I've read about some Macbook hacks for the SSD user, but still wonder how I should make my SSD's "new" state last longer.
I do still wonder if I should use the "Secure Empty Trash" in Snow Leopard? At first, I've partitioned my drive for all available space, but then reduced the HFS partition to 68 Gb or so - will this prolong my drive's "new" state because of Intel's algorithms?
Should I use "Erase Free Space" in Disk Utility?
Thanks!
I would say NO. Secure Erase (of used or free space) will generate (many) more writes which can only make the problem worse, "using up" your reserve space much more quickly than before. At least that's what my intuition says. Maybe somenoe who knows for sure will chime in.
. . . yet that article was never bumped to the top. Even the G1s were priced higher! Props to Anand for keeping the SSD companies honest on performance, but why can't the same be done for pricing?
If I were going to release/retail a new SSD, I would definitely call Anand ahead of time and tell him an expected price about 50-75% of the real one so he would blindly print it and create demand.
Another great article, thank you for all the hard work and insightful analysis. Articles like yours help keep the industry honest and responsive.
I can't wait until they become cheap enough for me to buy two for a RAID 0, and of course for Intel to release new drivers to support the TRIM commands.
Ok so with all this new data in mind, what's currently the fastest, least performance degrading 256GB SSD you can get for under $1000? I'm currently running Snow Leopard on a dual quad core Mac Pro... I'm not sure Snow Leopard supports any of the new technologies like TRIM or not... but whatever the best 256GB SSD I can get, that's what I'd like....
I just wanted to add a word of warning over drivers and SSD. I just built a new i5 750/P55 system with a Corsair X64 drive (Indilinx). While the system worked fine in bits it frequently froze. I tried lots of things to cure it, but in the end the fix was simple. I HAD to use the Intel drivers to avoid intermittent freezing.
Reading around some of the othere references Its inconsistent and depends on how the drives implements the command set so your mileage may vary but I think its a big issue.
I think we could all do with an article on how best to configure an SSD based system. I decided that I wanted to move "Users" and all associated hidden directories to the D: drive (a 1TB HD) leaving the 64GB for Windows and Apps. This would mean it would never get too full.
I'm amazed at how difficult Microsoft make it to shift the "Users" directory to another drive. In fact they have a Knowledge base article saying not to - Linux easy, just map /home elsewhere.
In the end I followed some instructions to create a junction from C:\Users to D:\Users, but its non-trivial and involves opening a command window in the middle of the installation process.
Yes I'd like to have done it with an autounattend.txt but the MS KB articles says it won't really work.
> In the end I followed some instructions to create a junction from
> C:\Users to D:\Users, but its non-trivial and involves opening a
> command window in the middle of the installation process.
I can't see this sort of thing becoming any easier until MS ditches
the archaic idea of drive letters and switches to a more natural
unified file system such as is used with all UNIX OS variants.
(oh look, the pits of hell seem to be getting a tad icey...)
" "I think we could all do with an article on how best to configure an SSD based system"
"That would be great " "
Thirded.
The variety of do's and don'ts (and conflictions) are numerous:
-do disable pagefile. don't. oh, which OS?
-don't defrag disk. do. oh, which OS?
-use ram disk for temp files. don't. no, wait, do.
If the only update was the 2 short blurbs at the top of the first page I don't think this entire article should be reposted. A simple statement with a link to the previous article would be fine. I just wasted a couple minutes going through each page realizing nothing was new....not cool. I was psyched for another SSD article Anand!
On that intel firmware that provide trim support but bricked people's drives, now intel said it is a problem specific to Win7 64bit. Does this mean that if you update the firmware from within Win7 64bit there could be a problem, or does it mean even if you updated the firmware in Vista or Win7 32, and then install Win7 64 on it, it could still cause a problem? Does anyone know the answer?
I just bought a X25-M G2, I plan to use it in a new computer. Can you just update the firmware on an older computer first, and then install WIn7 64bit on it?
I was curious about the possible setup of these Kingston drives in a Raid 0 format. If they are roughly half the performance of an intel 80GB drive, it would be reasonable to think 2x of these Kingston drives would be comparable to a single Intel 80GB drive. However, I am not familiar with TRIM support on a Raid setup. Is there any info on this?
I have a x200 Thinkpad with a SSD inside. The SSD drive is a 128GB Samsung (OEM MMCQE28G8MUP-0VAL1). I noticed that the drive becomes slower after three month of use. So I use WIN XP SP3. Is it possible to use Trim or Garbage Collection in any kind of way on this older Samsung SSD drive? The Samsung support homepage and google couldn’t answer my question.
I wish Anand would acknowledge these extremely high prices and how wrong the prices in that article were (preferably in a follow-up). Especially since he is again making statements about Newegg's possible future pricing (of the Kingston).
Now, I've been trying to find these for that price at Newegg, and can't. The drive now shows up (just the 2.5" one, not the one including the bracket) but for $124.99 before rebate $104.99 AR.
I bet this is an error they will fix, as $104.99 Before Rebate would make $84.99 AR, which is what Kingston mentions in their press release.
Rebate looks good for 2 per person, so I plan on buying two for a new Windows 7 box.
Disclosure, I work for EMC (an enterprise information infrastructure - which includes all sorts of storage arrays).
Looking at some of the comments, I'm not sure if people understand the impact of the x-25 getting down to the prices they are (both via kingston OEM and the 80/160GB drives).
Flash will rapidly replace all high-performance disk use cases. There will only be room for very large SATA and SAS disks, and all high performance use cases will be dominated by Flash.
Some people don't understand that even at TODAY's prices, for some (many) use cases (just not consumer focused ones), they are more economical. For example, the kingston drive, measured in IOPs/dollar is 37x **cheaper** than a 15K SAS drive. That's the acquisition cost. When you think that it would take 37 15K SAS disks, consuming more power, space and cooling - you can immediately imagine the impact this has on the enterprise storage market.
The rapid price decline we've all seen over the last year (a 40GB MLC drive cost ~$900 at the beginning of the year) means that in the next year or so, we'll have SSD with the $/GB of a large, fast rotating magnetic media disk, but still 100x better random IOps/$
While perhaps PCM non-volatile storage will eventually replace flash, that will happen well after flash replaces spinning media.
Thanks for the great article! I am just curious if the Kingston UltraDrives can use the same firmware update as intel and can be considered the same drive or if it is something like with OCZ and Super Talent in that they have separate firmware updates. Would the Kinston be a cheaper alternative to the Intel?
Apparently Intel pulled the iso. Here's what you get after you click on the link,
"02HA Firmware Upgrade for Windows 7* Systems - Unavailable
Intel has been contacted by users with issues with the 02HA firmware upgrade on Windows 7* systems and are investigating. We take all sightings and issues seriously and are working toward resolution. We have temporarily taken down the firmware update while we investigate.
Good article Anand, and I like the new tests you guys came up with. I have a request for a new type of test and I know others in the various hardware forums have these same questions.
SSDs are too small for many of us to use as the lone drive in our system, and we thus have to combine an SSD as an OS/app drive w/ a traditional hdd for our file storage. Given that this is a reality for many of us eager to jump into SSDs, it'd be great if we could see a test that demonstrates a real-world scenario of users loading DBs/pictures/videos etc from our storage drive while still running our apps/OS from the ssd.
I know it's probably not something you guys could do regularly because the amount of testing would be a huge burden, but perhaps you fellas could do a one-off article just to highlight the differences? Maybe show a couple different usage scenarios such as:
~ a budget/midrange P55 setup w/ OS/apps on a small Indilinx ssd and a single 1TB drive for storage
~ an older Core 2/AMD system with the new Kingston offering used for OS/apps and the storage drives being pair of 640GBs in RAID 1.
I think this kind of article would be very useful to your readers and I know there's lots of us in various forums who are hesitant to jump into a hybrid ssd/hdd setup because we're unsure of exactly how it will affect us.
Still not a lot of info out there showing real-world usage scenarios w/ an ssd as the os/app drive and a regular 7200rpm or two as storage, but perhaps for others interested in this kinda testing something like this will suffice. From the MSI p55-gd65 review-
I do get a feeling that the 2009 test suite is heavely favoring the g2 drives
It is plain and simple what takes time. Not iops, we need seconds. And seconds where you can feel it; fx. seconds for diferent types of filecopying, handling large files, working while making backup, working while having virus scan.
The g2 is way overhyped, the samsung way underrated
Reading the results you can have the impression the velociraptor doesnt work for desktop use
The need for trim is overhyped and having severe consequenses
And the results from defective and bad bios updates from indilinx and espec. Intel, was to be seen in advance.
Are we going to se some new benchmark suite when gfx lrb arives??
The end results. Ordinary desktop people having hdd for servers and gfx for research.
It doesnt make sense.
Some bad thought creap into my mind. Help me. Is the Intel marketing a problem, if you dont behave? :) - or did they just give you the random 4k engineering hammer?
The 4k thing was a way to quantify the stuttering problem on some drives, as if he had just said "the drives stutter" but only shown objective results that look good (due to sequential write speeds), then people would have claimed he was making stuff up to be an intel fanboi as well.
