Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Does anyone know how good the standard 128 SSD in the Unibody 17" MBP is? and if I was got get the 320GB 7200rpm HDD could I change it to the SSD at some point in the future?
 
The article went on to say that unless you erase the whole disk and rewrite it, the fact that nothing gets actually rewritten to 'defragment' the info on a ssd, makes the driver perform slower the more data you write to it.

I'm not trying to be an authority on the drives, but you gotta read the article as it was an eye opener for me and made me decide to go with a 7200rpm drive for now until the ssd prices are justified in my eyes.

I agreed with you, and also said already that the exact same thing happens to HDD (just for a different reason, and its less apparent since they are much slower than sdd)

I got mine (which is like the crappiest SSD out there, with the infamous JMicron controller, and is the cheapest of its size {at the moment I bought it.}) and I filled it up until 1GB was free (~117GB used, 119GB after formated) and it really does not feel slugish, now it does not feel like when I first put it inside my laptop but I assumed it was due to "system wear" and that formatting it would go away, I do not know if that will happen anyway it feels at least 99% of the initial speed (almost non-noticeable)


Cheap SSDs with jmicron controllers are a NONO. Do not get them. These are typically the SSDs with the lowest prices that make you think, hey, these are affordable. They have good read performance (average for SSDs) but they are horrendous for random writes (which reflects real world usage). Things gets so bad that the your computer may 'freeze' until the writes are completed. Goole 'SSD' and 'stuttering' to get a feel for the problem. The problem is not if you'll experience the stuttering, but when. Give them to people you hate. As they are, they are in no way acceptable for desktop usage.

Thats half the thruth. While JMicron sdd are eclipsed by the intels and samsungs they still offer performance "WAY" better than [LAPTOP HDDS] (again do not compare to 1TB 7200 rpm 32mb cache desktop HDD or velociraptor because there is no way you can fit that inside a MBP, thats like comparing a honda with a ferrari, keep it in contest)
In the other side the intel drives cost half of what your computer worth (usually, unless you got a fully loaded 17") and its hard to justify paying all that money for ~70GB after formating (or ~150 if you go for the $800+ one) Also most of the problems with those drives are being reported by windows users I have yet not seen somone using OS X complaining about stuttering yet (I also own one and really I have not experienced none) and on benchmarks at least ramdom writes (which you said its what kills these drives are pretty low, still are slightly [3-5%] faster than my old 5400rpm (I guess it would be lower than 7200rpm or desktop HDD, but again is not fair to compare to desktop HDD, and while you can indeed get 7200rpm for laptops some of them are just as bad and will make tons of noise, vibrations, etc etc etc


Anyway if you want SSD now I say get the crappy one and the smallest you can live with, because Intel's and Samsung's and toshiba's are likely to come down (at least to half its current price, if not more) within 6 (12?) months or so.
 
Thats lotsa of help xoggyux, thank you. Did you get the standard SSD thats BTO with the MBP?

Also will it pretty much be a straight swap out once the better SSD's come down in price.
 
Thats lotsa of help xoggyux, thank you. Did you get the standard SSD thats BTO with the MBP?

Also will it pretty much be a straight swap out once the better SSD's come down in price.

I do not know what you mean with BTO (you mean standard?, if so no. I bought it off newegg @ 220 shipped, the swap is very easy in mid 2008, and for that matter all C2D if I am not mistaken, and even easier in new unibodies)

And yes, thats my whole point, I know the $800 are better, just not 4 times better :D (they cost at least 4 times more)
 
Sorry mate, BTO = Built To Order, you can have the SSD as an optional HD upgrade for 12GB @ £240.

Just wondering what type/brand of SSD it is and if it's worth buying. At moment thinking that I should just get the normal 7200rpm 320GB and wait for better/cheaper SSD to come out and then swap it over.
 
Sorry mate, BTO = Built To Order, you can have the SSD as an optional HD upgrade for 12GB @ £240.

Just wondering what type/brand of SSD it is and if it's worth buying. At moment thinking that I should just get the normal 7200rpm 320GB and wait for better/cheaper SSD to come out and then swap it over.

whatever you do, do not buy off apple
its true that you might be tempted to avoid opening the case and after all apple's are ~$100 more only, however when you upgrade yourself not only you save a bit of money (not much anymore since apple dropped their prices recently, but still its something) but you also get to keep the internal HD that came with your laptop that with ~$15 or less you can convert into an external HDD (if you check my thread you can read my review of rocketfish enclosure that even comes with eSATA which its even bootable, I booted OS X off it, I will probably try boot camp off it as well though I have no idea if it will work) also new unibodies HDD upgrade (as read from ifixit) are like changing the batteries of a remote control, if chicken had opposable thumbs they could do it.

