon as OS X supports it.

Sorry! Should have mentioned to run the command as an admin. Nevertheless, this is great news, thanks for your help!

 

[KPOM]

Sorry! Should have mentioned to run the command as an admin. Nevertheless, this is great news, thanks for your help!

I have only one username set up, and it is an administrator account. Oddly, though, I still got the error message when I opened up a command line. I rebooted into Safe Mode with Command Prompt and it worked fine.

 

[Dammit Cubs]

What people should keep in mind is that larger SSDs are faster. We need to see more benchmarks, but i bet the 128gb drive will show significantly better numbers than the 64gb in the 11.6", which itself is already very fast.

This is another reason why I think the stock 13.3 is the one to go for. Especially if you only stick with 2gb of ram, having a faster ssd will make swapping less painful.

This was already debunked. the 128GB perform the same as the 64GB. That makes sense because regardless of the size, there are 4 die on the board. wither 4 16GB flash die or 4 32GB flash die so the its the same number of I/0's in terms of parallel data reading.