Human-observed delay time is really all everything comes down to. 4kb performance and IOPS are interesting to look at, but they're severly inadequate. If I remember right, RAID0 tends to look awesome in benchmarks, but offers virtually zero real-world benefit.
So screw the benchmarks get and out a stopwatch. Measure what matters: boot time, time to launch various applications, media encoding time, and the time to copy files. Quantified seconds are a helluva lot more meaningful than IOPS.
Yep, something like that, but with more categories. I think this is a good set:
1. Boot time
2. Time to launch applications
a) Firefox
b) Google Earth
c) Photoshop
3. Time to open huge files
a) .doc
b) .xls
c) .pdf
d) .psd
4. Game framerates
a) minimum
b) average
5. Time to copy files to & from the drive
a) 3000 200kB files
b) 200 3.5MB files
c) 1 2GB file
6. Other application-specific tasks
Imagine if we had a table of this data and all of the SSDs and a conventional HDD. That would be so much more useful than all the charts in the existing article.
Finally, an SSD that I can afford and makes sense to me. Anyone can afford a POS SSD drive, but this one doesn't look crippled too much. The writes I'm not too worried about, but the stuttering and performance degrading is not what I want to pay a premium for.
I'm hoping by Jan-Feb (when I could afford one) that Kingston will have the trim support and upgraded firmware. Then I'll jump on it!
Just picked mine up yesterday.. Kingston 40gb bootdrive desktop kit.. Has 2CV102HA firmware .. i.e. TRim support already.. be on the lookout for it. .dont know if its all drives now or just some.
What I read from Intel about the Optimizer suggests it is not needed in Windows 7 or not needed to actually run it. Anyone understand what Intel is saying here:
"When using the latest Microsoft Windows* 7 operating system with Microsoft AHCI storage drivers the OS will contain native support to execute the Intel® SSD Optimizer on an Intel SSD without requiring any user interaction."
"Microsoft Windows* 7
Microsoft* AHCI
Native OS support (Intel® SSD Toolbox not required)"
I don't like how Kingston labelled this drive as v-series, this may lead people to think that v-series 64gb and 128gb are good as well, while they end up buying a JMicron crap...I know that the v-series stands for value but I'd expect a kind of consistancy in series parts.
I'd expect 64 and 128 drives to perform equally or better than the 40gb part, but it wouldn't...perhaps kingston had to find a sloppy way to get rid of those unsold "value" drives...
Regarding ssd reliabilty brought up in the recent posts, I'd be pretty confident to put an SSD in a home pc anytime.
Man, they are used for SERVERS (well, SLC drives actually, but the story is the same), one of the most mission critical environment out there.
Dangers are ahead only if you want to mess trying TRIM, fw updates or any other topic brought up every now and then, but if you wait a while and resist early-adopting new features for a couple days, problems are issued quickly. Anyway backup is just a click away.
I would like to see tests on the JMicron-based Kinston V Series. Supposedly their newer controller resolves the stuttering and random access bottleneck of the gen-1 SSDs.
I installed one in a budget build for a customer and was very impressed with performance in a modest use case scenario (with Windows xp and an Atom 330). I would have no reservations using these drives in future builds, although the new 40GB model is an interesting proposition.
Seems not enough testing was done for them to have data corruption, again. These kind of issues shouldn't really surface if proper testing were carried out :/
"Would I recommend waiting until next year to buy? This is one of the rare cases where I'd have to answer no."
Given all of the serious reliability issues -- including today's latest Intel firmware debacle -- it's quite clear that SSD are not ready for prime time. Over and over again, we hear stories about disks not living up to their potential due to bad drivers or firmware, and there have been far too many cases of total failure with attendant catastrophic data loss.
How can Anand recommend using a SSD on a primary machine?
Wait at least a year for the bugs to be worked out if you value your data.
Bullshit. What reliability issues? What firmware debacle? The handful people with potentially bricked devices after the flash hardly qualify as a debacle. I would guess the success rate is over 90%. Flashing is always a risk, you know...
Intel did the right thing by pulling the firmware, though. They'll look into it. But i wouldnt be surprised if it wasnt an error on their end - they spent lots of time making this firmware. Much, much more than the garbage the competition throws and its customers every week.
I've heard this argument for years now, along with quite a few others that have died away. "SSDs are not ready for prime time". Wich SSDs, and for what usage are they not ready compared to HDDs?
Unless you yourself upgrade the firmware of an SSD whitout first waiting and making sure the upgrade is safe, there is no risk by using SSDs that is not greater with HDDs. SSDs generaly are much more reliable and rugged, and when they near the end of their natural life, you will see the raw data failrate predictably increase until it hits the point where the ECC can't do the job anymore, before wich point you take backup and buy a new one.
Anyone who keeps invaluable data on just one physical medium whitout backup would be a fool to think it is safe. Use the SSDs for OS + programs and have a RAID in a redundant mode with offsite backup for your valuable data.
Anand is IMHO right to recommend buying an SSD now and not waiting because we have passed the point where the ratios between price, capacity, performance, and reliability make them far superior to harddisk for boot drives. I've had SSDs for over a year in my computer, and payed a hefty price for being an early adopter, but it was well worth it. The prices, specs, and reliable market today almost make me laugh out loud when people say "SSDs are not ready yet". Claiming "(many) people are not ready for SSDs yet" would be far more accurate.
If you actually look at the number of people with the issue and filter out the chit-chat, its not "alot" actually. However, my flash went just fine.
Thats the risk with flashing firmware regardless of what device it is. Theres always the chance to brick it. Thats also why I dont understand why some people flash every fucking piece of hardware whenever theres a new firmware available - ok, this is a different case, here it makes perfectly sense.
Has anyone had problems flashing a drive before they put any data on it? I have a new G2 still in the box that I just haven't yet had time to do anything with, was planning on flashing it then loading Win7 and Ubuntu 9.10 this weekend, have there been any reports of issues when loading an OS after flashing?
In 'The SSD Relapse' you state that the G2 "doesn't drop in performance when used...at all." Yet in 'The SSD Improv' the 80GB G2 with TRIM firmware drops more than 60% in the 4KB random write test.
Granted, the charts and figures used in 'Relapse' to back its claim are for the 160GB G2 and these results are for the 80GB G2. What do you attribute this difference in performance to, the new TRIM firmware?
As a Mac OS X user, at this point, I feel like I can only really consider drives that perform well without TRIM, especially in the <= 80GB range, so would appreciate your feedback.
I believe it is because in that particular test he is writing to the ENTIRE 80gig drive, so it's not that TRIM isn't working per se, rather that there is no free space to allow TRIM to do anything at all.
As an addendum to my previous post, you can see that after file deletion the performance goes back to virtually new. So in a sense the performance of the drive never goes down, UNLESS you simply delete a partition rather than erasing the data on that partition FIRST and then deleting it.
I think that's probably something that needs to be implemented with an updated driver, or at least a warning box that comes up saying "deleting this partition without formatting will hinder performance".
I don't know if you have done extensive benchmarking with SSDs, but using IO QD = 3 for 4KB random in IOmeter don't yield representative results for intels SSDs. I have been benchmarking SSDs with some other guys for over a year now, and we have found that while Samsung and Indilinx scale from QD 1-4 and then flat out, Intels x25-M scales well all the way to around QD 10-16, and actually is capable of over 140 MB/s at 4KB random read at QD=64 in fresh state, and still over double your messured 60-64 MB/s in used state.
In our benchmark thread, the records with x25-M from ICH10R before TRIM firmware are:
IOmeter
4KB random read QD=64: 40913 IOPS = 163,5 MB/s
4KB random write QD=64: 19360 IOPS = 77 MB/s
PCmark vantage HDD score: 43107.
The same guy that got the random read and PCmark scores above also got PCmark vantage HDD score 120374 with 3 x25-M gen1 from ICH10R.
In other words, the results posted for x25-M in this review are either in "used state" or below par.
You know Anand trys to simulate REAL situations and a QD of 10-16 is absolutly unrealistic for a home user - don't even talk about a QD of 64. Maybe the Intel SSDs shine there, but it's just of no interest.
In his traced "light" benchmark with mostly single-tasking the average queue depth is 6.09. If you take into account that most of the sequential read and write operations (although they are under 30%) don't generate a queue, and that 4-16KB IOs often come in bursts with more queue depth, you can easily get a QD of more than 10 with relatively mild multitasking if it involves disk access. A harddisk's cache is regulary used as a buffer for the write portion of these, but if the cache is full your system almost freezes because of the harddisks low IOPS. An example would be if you try installing windows updates or a program while running a virus scan and listening to music or extracting a compressed archive, you will quickly notice if you try this on a harddisk.
Ok I was a bit surprised and tried it myself with my 160gb Intel G2 (no firmware update yet)
Copy 10gb file onto drive, MSE full scan and extracting a 4gb file with 7z while the usual stuff (FF, PS, eclipse and winamp) running in the backgorund and I got a maximum QD of under 8, which is higher than I expected it to be, but still far away from a QD of 16.
And when doing more multitasking stuff I'd probably tax the rest of the system more than the SSD so at least for my usual behavior a QD >8 is unrealistic, but I see your point.