And if you should or not wait, thats really up to you, you will get tons of replies saying "DO IT, IT WORTH IT" and some many other saying "TOO EXPENSIVE WAIT A BIT LONGER" anyway my opinion that $200 is not too much for an experiment that would likely boost your performance and give you extra storage (as you can keep internal drive as external) even if you dont like it you could use it as flash memory in the worse case scenario. Anyway I think in after spending $2000+ in your laptop $200 makes little difference in your pocket. And waiting yes.. BUT its always an issue to "upgrade" 1year+ old computer (at that time you will have nightmares asking yourself if it really worth upgrading it or just wait a bit longer until its time to buy the next, etc etc etc)
 
Just got my Intel SSD and installed it today. Its soooo fast WOW! Very happy with the speed!:D:D:D The only con is that there very expensive.
 
Are there any statistics about SSD reliability/longevity? I'd hesitate to invest a great deal of important data in a new, unproven technology. At least with platter hard drives you know you have to have 2+ backups...

There is no such statistics. Storage manufacturers give some nice but largely irrelevant to real life figures.

Everybody (or was it just me?) thought that there was such statistics for HDD, but no, it is not available for them too. It's just every body has lots of experience with them. (Under non-stop load, 2.5" HDD would die within 1.5 years.)

I have 256GB SSD and it is very nice performance boost. Also my lappy most of the time remains very very cool. Never bothered to check battery life, but I think for most laptop workloads SSD should be also beneficial: main power drain occurs only during write operations.

Now to the bad news. There is nobody who can tell any tiny piece of reliability information for SSDs. Anand wrote a lengthy article on SSD which also (mainly between lines) has reliability information.

Generally on SSD one has to keep lots of free space, otherwise wear leveling wouldn't be any effective. (e.g. if you have only 5% free and do lots of work - the 5% of SSD would be wearing out, while rest of SSD remains untouched.)

Modern flash media is capable of surviving ~1Mln of erase/write cycles. But the problem as some sites have identified, is that quality of flash isn't uniform: some cells would survive 5Mln cycles, some - only 100K. Controllers on SSD try to leverage that. But again, if you have very little of free space on SSD, the wear level algorithms would be very inefficient leading that those bad cells might fail pretty soon.

I would advise to read Anands article in full to understand what you are buying into.

Or. If you lazy and have some money: better save your money now and buy plain HDD.

Or. If you lazy and have lots of money (hehe like me): buy SSD, but better be ready to replace it somewhere within one year. At least that's my personal plan: I hope that in one year prices would come down and quality would go up. And probably manufacturers - directly or indirectly - would publish some reliability information.

P.S. In the second article I linked, there is a reliability figure from Mtron:

In the data sheet for their 32G product - which incidentally has 5 million cycles write endurance - they quote the write endurance for the disk as "greater than 85 years assuming 100G / day erase/write cycles" - which involves overwriting the disk 3 times a day.

The catch here is that they rewrite *whole* disk 3 times a day. On busy SSD of mere mortal, generally 95% of space is written once with some static data and only 5% are constantly (over)written.

Samsung MLC drives have slightly lower figure, because Mtrons I have checked have 1.5Mln erase cycles, while Sammys have 1 Mln cycles. But since the figure is useless anyway, it doesn't matter much. As Anand points out, Samsungs are number one in quality department. Intel is faster, but Sammy's firmware is simply better tested and more robust.

Huuuu.... I'm done.
 
I saw a study (i don't known if it was in cnet) where ssd was fast when new but later and with a lot of use they become pain slow, may be because fragmentation or other cause. And yes, ssd are noticeable affected by fragmentation even when the access time is almost zero.
 
I saw a study (i don't known if it was in cnet) where ssd was fast when new but later and with a lot of use they become pain slow, may be because fragmentation or other cause. And yes, ssd are noticeable affected by fragmentation even when the access time is almost zero.

Those were very likely quotes from Anand's article. Also note that most of the tests are performed on Windows and NTFS is highly susceptible to fragmentation. Not so Mac OS X implementation of HFS+.

Fragmentation has the effect that to rewrite even small block of file (fragment) whole SSD block has to be erased. Erase operation is the most expensive operation on SSD. As long as file system doesn't fragment too much (what is case of Mac OS X) this is not a problem.

Also Anand noted that in no event one should try to defragment SSD - that is quite expensive operation for SSD as all used blocks would be erased at least once - or more. The only solution to bring old performance back is to back up drive, reset it's state (using special manufacturer supplied tools) and restore content from back up.

Thinking about it now... If one plans to use BootCamp, probably it is better to avoid SSD for now. Windows 7 is expected to come with number of SSD improvements, but WinXP and Vista might seriously decrease life span of SSD.
 
