[iMac 2011] BTO Internal SSD or external TB SSD, now that it's bootable?

[Sylonien]

Now that its' finally confirmed.

Thunderbolt Supports Booting From External Disk

Originally waited and waited for Apple SSD as the waiting time was 4-6 weeks, now it's gone down to 2-4 days. Almost went and ordered but now that I read it's bootable. Not sure which one to get.

I was gonna put up with the slow Apple SSD cause I didn't want to take my machine apart or take it to repair shop (fear of them scratching, dust, possible fan issues etc). Just wanted peace of mind.

But since it's bootable now, would it be better in terms of speed or cheaper? to go external for equivalent 240Gb SSD? Maybe even the option to run RAID 0 SSD? Example of the Little Big Disk seems to be quicker than the Apple SSD?

write at up to 352.5 MB/s, while read speeds reached 827.2 MB/s.

Another question I suppose is that no external enclosure is available yet. Except for the Promise one's.

(side tracking - since the Promise has swappable drives, in theory could you load that beast up with SSD drives instead of Hard drives? Wouldn't the 6 x SSD drive blow the Thunderbolt cable and port in to the stratosphere? )

 

[philipma1957]

Now that its' finally confirmed.

Thunderbolt Supports Booting From External Disk

Originally waited and waited for Apple SSD as the waiting time was 4-6 weeks, now it's gone down to 2-4 days. Almost went and ordered but now that I read it's bootable. Not sure which one to get.

I was gonna put up with the slow Apple SSD cause I didn't want to take my machine apart or take it to repair shop (fear of them scratching, dust, possible fan issues etc). Just wanted peace of mind.

But since it's bootable now, would it be better in terms of speed or cheaper? to go external for equivalent 240Gb SSD? Maybe even the option to run RAID 0 SSD?

Another question I suppose is that no external enclosure is available yet. Except for the Promise one's.

(side tracking - since the Promise has swappable drives, in theory could you load that beast up with SSD drives instead of Hard drives? Wouldn't the 6 x SSD drive blow the Thunderbolt cable and port in to the stratosphere? )

yes but the 4 drive one would be good enough! t-bolt can really smoke.



BTW this is my plan below:


Here is a 1 drive fw800 out of site out of mind. I will get this t-bolt


http://www.lacie.com/dk/products/product.htm?id=10549


or this one


http://www.sonnettech.com/news/nab2011/

mount behind a 2011 imac like the one below

 

[BigBeast]

I'm waiting until 2.5" portable enclosures come with TB. Then I can take my TB SSD anywhere and boot from THAT. Already got the 2011 iMac. All that's needed now is the 2.5" TB enclosure...

Also- I don't want an SSD TB raid like Lacie's Little Big- I want a thin 2.5" portable SSD that is powered by TB. Just sayin...

 

[Icaras]

I'm waiting until 2.5" portable enclosures come with TB. Then I can take my TB SSD anywhere and boot from THAT. Already got the 2011 iMac. All that's needed now is the 2.5" TB enclosure...

Also- I don't want an SSD TB raid like Lacie's Little Big- I want a thin 2.5" portable SSD that is powered by TB. Just sayin...

Agreed, I'm waiting for the exact same thing, however the Sonnet Fusion F2TBR looks very nice and compact, even though it's a RAID drive.

I'm a fan of G-Technology so I'm really eager to hear some announcements on real TB products.

 

[theSeb]

I was gonna put up with the slow Apple SSD cause I didn't want to take my machine apart or take it to repair shop (fear of them scratching, dust, possible fan issues etc). Just wanted peace of mind.

The real world difference between "slow" and "fast" SSD drives is negligible. I've posted graphs from anandtech a couple of days ago to show this. It seems that most of these "fast" drives come with a lot firmware issues and headaches. Consider this in your decision process. I would recommend something like the intel 320 if you're buying aftermarket.

Wouldn't the 6 x SSD drive blow the Thunderbolt cable and port in to the stratosphere? )

Yes, it seems that it is possible to replace the drives in there with SSDs or SAS drives. How would you feed it to "blow the cable and port"? With another similar device? If you were copying to the Pegasus, you would be limited by your source, unless your source is a similar device with a RAID array.

 

[davidcmc]

The main reason that keeps me from buying an iMac is the horrendous and complicated process to simply swap the original HDD with my OCZ SSD.
I don't want to give Apple my money for a bad and overpriced SSD.

I think it's taking too much time for these companies to release a simple 2,5" enclosure with a Sata -> TB port.

 

[theSeb]

The main reason that keeps me from buying an iMac is the horrendous and complicated process to simply swap the original HDD with my OCZ SSD.
I don't want to give Apple my money for a bad and overpriced SSD.

I think it's taking too much time for these companies to release a simple 2,5" enclosure with a Sata -> TB port.

Overpriced? Yes
Bad? In what way?

Edit: I love these graphs

 

[davidcmc]

Bad? In what way?

With "bad" I mean relatively slow.
There are cheaper and better SSDs out there, not to mention the larger space that they have.