There are probably some people who could get a QD of 10-12 in a normal workload, so that could be interesting for them :)
There is a big difference between average queue and queue during operations. The IO queues are relatively low for all the background operation the OS does, and sequensial read/write, but when you do things like open administrative tools, open a program, or install something, you will se queue spikes of easily 50-60.
After i commented this article i had a discussion with some other benchmarkers, and we found that normal use generates IO queue spikes almost whenever you interact with the OS. It doesn't take much effort to generate a queue above 100. Running a couple of virus scans on folders of small files while installing a program and opening a folder with lots of pictures will get you a massive queue.
Nice work again Anand - I just don't bother reading SSD reviews/features on other sites any more.
HOWEVER...
Many people reading this will be developers wondering which controller best favours typical development activities, eg. checking out and compiling large projects. It's a hard one to guess - lots of random reading, sequential reading, random writing (temp files) and sequential writing (output files).
So how about a "typical developer" benchmark of, say, compiling the Linux kernel while concurrently encoding some H.264, grepping all the sources for a string, and watching some por..er, I mean, youtube.
Has anyone else evaluated the best SSD type for s/w development?
Not a reply to your question, but I've taken on the habit of mounting my temp files to a ramdisk. Besides the speed improvement, this should help to reduce unnecessary write cycles to the SSD. Similarly, I run without swap (or page file), RAM being as affordable as it is.
So - "Your best bet is to install Windows 7 with your I/O controller in AHCI mode (for Intel chipsets)" this means you set in BIOS sata controller as AHCI not as IDE ?????
Or just leave as IDE and after Win7 install, update intel storage driver as normal...? which is the best solution?
Can I use Intel IMSM RAID on just the data storage drives in my PC and retain TRIM functionality on the W7 Pro OS (X-25M G2) SSD?
My new PC build plan was to use a single 80GB G2 for the OS and programs and two 2TB HDDs in a RAID1 array for my data/photo/video/music storage. Although I read the relevant part of this article 3 times I am still not certain if this is doable. Yes/No?
"Thank you for contacting Intel Technical Support.
I understand your question about the Intel(R) Matrix Storage Manager and the SSD. Since the drives you will be using in RAID are non-solid state disks, and the SSD you will be using will not be part of the RAID, there should not be any problem with TRIM."
I would not expect this to be possible. It sounds like you have 3 drives hanging off of the same SATA controller. You want 2 of them in RAID 1 using the IMSM driver, and one of them using the MS driver.
In my experience it would be impossible to get Windows to load 2 different drivers for different ports on the same controller.
Even if they were physically independent but identical controllers (for example if you had 2 identical SAS cards plugged into the same system), as soon as you update the driver on one controller, a reboot results in all instances of that controller using the updated driver.
"You no longer have to boot to DOS and secure erase your SSD before installing Windows, just quick format the partition before installing Windows 7."
I always thought the "quick format" doesn't wipe the entire drive as clean as a full/long format. Does the above mean we only need to do a quick format to wipe a drive clean to install Windows 7?
Also, I noticed Superfetch, Prefetch, and Ready Boost are enabled in Windows 7 x64 with my X25-M G2 160GB drive installed.
"SSDs are made up of millions of NAND flash cells. They can be written to in groups called pages (generally 4KB in size) but can only be erased in larger groups called blocks (generally 128 pages or 512KB). These stipulations are partially the source of many SSD performance issues."
but isnt it possible to erase data in 4KB, the same way they are written?
Where does speed matter most?
Getting some sense into the g2 hype
Now you can have a job with a Dell and a hd with samsung controller without quitting the job
Thanx Anand for always improving your site and letting us learn
Thats whats move this boys-toy business forward
Thanks. I assume Intel will update their drivers to support RAID TRIM eventually, but if you don't have an 'R' tybe Intel ICH, does Win7 support TRIM if you set up a software RAID array in the OS still using the MS AHCI drivers?
Another great SSD article from Anand (thank you so much)!
But I'm really disappointed that you have to use the Microsoft Win 7 disk driver to get TRIM. It has a huge problem.
Based on my experience, I would normally strongly recommend that the first thing you do when installing Win 7 is replace the Microsoft driver with the latest Intel Matrix Storage Manager.
I had exactly the same crippling problem on a Dell Dimension 9100 which baffled me until I replaced the Microsoft driver with IMSM - problem disappeared instantly.
Okay it was bit of an old PC but of the 3 I have been testing with Win 7 for the last couple of months, it was the only one that used the Microsoft driver instead of IMSM (it was an XP upgrade ie OS reinstall, the others upgrades from Vista). So I don't know how widespread the problem with the Microsoft driver is, it hit my only PC that was a candidate, but I'd be pretty wary.
Otherwise Win 7 has been great.
So it looks like you have a choice: IMSM with no TRIM or a buggy Microsoft driver with TRIM.
Once the Intel firmware is updated to v1.4 with trim support does Windows 7 start using trim? i.e I have already installed Windows 7 before dong the firmware update.
Does ghosting an image back to the SSD have any impact on trim??
That Kingston for $85 is looking tempting for sprucing up a circa-2004 laptop of mine that I use daily. But the laptop is ide. I know of others in the same boat.
Have you heard of anyone producing an inexpensive quality ssd for ide? The Kingston seems like the perfect candidate since it's anyways not as fast as the regular Intels.
Make sure that the drive interfaces match. If it's from 2004, it might still be PATA. I'd imagine this is SATA only.
That said, it'll be interesting to see if soldering more flash chips on = upgrade in space. Shouldn't be too hard to find the flash chips (2 * Kingstons < Intel), but I'll need to find a good soldering jock.
Ummmm... maybe I wasn't making myself clear.
I wasn't referring to upgrading the drive's capacity.
I was referring to Kingston producing a PATA drive that would work with my PATA laptop. I know PATA doesn't work with SATA. (I loosely referred to PATA as IDE, which people often do...)
thanx
Another great SSD article, Anand. Found a problem: Page 3, "Wipe When You Can’t TRIM", when you said "I went into a deep explanation of the relationship between free space and the performance of some SSDs here.", it sounds like there should be a link to a previously written article, but there is no link.
OS X doesn't have TRIM support yet unfortunately. I'm trying to get Apple to at least acknowledge that they will be supporting the feature but haven't had any luck yet. I haven't tried some of the erase options in Disk Utility to see if they do the equivalent of a secure erase. I'll try barking up that tree shortly.
Newegg will just gouge the crap out of prices regardless. $100-130 will be more like $150-180. Remember when you said Intel would be dropping SSD prices with the G2 to $230 and $450 for the 80g and 160g respectively?
I picked up my 80GB G2 a couple weeks ago from Newegg for $239 shipped, for the retail boxed one; the deal also included a Icy Dock 2.5" to 3.5" adapter and a ThermalTake Element S case for $30 with a $30 MIR.
Ya they had them for about 4 hours in an AM Shell Shocker for $240, which is their MSRP. They also had them for that price for about 8 hours at launch and since then its been gouged as high as $400 for a G2. All the other retailers that are selling for around the $240 MSRP are drop-shipping directly from Intel so they typically don't have stock in-hand and are estimating anywhere from 2-8 weeks for delivery....
With your recommendation to keep at least 20% of a drive as free space, should it be within a partition, or just leave some unpartitioned space on the drive?
The 40GB 5-channel version has read speeds of 170MB/s. Does this mean that if the 10-channel G2 weren't inhibited by SATA 3Gb/s, it would probably get 340MB/s? Or is there diminishing returns on adding more channels.
Also, any price cuts on the horizon or just stagnating until drives start shipping with SATA 6Gb/s support.
It's possible. All of the high end SSDs are bound by the interface for sequential read speed. I'm not sure how fast the controller could go if it weren't bound by the interface, but over 300MB/s sounds quite likely.
Flash prices haven't really gone down lately from what I'm hearing. I don't know of any price cuts on the horizon unfortunately. Next year.
i don't think you will see any price cuts since intel can't make enough G2s.
I tend to agree with others regarding SATA 3. If intel is not rushing to enable SATA 3 on their ICH anytime soon, why would they be coming out with SATA 3 devices?
"Is Intel still my overall recommendation? Of course. The random write performance is simply too good to give up and it's only in very specific cases that the 80MB/s sequential write speed hurts you."
of the last review, is now a
"The write speed improvement that the Intel firmware brings to 160GB drives is nice but ultimately highlights a bigger issue: Intel's write speed is unacceptable in today's market."
Personally my pick is still the X25-M, it's had a better track record than any of the other controller makers. The problem for Intel is that Indilinx performance and feature set have both gotten a lot better.
You now get TRIM, reliable performance, competitive pricing and generally everything but a super dynamic controller that's constantly trying to maximize performance. Depending on the usage scenario though, the latter may not matter as much (see my Photoshop results from the last review or the heavy trace results from this review).
My position hasn't really changed, just highlighting the importance of the issue so that hopefully Intel's value SSDs next year aren't just rebadged G2s :)
I think his previous review is consistent with this one. His ultimate recommendation between Intel and Indlinx drives has remained unchanged. That is, they are close enough that the size differences should be the decider for a potential buyer.