I saw a study (i don't known if it was in cnet) where ssd was fast when new but later and with a lot of use they become pain slow, may be because fragmentation or other cause. And yes, ssd are noticeable affected by fragmentation even when the access time is almost zero.

SDD should not be susceptible to fragmentation (thats HDD's problem due to its nature of a reading head moving along a moving disk) does its true SSD become slower overtime is due to something else (I read the explanation but the author took 3 pages and lots of pictures to explain, I am not going to say it here :D)
 
SDD should not be susceptible to fragmentation (thats HDD's problem due to its nature of a reading head moving along a moving disk) does its true SSD become slower overtime is due to something else (I read the explanation but the author took 3 pages and lots of pictures to explain, I am not going to say it here :D)

If ssd was unaffected by fragmentation then sequential reading time will be near, if not the same, as random reading time, but it is not true.

And googling i can read that hfs+ can (and is) affected by fragmentation, in fact i don't think that exist a file system that can't be fragmented.

http://www.pcper.com/article.php?aid=669&type=expert&pid=7

SSD is quicker but not for every single case and depend much in the kind of usage, or you could say, a mixed bag.
 
If ssd was unaffected by fragmentation then sequential reading time will be near, if not the same, as random reading time, but it is not true.

As I was highlighting above, fragmentation on SSD is bad for write performance, not for read performance.

In HDD, to read fragmented file, you need to perform lots of seeks (reposition head of HDD to another cylinder) to access its every fragment. And seeks are expensive (milliseconds). More fragments -> more affected blocks -> more seeks.

Is SSD, to rewrite fragmented file, every block where fragments of file reside might had to be erased/written anew. Normal writes are fast on SSD. But erase is very expensive (dozens/hundreds microseconds; The Operation which actually decreases the longevity of SSD). More fragments -> more affected blocks -> more erase operations.

And googling i can read that hfs+ can (and is) affected by fragmentation, in fact i don't think that exist a file system that can't be fragmented.

http://www.pcper.com/article.php?aid=669&type=expert&pid=7

It's all relative. On NTFS to see file ~1-10MB with *hundreds* or even *thousands* of fragments is not that unusual.

You can use Windows standard defragmenter to check that: files with most fragments are listed in report given after analyze operation. (It's pretty much only what it's good for anyway.)

P.S. Microsoft broke built-in defragmenter of OS/2's HPFS (ancestor of NTFS) when IIRC they introduced extended ACLs in around NT 3.5 times. HPFS had a fragmentation resisting logic (and I believe that IBM was first who actually implemented that in file system). And M$ still can't fix it... No comments.
 
I hear the soilid state drives are rugged, somewhat lighter and have no moving parts. Also super fast. But I hear they can wear out easily. Take alot of power. And when they fail all is lost:eek: So what should I do?

I read somewhere here that another advantage of SSD is the fact NOT all is lost. the poster stated that you wouldn't be able to write anymore, but reading should be possible.

only read this once so i'm not sure if this is correct.
 
I read somewhere here that another advantage of SSD is the fact NOT all is lost. the poster stated that you wouldn't be able to write anymore, but reading should be possible.

only read this once so i'm not sure if this is correct.

this is correct. the SSD figures out that a block is bad when it tries to write to it. if it's bad, it will not write anything to that block and the data preserved. with a hard drive, you figure out the data went bad during reading... and obviously at that point its too late.
 
The 128GB SSD from apple is the Samsung 128GB. The Corsair is a re- badged Samsung. There is a thread here that shows all the stats here somewhere. FOo me... @ that price - its fast enough and it's SUPER low power to boot. Ill upgrade to a better product in 6mo. I got mine on a BTO MBP 15" thats on its way from China.


Does anyone know how good the standard 128 SSD in the Unibody 17" MBP is? and if I was got get the 320GB 7200rpm HDD could I change it to the SSD at some point in the future?
 
I've been using G.Skill Titan 128GB SSD in my MBP and have been extremely happy. It costs around $300 on sites like newegg.com, has really good read/write speeds and stays cool. I don't think I'll ever buy non-SSD laptop drive again.
 
My OCZ 120GB Vertex is pretty fast. I just instlaled my os and restore my time machine data. Apps open instantly. I love it.
 
I'm looking forward to getting a Sandisk G3 SSD, which should be out sometime Q3/Q4 this year. This says it all:

http://gizmodo.com/5126848/sandisks-g3-ssds-deliver-40000-rpm-speeds-without-breaking-the-bank

60GB $149
120GB $249
240GB $499

Their performance has a virtual RPM equivalent of a 40,000RPM disk (doesn't exist) and they made their own controller that uses an in-house optimized algorithm called ExtremeFFS to greatly enhance read/write longevity.

This will be a good day, because when it launches, it should push SSD prices down.

2009 is definitely going to be the year of SSD.

And no, I don't work for Sandisk. Just excited about what's coming.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.