Also, lack of TRIM support is not a big deal.
There's an article in some tech site (I think it's AnandTech) showing how SSDs, even without TRIM support, don't suffer performance degradation in OS X.
Their final statement was that OS X file system (HFS+) seems to work differently from NTFS and doesn't cause performance degradation.

 

[Hellhammer]

Considering that TB cable alone is 50 bucks, going with the Apple SSD doesn't sound that expensive anymore. Also, the 4TB Pegasus box is 999$, whereas a similar box from OWC with FW800 and eSATA is 549$, so again, TB seems to cost a hefty premium.

I wouldn't be surprised to see a simple 2.5" TB enclosure going for around 200$. If you want the SSD now and don't want to wait forever, just go with the Apple SSD. It's fast enough and reliable.

 

[theSeb]

With "bad" I mean relatively slow.
There are cheaper and better SSDs out there, not to mention the larger space that they have.

Did you see the graphs above? What do they suggest to you about synthetic benchmarks?

 

[88 King]

I found there are couple of problems using 2011 imac with OSX on external Firewire SSD after a week.

Firstly, Bootcamp will not install on external drive, so you have to run it virtually. Secondly, I haven't found a way to use Time machine to backup OS in my external drive.

So booting from an external Thunderbolt drive could have the same problems.

It will be cheaper to swap out the internal hard drive with a SSD, and use the internal drive via Firewire.

The only reason I can see using Thunderbolt external enclosure is if you using multiple SSD drives for storage.

 

[Icaras]

I wouldn't be surprised to see a simple 2.5" TB enclosure going for around 200$. If you want the SSD now and don't want to wait forever, just go with the Apple SSD. It's fast enough and reliable.

I have a feeling the $200 price tag would be for 128GB enclosures...

 

[davidcmc]

Did you see the graphs above? What do they suggest to you about synthetic benchmarks?

Now please post random 4k access.
System startup, game loading and similiar things are insignificant and usually perform the same way in most SSDs, since they don't demand exclusively read/write power, but also some other factors.

I want to see file operations, not CPU/RAM demanding applications/situations.

 

[Badger^2]

I found there are couple of problems using 2011 imac with OSX on external Firewire SSD after a week.

Firstly, Bootcamp will not install on external drive, so you have to run it virtually. Secondly, I haven't found a way to use Time machine to backup OS in my external drive.

Those problems are specific to you.

Booting a SSD through firewire is a huge limit on the speed of the SSD. Like driving a Ferrari around in first gear.

And who needs bootcamp? Thats what parallels is for.

Unless you need the hourly backup and file retrieval that TM offers -- there is other backup software you could use.

 

[theSeb]

Now please post random 4k access.
System startup, game loading and similiar things are insignificant and usually perform the same way in most SSDs, since they don't demand exclusively read/write power, but also some other factors.

I want to see file operations, not CPU/RAM demanding applications/situations.

*face palm*

 

[88 King]

Those problems are specific to you.

Booting a SSD through firewire is a huge limit on the speed of the SSD. Like driving a Ferrari around in first gear.

And who needs bootcamp? Thats what parallels is for.

Unless you need the hourly backup and file retrieval that TM offers -- there is other backup software you could use.

Those are my problems, but its does not mean I'm alone with them.

Day to day use of the ssd with Firewire is very quick, been using Adobe Master Collection last few days mainly playing around with After Effects; Photoshop and Dreamweaver, its much faster than stock hard drive.

Beside a Ferrari in 1st gear is still faster than most cars out there, and no body drive around their Ferrari at top speed all the time.

Bootcamp is useful for those want to play latest game I suppose. Parallel or VMware will not offer the same performance of running current games as in Bootcamp.

 

[Icaras]

And who needs bootcamp? Thats what parallels is for.

It really amazes me how you constantly come up with the most incredibly ridiculous generalizations.

That's right. NO ONE needs bootcamp. Scratch every windows gamer out of the entire Mac community now!

Bootcamp is useful for those want to play latest game I suppose. Parallel or VMware will not offer the same performance of running current games as in Bootcamp.

That, and being able to run those games in decent enough fps. I would never use virtual machines on a Mac to run modern Windows games. EVER.

 

[Sylonien]

Thanks for the info and thoughts.

Well I was gonna put up with the Apple "slower" SSD cause like said I'm not confident in opening it up myself (for various reasons) and don't want to take it to a shop which would cost more I believe.

Agility 3 240Gb costs £370, then probably cost like a £100 to get it fitted? Totaly £470~.

Apple 1TB Serial ATA Drive + 256GB Solid State Drive = +£480.

Can't be bothered to faff about if real world difference isn't huge? And messing around installing it on the external drive.

Mind you I will probably still get a 60Gb SSD RAID (or Lacie) external TB for using it as a scratch disk.

 

[davidcmc]

*face palm*

Look once again at the tests you have posted and do that face palm in front of a mirror.
Chrome build, WoW startup, Firefox startup. That must be the best way to compare hard drives... Oh yea.