As it stands, in most normal usage scenarios the statement from his last article still applies. There are very few times where you are going to be constantly tapping the 80MB/s barrier. More often you will be limited by the other SSD's lower random write speeds.
What his own benchmark can show is that in uncommon circumstances you can sway the bar back towards parity between the brands. I would argue, however, that the "heavy" usage scenario is not a practical one for an SSD. Just read his description again and look for all of the places the average user would NOT be putting that data onto an SSD:
[quote]
In this test we have Microsoft Security Essentials running in the background with real time virus scanning enabled. We also perform a quick scan in the middle of the test. Firefox, Outlook, Excel, Word and Powerpoint are all used the same as they were in the light test. We add Photoshop CS4 to the mix, opening a bunch of 12MP images, editing them, then saving them as highly compressed JPGs for web publishing. Windows 7’s picture viewer is used to view a bunch of pictures on the hard drive. We use 7-zip to create and extract .7z archives. Downloading is also prominently featured in our heavy test; we download large files from the Internet during portions of the benchmark, as well as use uTorrent to grab a couple of torrents. Some of the applications in use are installed during the benchmark, Windows updates are also installed. Towards the end of the test we launch World of Warcraft, play for a few minutes, then delete the folder. This test also takes into account all of the disk accesses that happen while the OS is booting.
The benchmark is 22 minutes long and it consists of 128,895 read operations and 72,411 write operations.
[quote]
How many people are going to be opening photoshop pictures FROM the SSD and saving them TO the SSD? Same goes with Windows Picture Viewer. Most people will have their pics on a standard large mechanical HD which makes this a moot point. Downloading large files from the net? Not a chance that would be going to my SSD! Bottom line is this is a great HELL test, but nowhere close to real-world usage.
The light-usage probably fits 90% of people 90% of the time (completely made up statistic).
With all of that said, I wish Intel's 80gig drive had closer sequencial write performance as the Indilinx drives, as it would make the decision a complete no-brainer. Looking at all the data, and planning on buying a ~$250 drive this holiday season has me very likely picking up an 80gig Intel as for me the price is about as steep as I'm willing to go for essentially a boot drive, with performance that for the majority of the time is at the top of the pack. It would be WAYYYY more difficult to decide if both Intel/Indilinx had the same drive sizes but fortunately we don't have to worry about that (at least for now....)
You have listed the 80gig Intel MLC drive twice in each chart. I'm assuming one of them shouldn't say (TRIM)?
Other than that, GREAT article! Those Kingston drives look to be the PERFECT inexpensive replacement for netbooks/cheaper laptops as the performance is vastly superior to cheap mechanical HD's.
"The G1 is more resillient (SIC) than the G2 when it comes to performance degradation over time since it doesn't have TRIM."
Is it just me or does this sentence make no sense? I thought the whole purpose of TRIM was to make the SSD more resilient to performance degradation. Seems like it should be:
"The G1 is less resilient..."
"None of the other Indilinx manufacturers have Windows 7 TRIM support yet."
Crucial has the TRIM firmware (1819) available for download for their m225 series SSDs.
Since I've been wondering about the free space issue, will there be a test where a drive is benchmarked vs. the amount of free space available? My guess is that the graph will be an exponential decay with the "noticeable" regime at 20% free space. Anyone cares to do this (say, at 90%, 80% ... close to zero free space)?
The only option you, a G1 owner, have to make your drive's performance back to ~100% is to do a full format of the drive. I'm pretty sure it would have to be done very frequently, each time you write 80/160GB of data to the drive... nothing else you can do, sorry =(
Intel's wipe ultility (SSD Toolbox) only works with the G2 (for use with Windows XP/Vista) and TRIM only works with the G2 as well (for use with Windows 7-to replace the Toolbox)... =(
Another great article by the one and only Anand Shimpi, king of SSDs
So, basically Intel made a very stupid move by only allowing SSD Toolbox AND TRIM on G2 drives. I would like to know if its due to architectural changes (perhaps?), maybe you know it but can't reveal it to anyone... if I weer you i would edit out the part when you say the G3 are coming next year... or else Intel will kill you, they just lost a custumer for the G2 (me)
But would it make sense for them to release the SATA 6 G3 drives next year if their on-board ICH10 and MCH P55 does not support it? Maybe the P57 and X68 will? Who knows?
Its absolutely ridiculous how intel is screwing those customers who bought their x25m g1s by specifically going out of their way to prevent trim support. At least give a manual wiper tool, you jerks. Geez!
I know that Toms Hardware has commented on the fact that Intel is not providing a firmware update for the X25-m first generation drives... but I think TH is in a position to rake Intel over the coals for this terrible decision! Toms Hardware caters to computer enthusiasts, exactly the type of customers who buy first generation hardware like the X25-m G1... so for Intel to NOT issue a TRIM firmware update for X25-m G1 drives should ahve Toms Hardware users up in arms, and certainly questioning whether Intel's support for the G2 drives will stand the test of time!
If you think Intel is wrong to abandon the G1 drives by not providing a TRIM firmware update, please consider signing this petition!
Agreed 100%, I purchased an x-25m G1 at a premium thinking that intel would take care of me once TRIM support was available. Since I heard that intel wouldn't be giving any G1's trim for no reason other than to screw over the consumer into buying an updated drive, I have only purchased Indilinx based SSDs (over 4 now) and I recommend that everyone else do the same!
They also promised that the 80gb G2 SSDs would be sold at a price point of $229, but due to the demand they are gouging everyone selling them at price points of 299 and above. (can't blame them, it's economics, but don't PROMISE it at $229 and not fulfill it.)
People buy stuff where the performance is, but I won't buy an intel SSD again unless they release something that performs significantly better than the competitor, but the point is that you should buy indilinx based drives because they are just as good and you won't be supporting Intel's poor decision with regard to the G1 customers.
great SSD articles you guys have been putting out latly. but it would be great to see a corsair SSD in these benchies. alot of people talk about them and they seem blazing fast even faster then the intels from what iv read around on google but id like to see one of my populer tech websiteas to confirm this. i like this article here but any chance you can throw a corsiar in the lineup :)
You pay the price for being early adapter. I mean even now with the G2 a colleague of mine just bought back in August, the thing still had to be RMA'd.
I'd never use something like this for work related use unless it's been tested by many guinea... err I mean other consumers before it's proven to be cheap and reliable.
If you don't have the software to image and restore your drive, get it now! It's an order of magnitude less expensive* than the SSD, and it's the fastest, cleanest, and most reliable way to recover from a failed hard drive.
*If you run OS X, you can use Time Machine or Disk Utility. If you run Linux or BSD, you can use dd from the terminal.
Do what I did. Buy a Gen 2, clone your drive, and sell the Gen 1 on ebay while it's still worth as much as it is. You will easily get close to $200 for the 80 gig version. I lost $30 upgrading, which was very worth it for TRIM support and the newest hardware.
Acronis True Image is probably the easiest for Windows, especially if you're cloning to identical sized drives. With the trial you can do this for free, I believe Seagate and WD have simplified free versions if you have one of their drives installed.
Don't mess with any of the backup/restore options, go to the Utilities and choose Clone drive, then select destination and source drives. Takes about 10-20 minutes depending how much data you have and is effortless.
No kidding. I clone my drive every once in a while on my fileserver with dd. Why it's made so difficult in every other OS is beyond me. I really shouldn't need to buy software in order to make a copy. It's a pretty basic function, really.
Just that the aforementioned software isnt primarily a cloning software but a backup and image software. Cloning a disk is just a minor function of it. And no, cloning a disk isnt a worthwhile backup strategy for most people.
That said, I backup my system and data on a daily basis, but I only clone a drive once maybe twice a year. I wouldnt mind if it was built into windows, but its far from a necessity.
No kidding. Early adopters really shouldn't complain, that's what you get for being the "elite." LMAO.
As I've said before, just wait til early next to summer and the tech will be more mature. Yea, go ahead and make your argument that you can wait til next year but by then they'll be new products. That's cool...but I think ppl saying this have no clue about what's really happening around the technology.
Regardless, sounds nice to be saying it so woohooo be the first!
If they were not able to put it in due to hardware problems its one thing... but those drives cost 665 dollars in the beggining. I bet that if you had one you would not be happy :|
Agreed, Intel will certainly meet more resistance in the future for their emergent technologies and undoubtedly lost the support of many of their most dedicated supporters with this decision.
If there were some technical reason or incompatibility behind the lack of TRIM support on G1 that'd be one thing, but from what has been published by AT and others, its nothing more than indifference on Intel's part preventing support.
Eloquently stated, i bought two of the G1 drives. I won't be suckered into supporting any of intels new tech early again and I'll be taking a much closer look at their competitors products when purchasing more established tech.
Nanochip went belly-up and the IP was auctioned off. Ovonyx turned into Numonyx and PCM is still an emerging alternative to other technologies. The biggest issues relate to the limitatios in process technology, i.e, at about 20 nm geometry, the required write currents are expected to actually melt the metal interconnects. There are other technologies such as resistive memory as developed by 4-d-s and organic memories such as the ones from Zettacore that may become very attractive.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
162 Comments
Back to Article
RAWIRON - Saturday, December 19, 2009 - link
I've read about some Macbook hacks for the SSD user, but still wonder how I should make my SSD's "new" state last longer.I do still wonder if I should use the "Secure Empty Trash" in Snow Leopard? At first, I've partitioned my drive for all available space, but then reduced the HFS partition to 68 Gb or so - will this prolong my drive's "new" state because of Intel's algorithms?
Should I use "Erase Free Space" in Disk Utility?
Thanks!
kunedog - Saturday, December 26, 2009 - link
I would say NO. Secure Erase (of used or free space) will generate (many) more writes which can only make the problem worse, "using up" your reserve space much more quickly than before. At least that's what my intuition says. Maybe somenoe who knows for sure will chime in.kunedog - Friday, December 4, 2009 - link
The Kingston 40GB is available for $130:http://www.newegg.com/Product/Product.aspx?Item=N8...">http://www.newegg.com/Product/Product.aspx?Item=N8...
I wonder how many, if any, were sold at $85 (or even $115). I also wonder how this article is repeatedly bumped for minor (and predictable) updates while the incorrect pricing predictions are never acknowledged. The Intel MLC SSDs continue to list for much higher than the given "expected" launch price . . .
http://www.anandtech.com/storage/showdoc.aspx?i=36...">http://www.anandtech.com/storage/showdoc.aspx?i=36...
. . . yet that article was never bumped to the top. Even the G1s were priced higher! Props to Anand for keeping the SSD companies honest on performance, but why can't the same be done for pricing?
If I were going to release/retail a new SSD, I would definitely call Anand ahead of time and tell him an expected price about 50-75% of the real one so he would blindly print it and create demand.
mohsh86 - Thursday, December 3, 2009 - link
lol the promised to deliver it by the end of Novmber, they didn't but they changed the date of release :P i've been checking every day..The Firmware Update tool 1.5 is now available, posted (according to intel in 30/11) the SSD toolbox still unavailable..
did any one try it ?!
i have the firmware the one before the bricking firmware, i guess it has the letter g, should i upgrade ?!
mohsh86 - Tuesday, December 1, 2009 - link
It finally came ! The End of November, The 1st of December..No Trim Firmware, No Intel SSD Toolbox, no RAID support for toolbox
Dverez - Tuesday, December 1, 2009 - link
The new Firmware is out and works properly.To test SSD this is a good program, I think.
(http://alex-is.de/PHP/fusion/downloads.php?cat_id=...
mohsh86 - Tuesday, December 1, 2009 - link
It finally came ! The End of November, The 1st of December..No Trim Firmware, No Intel SSD Toolbox, no RAID support for toolbox
Sind - Monday, November 30, 2009 - link
Wheres the firmware update, it's Nov 30th!xpclient - Wednesday, November 25, 2009 - link
Please test 2 X25-Ms with Intel Matrix RAID10 and other Matrix RAID arrays the moment Intel drivers with TRIM come out.Mygaffer - Tuesday, November 24, 2009 - link
Another great article, thank you for all the hard work and insightful analysis. Articles like yours help keep the industry honest and responsive.I can't wait until they become cheap enough for me to buy two for a RAID 0, and of course for Intel to release new drivers to support the TRIM commands.
mrt2 - Monday, November 23, 2009 - link
Ok so with all this new data in mind, what's currently the fastest, least performance degrading 256GB SSD you can get for under $1000? I'm currently running Snow Leopard on a dual quad core Mac Pro... I'm not sure Snow Leopard supports any of the new technologies like TRIM or not... but whatever the best 256GB SSD I can get, that's what I'd like....rdhir - Wednesday, November 18, 2009 - link
WARNING on Intel Matrix driversI just wanted to add a word of warning over drivers and SSD. I just built a new i5 750/P55 system with a Corsair X64 drive (Indilinx). While the system worked fine in bits it frequently froze. I tried lots of things to cure it, but in the end the fix was simple. I HAD to use the Intel drivers to avoid intermittent freezing.
Because I had made the SSD the C:drive and used a 1TB Samsung F2 as the D: drive, I had not wanted to lose TRIM support, until the new AHCI drivers came out, but MS drivers don't work. reference article is here http://derek858.blogspot.com/2009/01/windows-7-int...">http://derek858.blogspot.com/2009/01/windows-7-int...
Reading around some of the othere references Its inconsistent and depends on how the drives implements the command set so your mileage may vary but I think its a big issue.
I think we could all do with an article on how best to configure an SSD based system. I decided that I wanted to move "Users" and all associated hidden directories to the D: drive (a 1TB HD) leaving the 64GB for Windows and Apps. This would mean it would never get too full.
I'm amazed at how difficult Microsoft make it to shift the "Users" directory to another drive. In fact they have a Knowledge base article saying not to - Linux easy, just map /home elsewhere.
In the end I followed some instructions to create a junction from C:\Users to D:\Users, but its non-trivial and involves opening a command window in the middle of the installation process.
Yes I'd like to have done it with an autounattend.txt but the MS KB articles says it won't really work.
mapesdhs - Thursday, November 19, 2009 - link
> In the end I followed some instructions to create a junction from
> C:\Users to D:\Users, but its non-trivial and involves opening a
> command window in the middle of the installation process.
I can't see this sort of thing becoming any easier until MS ditches
the archaic idea of drive letters and switches to a more natural
unified file system such as is used with all UNIX OS variants.
(oh look, the pits of hell seem to be getting a tad icey...)
Ian.
MadAd - Thursday, November 19, 2009 - link
"I think we could all do with an article on how best to configure an SSD based system"That would be great
CuriousMike - Monday, November 23, 2009 - link
" "I think we could all do with an article on how best to configure an SSD based system""That would be great " "
Thirded.
The variety of do's and don'ts (and conflictions) are numerous:
-do disable pagefile. don't. oh, which OS?
-don't defrag disk. do. oh, which OS?
-use ram disk for temp files. don't. no, wait, do.
ol1bit - Wednesday, November 18, 2009 - link
I can't believe they don't have firmware for my drive to trim it... I just bought the sucker 3 months ago.....Intel X25-M SSDSA2MH080G1 80GB SATA II MLC Internal Solid state disk (SSD)
dagamer34 - Wednesday, November 18, 2009 - link
No TRIM for G1 drives. Sorry.yacoub - Tuesday, November 17, 2009 - link
An old article reposted as new simply because of a couple-sentence-long update? Interesting.Also, $85? No, try $130 and no MIR. Thanks ScrewEgg!
http://www.newegg.com/Product/Product.aspx?Item=N8...">http://www.newegg.com/Product/Product.a...pk=kings...
7Enigma - Tuesday, November 17, 2009 - link
If the only update was the 2 short blurbs at the top of the first page I don't think this entire article should be reposted. A simple statement with a link to the previous article would be fine. I just wasted a couple minutes going through each page realizing nothing was new....not cool. I was psyched for another SSD article Anand!max22 - Tuesday, November 17, 2009 - link
I installed Intel's pulled firmware fine on my G2 drives. Nothing happen to them at all. I think the whole issue has been blown way over the top.dcljpc - Saturday, November 14, 2009 - link
On that intel firmware that provide trim support but bricked people's drives, now intel said it is a problem specific to Win7 64bit. Does this mean that if you update the firmware from within Win7 64bit there could be a problem, or does it mean even if you updated the firmware in Vista or Win7 32, and then install Win7 64 on it, it could still cause a problem? Does anyone know the answer?I just bought a X25-M G2, I plan to use it in a new computer. Can you just update the firmware on an older computer first, and then install WIn7 64bit on it?
Thanks.
jaydops - Friday, November 13, 2009 - link
I am looking to purchase a ssd for a Dell Studio 15 i7 with 4gb ddr3 1333ghz.Anand recommends Indilinx (OCZ or SuperTalent) or Intel.
1) Are all OCZ drives good or only Vertex drives? If only Vertex then those are really expensive on newegg.
2) Why was the Corsair P256 not recommended? It seems to have good numbers in the charts. And newegg has a reasonably priced 128gb version here: http://www.newegg.com/Product/Product.aspx?Item=N8...">http://www.newegg.com/Product/Product.aspx?Item=N8...
3) The Dell Studio 15 has a sata connection, will that work ok with sata2 ssds?
crobb100 - Tuesday, November 10, 2009 - link
No Kingston SSD for sale at NewEgg so far. Where'd it go?dullard - Tuesday, November 10, 2009 - link
It is at Newegg. But it isn't at the MSRP.http://www.newegg.com/Product/Product.aspx?Item=N8...">http://www.newegg.com/Product/Product.a...SSDNow%2...
I'm considering getting it for a new Win 7 build. But are SSDs ready for prime time yet? Or should I wait a bit longer?
lordstryker - Monday, November 9, 2009 - link
I was curious about the possible setup of these Kingston drives in a Raid 0 format. If they are roughly half the performance of an intel 80GB drive, it would be reasonable to think 2x of these Kingston drives would be comparable to a single Intel 80GB drive. However, I am not familiar with TRIM support on a Raid setup. Is there any info on this?spaceB - Monday, November 9, 2009 - link
Hi all,I have a x200 Thinkpad with a SSD inside. The SSD drive is a 128GB Samsung (OEM MMCQE28G8MUP-0VAL1). I noticed that the drive becomes slower after three month of use. So I use WIN XP SP3. Is it possible to use Trim or Garbage Collection in any kind of way on this older Samsung SSD drive? The Samsung support homepage and google couldn’t answer my question.