LOL.
I don't know why I still post in this forum.

 

[clyde2801]

Give our larcenous friends in the People's Republic some time to get their hands on thunderbolt parts. You'll start seeing them pop up on eBay from anywhere from $5-$30. Just as long as I'm not the first person to buy one.

 

[theSeb]

Look once again at the tests you have posted and do that face palm in front of a mirror.
Chrome build, WoW startup, Firefox startup. That must be the best way to compare hard drives... Oh yea.

LOL.
I don't know why I still post in this forum.

The whole point is that you have missed the point. You've been checking out the reviews, without understanding what you're looking at, and you have seen that drive A can do 3 gajillion 4K random reads a second but drive B can only do 1 gajillion reads a second. The problem though is that back in the real world in our current paradigm 1 gajillion is enough and you won't see a real world difference between drives A and B in typical tasks. Anandtech make this quite clear in their reviews.*

I would rather have the 1 gajillion drive that is more reliable instead of one of the 3 gajillion drives with more issues than I have time to list. Sandforce, I am looking at you.*

 

[Hellhammer]

Lamers? Is that how kids talk these days? I hope you realise that it isn't a real word.*

The whole point is that you have missed the point. You've been checking out the reviews, without understanding what you're looking at, and you have seen that drive A can do 3 gajillion 4K random reads a second but drive B can only do 1 gajillion reads a second. The problem though is that back in the real world in our current paradigm 1 gajillion is enough and you won't see a real world difference between drives A and B in typical tasks. Anandtech make this quite clear in their reviews.*

I would rather have the 1 gajillion drive that is more reliable instead of one of the 3 gajillion drives with more issues than I have time to list. Sandforce, I am looking at you.*

There is the AnandTech Storage bench that shows you otherwise:

http://www.anandtech.com/show/4341/ocz-vertex-3-max-iops-patriot-wildfire-ssds-reviewed/6

You can read the specs of the test in the article but even the heavy test includes basic operations like photo-editing, game loading, downloading etc. The difference isn't noticeable when you have just one test because the difference is so small but when you add hundreds of those tasks after each other, the difference starts to add up like the above graphs show.

 

[theSeb]

Hellhammer, I am not the first to question SSD benchmarks; anandtech forum is full of similar posts. So the difference between between the vertex 3 and the intel 320 is 120 seconds of disk busy time. Considering what actually happens in that benchmark I still have to question exactly how many times a user would do all that to notice a significant difference in their workflow between the two drives. In addition, I reckon those 120 seconds would be wiped away quickly considering all of the issues users are having with sandforce-based drives.

 

[davidcmc]

There is the AnandTech Storage bench that shows you otherwise:

Image

Image

http://www.anandtech.com/show/4341/ocz-vertex-3-max-iops-patriot-wildfire-ssds-reviewed/6

You can read the specs of the test in the article but even the heavy test includes basic operations like photo-editing, game loading, downloading etc. The difference isn't noticeable when you have just one test because the difference is so small but when you add hundreds of those tasks after each other, the difference starts to add up like the above graphs show.

He bases everyone's necessity by his own necessity.
His workload is light and he thinks everyone's workload is all about starting up Firefox, WoW and booting Windows.
He may not have heard the word "virtualization", or even tried to copy large 20Gb files every 30 minutes.

 

[theSeb]

He bases everyone's necessity by his own necessity.
His workload is light and he thinks everyone's workload is all about starting up Firefox, WoW and booting Windows.
He may not have heard the word "virtualization", or even tried to copy large 20Gb files every 30 minutes.

The light benchmark is this:

"There's also a new light workload for 2011. This is a far more reasonable, typical every day use case benchmark. Lots of web browsing, photo editing (but with a greater focus on photo consumption), video playback as well as some application installs and gaming. This test isn't nearly as write intensive as the MOASB but it's still multiple times more write intensive than what we were running last year."

This is the heavy workload:

"2) I tried to cover as many bases as possible with the software I incorporated into this test. There's a lot of photo editing in Photoshop, HTML editing in Dreamweaver, web browsing, game playing/level loading (Starcraft II & WoW are both a part of the test) as well as general use stuff (application installing, virus scanning). I included a large amount of email downloading, document creation and editing as well. To top it all off I even use Visual Studio 2008 to build Chromium during the test.
"

How exactly does a faster ssd speed up your movie watching or browsing the web? How much faster is editing a document on one ssd versus another or saving it? Why do these benchmarks not show just how much time you can save in a real workflow? Who would do all that in a real workflow? If I am working in Xcode and compiling my code at various intervals then that's all I am doing.

This is why so many have asked for real world benchmarks and that's why anand included the graphs that I attached earlier in this thread. It shows that building chromium on a ssd vs a mechanical disk is a huge difference. But what's the difference between ssd 1 and ssd 2? It's minimal. So in my real workflow how much time am I going to really save by choosing one SSD over another? Never mind the reliability issues and that whole can of worms.