Big thx in anvance!
GreeTz
space
maraz - Wednesday, November 4, 2009 - link
Hi,I was very interested to see results from trace-driven I/O tests. Is the AnandTech Storage Bench mentioned in this article generally available?
Best regards,
Manolis.
Steney - Monday, November 2, 2009 - link
What would the performance of two of these drives in a RAID 0 array be like? Would there a single 80GB drive for $170 that would be better?abuda - Monday, November 2, 2009 - link
just found this stuff in mwavehttp://www.mwave.com/mwave/SKUSearch_v3.asp?px=IM&...">http://www.mwave.com/mwave/SKUSearch_v3.asp?px=IM&...
INTEL 160GB SSD G2 just US 599 with free shipping
kunedog - Monday, November 2, 2009 - link
That's only $159 MORE than the predictions we got from Anand over 3 months ago:http://www.anandtech.com/storage/showdoc.aspx?i=36...">http://www.anandtech.com/storage/showdoc.aspx?i=36...
Wow, indeed.
I wish Anand would acknowledge these extremely high prices and how wrong the prices in that article were (preferably in a follow-up). Especially since he is again making statements about Newegg's possible future pricing (of the Kingston).
Celeus - Monday, November 9, 2009 - link
No comment regarding Anand, but the quote of $85 seems to be based on a press release from kingston.http://www.kingston.com/press/2009/flash/10c.asp">http://www.kingston.com/press/2009/flash/10c.asp
Now, I've been trying to find these for that price at Newegg, and can't. The drive now shows up (just the 2.5" one, not the one including the bracket) but for $124.99 before rebate $104.99 AR.
I bet this is an error they will fix, as $104.99 Before Rebate would make $84.99 AR, which is what Kingston mentions in their press release.
Rebate looks good for 2 per person, so I plan on buying two for a new Windows 7 box.
virtualgeek - Saturday, October 31, 2009 - link
Disclosure, I work for EMC (an enterprise information infrastructure - which includes all sorts of storage arrays).Looking at some of the comments, I'm not sure if people understand the impact of the x-25 getting down to the prices they are (both via kingston OEM and the 80/160GB drives).
Flash will rapidly replace all high-performance disk use cases. There will only be room for very large SATA and SAS disks, and all high performance use cases will be dominated by Flash.
Some people don't understand that even at TODAY's prices, for some (many) use cases (just not consumer focused ones), they are more economical. For example, the kingston drive, measured in IOPs/dollar is 37x **cheaper** than a 15K SAS drive. That's the acquisition cost. When you think that it would take 37 15K SAS disks, consuming more power, space and cooling - you can immediately imagine the impact this has on the enterprise storage market.
The rapid price decline we've all seen over the last year (a 40GB MLC drive cost ~$900 at the beginning of the year) means that in the next year or so, we'll have SSD with the $/GB of a large, fast rotating magnetic media disk, but still 100x better random IOps/$
While perhaps PCM non-volatile storage will eventually replace flash, that will happen well after flash replaces spinning media.
Great article Anand!
I did a post on it here...
http://virtualgeek.typepad.com/virtual_geek/2009/1...">http://virtualgeek.typepad.com/virtual_...ill-chan...
iwodo - Friday, October 30, 2009 - link
If V Series stays at $85, then making a Raid with 2 40GB V will be an VERY attractive option.So may be do a review on 2 40GB V series with Intel Software Raid? as well as other raid card?
excalibur3 - Thursday, October 29, 2009 - link
Thanks for the great article! I am just curious if the Kingston UltraDrives can use the same firmware update as intel and can be considered the same drive or if it is something like with OCZ and Super Talent in that they have separate firmware updates. Would the Kinston be a cheaper alternative to the Intel?Saturn1 - Thursday, October 29, 2009 - link
Is there a way to get the manual trim to run from the toolbox if you do not have that last firmware update?linster - Thursday, October 29, 2009 - link
Apparently Intel pulled the iso. Here's what you get after you click on the link,"02HA Firmware Upgrade for Windows 7* Systems - Unavailable
Intel has been contacted by users with issues with the 02HA firmware upgrade on Windows 7* systems and are investigating. We take all sightings and issues seriously and are working toward resolution. We have temporarily taken down the firmware update while we investigate.
Thank you for your patience."
Shadowmaster625 - Thursday, October 29, 2009 - link
Can you guys do some testing on the Runcore Pro IV 16GB PATA SSD? It supposedly uses the indilinx controller.coconutboy - Thursday, October 29, 2009 - link
Good article Anand, and I like the new tests you guys came up with. I have a request for a new type of test and I know others in the various hardware forums have these same questions.SSDs are too small for many of us to use as the lone drive in our system, and we thus have to combine an SSD as an OS/app drive w/ a traditional hdd for our file storage. Given that this is a reality for many of us eager to jump into SSDs, it'd be great if we could see a test that demonstrates a real-world scenario of users loading DBs/pictures/videos etc from our storage drive while still running our apps/OS from the ssd.
I know it's probably not something you guys could do regularly because the amount of testing would be a huge burden, but perhaps you fellas could do a one-off article just to highlight the differences? Maybe show a couple different usage scenarios such as:
~ a budget/midrange P55 setup w/ OS/apps on a small Indilinx ssd and a single 1TB drive for storage
~ an older Core 2/AMD system with the new Kingston offering used for OS/apps and the storage drives being pair of 640GBs in RAID 1.
I think this kind of article would be very useful to your readers and I know there's lots of us in various forums who are hesitant to jump into a hybrid ssd/hdd setup because we're unsure of exactly how it will affect us.
coconutboy - Thursday, October 29, 2009 - link
Still not a lot of info out there showing real-world usage scenarios w/ an ssd as the os/app drive and a regular 7200rpm or two as storage, but perhaps for others interested in this kinda testing something like this will suffice. From the MSI p55-gd65 review-http://www.anandtech.com/mb/showdoc.aspx?i=3655&am...">http://www.anandtech.com/mb/showdoc.aspx?i=3655&am...
krumme - Wednesday, October 28, 2009 - link
How about this collection:http://www.anandtech.com/storage/showdoc.aspx?i=34...">http://www.anandtech.com/storage/showdoc.aspx?i=34...
It measures where the speed matters most (sorry can not get link to work)
Is the world that different today than 2009?
We need to se results balanced like this
I do get a feeling that the 2009 test suite is heavely favoring the g2 drives
It is plain and simple what takes time. Not iops, we need seconds. And seconds where you can feel it; fx. seconds for diferent types of filecopying, handling large files, working while making backup, working while having virus scan.
The g2 is way overhyped, the samsung way underrated
Reading the results you can have the impression the velociraptor doesnt work for desktop use
The need for trim is overhyped and having severe consequenses
And the results from defective and bad bios updates from indilinx and espec. Intel, was to be seen in advance.
Are we going to se some new benchmark suite when gfx lrb arives??
The end results. Ordinary desktop people having hdd for servers and gfx for research.
It doesnt make sense.
Some bad thought creap into my mind. Help me. Is the Intel marketing a problem, if you dont behave? :) - or did they just give you the random 4k engineering hammer?
Anand - we need another ssd article !!! :)
Take care
strikeback03 - Thursday, October 29, 2009 - link
The 4k thing was a way to quantify the stuttering problem on some drives, as if he had just said "the drives stutter" but only shown objective results that look good (due to sequential write speeds), then people would have claimed he was making stuff up to be an intel fanboi as well.AnnonymousCoward - Wednesday, October 28, 2009 - link
Human-observed delay time is really all everything comes down to. 4kb performance and IOPS are interesting to look at, but they're severly inadequate. If I remember right, RAID0 tends to look awesome in benchmarks, but offers virtually zero real-world benefit.So screw the benchmarks get and out a stopwatch. Measure what matters: boot time, time to launch various applications, media encoding time, and the time to copy files. Quantified seconds are a helluva lot more meaningful than IOPS.
strikeback03 - Wednesday, October 28, 2009 - link
You mean something like this: http://www.anandtech.com/storage/showdoc.aspx?i=34...">http://www.anandtech.com/storage/showdoc.aspx?i=34...AnnonymousCoward - Wednesday, October 28, 2009 - link
Yep, something like that, but with more categories. I think this is a good set:1. Boot time
2. Time to launch applications
a) Firefox
b) Google Earth
c) Photoshop
3. Time to open huge files
a) .doc
b) .xls
c) .pdf
d) .psd
4. Game framerates
a) minimum
b) average
5. Time to copy files to & from the drive
a) 3000 200kB files
b) 200 3.5MB files
c) 1 2GB file
6. Other application-specific tasks
Imagine if we had a table of this data and all of the SSDs and a conventional HDD. That would be so much more useful than all the charts in the existing article.
buzznut - Wednesday, October 28, 2009 - link
Finally, an SSD that I can afford and makes sense to me. Anyone can afford a POS SSD drive, but this one doesn't look crippled too much. The writes I'm not too worried about, but the stuttering and performance degrading is not what I want to pay a premium for.I'm hoping by Jan-Feb (when I could afford one) that Kingston will have the trim support and upgraded firmware. Then I'll jump on it!
soltari - Thursday, December 10, 2009 - link
Just picked mine up yesterday.. Kingston 40gb bootdrive desktop kit.. Has 2CV102HA firmware .. i.e. TRim support already.. be on the lookout for it. .dont know if its all drives now or just some.http://img209.imageshack.us/img209/8616/img6915.jp...">http://img209.imageshack.us/img209/8616/img6915.jp...
http://img301.imageshack.us/img301/7214/img6920m.j...">http://img301.imageshack.us/img301/7214/img6920m.j...
box note on firmware and hdtune info screenshot..
Cavicchi - Wednesday, October 28, 2009 - link
What I read from Intel about the Optimizer suggests it is not needed in Windows 7 or not needed to actually run it. Anyone understand what Intel is saying here:"When using the latest Microsoft Windows* 7 operating system with Microsoft AHCI storage drivers the OS will contain native support to execute the Intel® SSD Optimizer on an Intel SSD without requiring any user interaction."
"Microsoft Windows* 7
Microsoft* AHCI
Native OS support (Intel® SSD Toolbox not required)"
strikeback03 - Wednesday, October 28, 2009 - link
No, that was part of the point of this article, that TRIM is supported natively in 7. It is not in earlier OSs, so they have the manual TRIM tool.drwho9437 - Wednesday, October 28, 2009 - link
Are all the Kingston SSDNOW drives X25 based in design with just different amounts of memory?Tuvok86 - Wednesday, October 28, 2009 - link
* Kingston SSDNow E Series- (Intel Controller, 50nm Intel SLC NAND, 16MB Cache)
* Kingston SSDNow M Series
- (Intel Controller, 50nm Intel MLC NAND, 16MB Cache)
* Kingston SSDNow V+ Series
- (Samsung Controller, Samsung MLC NAND, 128MB Cache)
* Kingston SSDNow V Series
40GB - (Intel Controller, 34nm Intel MLC NAND, 16MB Cache)
64/128GB - (JMicron Controller, Toshiba MLC NAND, No Cache)
Tuvok86 - Wednesday, October 28, 2009 - link
oops, I meant40GB - (Intel Controller, 34nm Intel MLC NAND, 32MB Cache)
Tuvok86 - Tuesday, October 27, 2009 - link
I don't like how Kingston labelled this drive as v-series, this may lead people to think that v-series 64gb and 128gb are good as well, while they end up buying a JMicron crap...I know that the v-series stands for value but I'd expect a kind of consistancy in series parts.I'd expect 64 and 128 drives to perform equally or better than the 40gb part, but it wouldn't...perhaps kingston had to find a sloppy way to get rid of those unsold "value" drives...
Regarding ssd reliabilty brought up in the recent posts, I'd be pretty confident to put an SSD in a home pc anytime.
Man, they are used for SERVERS (well, SLC drives actually, but the story is the same), one of the most mission critical environment out there.
Dangers are ahead only if you want to mess trying TRIM, fw updates or any other topic brought up every now and then, but if you wait a while and resist early-adopting new features for a couple days, problems are issued quickly. Anyway backup is just a click away.
clarkn0va - Wednesday, October 28, 2009 - link
I would like to see tests on the JMicron-based Kinston V Series. Supposedly their newer controller resolves the stuttering and random access bottleneck of the gen-1 SSDs.I installed one in a budget build for a customer and was very impressed with performance in a modest use case scenario (with Windows xp and an Atom 330). I would have no reservations using these drives in future builds, although the new 40GB model is an interesting proposition.
clarkn0va - Wednesday, October 28, 2009 - link
I do agree though that the V Series branding is confusing and misleading, in light of the different controller. True to Intel tradition.The0ne - Tuesday, October 27, 2009 - link
Seems not enough testing was done for them to have data corruption, again. These kind of issues shouldn't really surface if proper testing were carried out :/mantis2000 - Tuesday, October 27, 2009 - link
"Would I recommend waiting until next year to buy? This is one of the rare cases where I'd have to answer no."Given all of the serious reliability issues -- including today's latest Intel firmware debacle -- it's quite clear that SSD are not ready for prime time. Over and over again, we hear stories about disks not living up to their potential due to bad drivers or firmware, and there have been far too many cases of total failure with attendant catastrophic data loss.
How can Anand recommend using a SSD on a primary machine?
Wait at least a year for the bugs to be worked out if you value your data.
Griswold - Wednesday, October 28, 2009 - link
Bullshit. What reliability issues? What firmware debacle? The handful people with potentially bricked devices after the flash hardly qualify as a debacle. I would guess the success rate is over 90%. Flashing is always a risk, you know...Intel did the right thing by pulling the firmware, though. They'll look into it. But i wouldnt be surprised if it wasnt an error on their end - they spent lots of time making this firmware. Much, much more than the garbage the competition throws and its customers every week.
drwho9437 - Wednesday, October 28, 2009 - link
I find the statement that flashing is always a risk on a drive based on flash memory very ironic.I should think they could spare the space for a backup firmware if it is as you say.
GullLars - Tuesday, October 27, 2009 - link
I've heard this argument for years now, along with quite a few others that have died away. "SSDs are not ready for prime time". Wich SSDs, and for what usage are they not ready compared to HDDs?Unless you yourself upgrade the firmware of an SSD whitout first waiting and making sure the upgrade is safe, there is no risk by using SSDs that is not greater with HDDs. SSDs generaly are much more reliable and rugged, and when they near the end of their natural life, you will see the raw data failrate predictably increase until it hits the point where the ECC can't do the job anymore, before wich point you take backup and buy a new one.
Anyone who keeps invaluable data on just one physical medium whitout backup would be a fool to think it is safe. Use the SSDs for OS + programs and have a RAID in a redundant mode with offsite backup for your valuable data.
Anand is IMHO right to recommend buying an SSD now and not waiting because we have passed the point where the ratios between price, capacity, performance, and reliability make them far superior to harddisk for boot drives. I've had SSDs for over a year in my computer, and payed a hefty price for being an early adopter, but it was well worth it. The prices, specs, and reliable market today almost make me laugh out loud when people say "SSDs are not ready yet". Claiming "(many) people are not ready for SSDs yet" would be far more accurate.
sotoa - Tuesday, October 27, 2009 - link
The firmware was there this morning and now it's not available? Anyone else see this?mbreitba - Tuesday, October 27, 2009 - link
Looks like a lot of people are having problems with it, and Intel has pulled it :http://communities.intel.com/thread/7693?start=45&...">http://communities.intel.com/thread/7693?start=45&...
Griswold - Wednesday, October 28, 2009 - link
If you actually look at the number of people with the issue and filter out the chit-chat, its not "alot" actually. However, my flash went just fine.Thats the risk with flashing firmware regardless of what device it is. Theres always the chance to brick it. Thats also why I dont understand why some people flash every fucking piece of hardware whenever theres a new firmware available - ok, this is a different case, here it makes perfectly sense.
strikeback03 - Wednesday, October 28, 2009 - link
Has anyone had problems flashing a drive before they put any data on it? I have a new G2 still in the box that I just haven't yet had time to do anything with, was planning on flashing it then loading Win7 and Ubuntu 9.10 this weekend, have there been any reports of issues when loading an OS after flashing?UltraWide - Tuesday, October 27, 2009 - link
Great article, I like the definite conclusions and recommendations. Keep up the good work!nicolasv - Tuesday, October 27, 2009 - link
Hi AnandIn 'The SSD Relapse' you state that the G2 "doesn't drop in performance when used...at all." Yet in 'The SSD Improv' the 80GB G2 with TRIM firmware drops more than 60% in the 4KB random write test.
Granted, the charts and figures used in 'Relapse' to back its claim are for the 160GB G2 and these results are for the 80GB G2. What do you attribute this difference in performance to, the new TRIM firmware?
As a Mac OS X user, at this point, I feel like I can only really consider drives that perform well without TRIM, especially in the <= 80GB range, so would appreciate your feedback.
Cheers
Nicolas
7Enigma - Tuesday, October 27, 2009 - link
I believe it is because in that particular test he is writing to the ENTIRE 80gig drive, so it's not that TRIM isn't working per se, rather that there is no free space to allow TRIM to do anything at all.7Enigma - Tuesday, October 27, 2009 - link
As an addendum to my previous post, you can see that after file deletion the performance goes back to virtually new. So in a sense the performance of the drive never goes down, UNLESS you simply delete a partition rather than erasing the data on that partition FIRST and then deleting it.I think that's probably something that needs to be implemented with an updated driver, or at least a warning box that comes up saying "deleting this partition without formatting will hinder performance".
GullLars - Tuesday, October 27, 2009 - link
I don't know if you have done extensive benchmarking with SSDs, but using IO QD = 3 for 4KB random in IOmeter don't yield representative results for intels SSDs. I have been benchmarking SSDs with some other guys for over a year now, and we have found that while Samsung and Indilinx scale from QD 1-4 and then flat out, Intels x25-M scales well all the way to around QD 10-16, and actually is capable of over 140 MB/s at 4KB random read at QD=64 in fresh state, and still over double your messured 60-64 MB/s in used state.In our benchmark thread, the records with x25-M from ICH10R before TRIM firmware are:
IOmeter
4KB random read QD=64: 40913 IOPS = 163,5 MB/s
4KB random write QD=64: 19360 IOPS = 77 MB/s
PCmark vantage HDD score: 43107.
The same guy that got the random read and PCmark scores above also got PCmark vantage HDD score 120374 with 3 x25-M gen1 from ICH10R.
In other words, the results posted for x25-M in this review are either in "used state" or below par.
Voo - Tuesday, October 27, 2009 - link
You know Anand trys to simulate REAL situations and a QD of 10-16 is absolutly unrealistic for a home user - don't even talk about a QD of 64. Maybe the Intel SSDs shine there, but it's just of no interest.GullLars - Tuesday, October 27, 2009 - link
In his traced "light" benchmark with mostly single-tasking the average queue depth is 6.09. If you take into account that most of the sequential read and write operations (although they are under 30%) don't generate a queue, and that 4-16KB IOs often come in bursts with more queue depth, you can easily get a QD of more than 10 with relatively mild multitasking if it involves disk access. A harddisk's cache is regulary used as a buffer for the write portion of these, but if the cache is full your system almost freezes because of the harddisks low IOPS. An example would be if you try installing windows updates or a program while running a virus scan and listening to music or extracting a compressed archive, you will quickly notice if you try this on a harddisk.Voo - Tuesday, October 27, 2009 - link
Ok I was a bit surprised and tried it myself with my 160gb Intel G2 (no firmware update yet)Copy 10gb file onto drive, MSE full scan and extracting a 4gb file with 7z while the usual stuff (FF, PS, eclipse and winamp) running in the backgorund and I got a maximum QD of under 8, which is higher than I expected it to be, but still far away from a QD of 16.
And when doing more multitasking stuff I'd probably tax the rest of the system more than the SSD so at least for my usual behavior a QD >8 is unrealistic, but I see your point.
There are probably some people who could get a QD of 10-12 in a normal workload, so that could be interesting for them :)
GullLars - Wednesday, October 28, 2009 - link
There is a big difference between average queue and queue during operations. The IO queues are relatively low for all the background operation the OS does, and sequensial read/write, but when you do things like open administrative tools, open a program, or install something, you will se queue spikes of easily 50-60.After i commented this article i had a discussion with some other benchmarkers, and we found that normal use generates IO queue spikes almost whenever you interact with the OS. It doesn't take much effort to generate a queue above 100. Running a couple of virus scans on folders of small files while installing a program and opening a folder with lots of pictures will get you a massive queue.
NeBlackCat - Tuesday, October 27, 2009 - link
Nice work again Anand - I just don't bother reading SSD reviews/features on other sites any more.HOWEVER...
Many people reading this will be developers wondering which controller best favours typical development activities, eg. checking out and compiling large projects. It's a hard one to guess - lots of random reading, sequential reading, random writing (temp files) and sequential writing (output files).
So how about a "typical developer" benchmark of, say, compiling the Linux kernel while concurrently encoding some H.264, grepping all the sources for a string, and watching some por..er, I mean, youtube.
Has anyone else evaluated the best SSD type for s/w development?
clarkn0va - Wednesday, October 28, 2009 - link
Not a reply to your question, but I've taken on the habit of mounting my temp files to a ramdisk. Besides the speed improvement, this should help to reduce unnecessary write cycles to the SSD. Similarly, I run without swap (or page file), RAM being as affordable as it is.lamo - Tuesday, October 27, 2009 - link
TRIM support aside, I wonder how would a pair of $85 Kingston drives in a RAID 0 striped array compare against a $225 X-25M dollar/performance-wise.cosminliteanu - Monday, October 26, 2009 - link
So - "Your best bet is to install Windows 7 with your I/O controller in AHCI mode (for Intel chipsets)" this means you set in BIOS sata controller as AHCI not as IDE ?????Or just leave as IDE and after Win7 install, update intel storage driver as normal...? which is the best solution?
LuMax - Monday, October 26, 2009 - link
Can I use Intel IMSM RAID on just the data storage drives in my PC and retain TRIM functionality on the W7 Pro OS (X-25M G2) SSD?My new PC build plan was to use a single 80GB G2 for the OS and programs and two 2TB HDDs in a RAID1 array for my data/photo/video/music storage. Although I read the relevant part of this article 3 times I am still not certain if this is doable. Yes/No?
LuMax - Tuesday, October 27, 2009 - link
I heard back from Intel today:"Thank you for contacting Intel Technical Support.
I understand your question about the Intel(R) Matrix Storage Manager and the SSD. Since the drives you will be using in RAID are non-solid state disks, and the SSD you will be using will not be part of the RAID, there should not be any problem with TRIM."
somedude1234 - Tuesday, October 27, 2009 - link
I would not expect this to be possible. It sounds like you have 3 drives hanging off of the same SATA controller. You want 2 of them in RAID 1 using the IMSM driver, and one of them using the MS driver.In my experience it would be impossible to get Windows to load 2 different drivers for different ports on the same controller.
Even if they were physically independent but identical controllers (for example if you had 2 identical SAS cards plugged into the same system), as soon as you update the driver on one controller, a reboot results in all instances of that controller using the updated driver.
DoveOfTheSouth - Monday, October 26, 2009 - link
I'd like to know this too.I'm considering a similar set-up (SSD for systems drive, RAID for data storage HDD) - seems best of both worlds.
Pandamonium - Monday, October 26, 2009 - link
The 2nd gen X18-M isn't available as far as I can tell. It kind of shafts us tablet and ultraportable users out of the SSD market for another round =/Cavicchi - Monday, October 26, 2009 - link
"You no longer have to boot to DOS and secure erase your SSD before installing Windows, just quick format the partition before installing Windows 7."I always thought the "quick format" doesn't wipe the entire drive as clean as a full/long format. Does the above mean we only need to do a quick format to wipe a drive clean to install Windows 7?
Also, I noticed Superfetch, Prefetch, and Ready Boost are enabled in Windows 7 x64 with my X25-M G2 160GB drive installed.
Voo - Monday, October 26, 2009 - link
Logically when a TRIM command is sent to the drive you lose all the data that was saved there.So, there's no difference between full/quick format or anything else as long as it triggers the TRIM command..
Beno - Monday, October 26, 2009 - link
"SSDs are made up of millions of NAND flash cells. They can be written to in groups called pages (generally 4KB in size) but can only be erased in larger groups called blocks (generally 128 pages or 512KB). These stipulations are partially the source of many SSD performance issues."but isnt it possible to erase data in 4KB, the same way they are written?
krumme - Monday, October 26, 2009 - link
Where does speed matter most?Getting some sense into the g2 hype
Now you can have a job with a Dell and a hd with samsung controller without quitting the job
Thanx Anand for always improving your site and letting us learn
Thats whats move this boys-toy business forward
flynace - Monday, October 26, 2009 - link
If you set up a software RAID-0 array within Windows 7 will it support TRIM?Or is TRIM not possible on any form of RAID (OS, BIOS, HW, etc.)?
And does aliging the partitions make any difference on these drives?
Voo - Monday, October 26, 2009 - link
Iirc Anand stated in some of his articles that there's no TRIM support for raid, because the controller had to forward the commands.But imho it sounds not like a unconquerable problem, so it's probably only a question of time.
flynace - Tuesday, October 27, 2009 - link
Thanks. I assume Intel will update their drivers to support RAID TRIM eventually, but if you don't have an 'R' tybe Intel ICH, does Win7 support TRIM if you set up a software RAID array in the OS still using the MS AHCI drivers?Willardjuice - Monday, October 26, 2009 - link
Did Intel update their Windows 7 sata/raid drivers to also include TRIM support? Or do we have to use the drivers provided by Microsoft?DoveOfTheSouth - Monday, October 26, 2009 - link
Another great SSD article from Anand (thank you so much)!But I'm really disappointed that you have to use the Microsoft Win 7 disk driver to get TRIM. It has a huge problem.
Based on my experience, I would normally strongly recommend that the first thing you do when installing Win 7 is replace the Microsoft driver with the latest Intel Matrix Storage Manager.
The following link describes the problem (long periods of system non-responsiveness) you can get with the Microsoft driver and at least some PCs, and how to fix it:
http://blogs.zdnet.com/Bott/?p=1431&page=3&...">http://blogs.zdnet.com/Bott/?p=1431&page=3&...
I had exactly the same crippling problem on a Dell Dimension 9100 which baffled me until I replaced the Microsoft driver with IMSM - problem disappeared instantly.
Okay it was bit of an old PC but of the 3 I have been testing with Win 7 for the last couple of months, it was the only one that used the Microsoft driver instead of IMSM (it was an XP upgrade ie OS reinstall, the others upgrades from Vista). So I don't know how widespread the problem with the Microsoft driver is, it hit my only PC that was a candidate, but I'd be pretty wary.
Otherwise Win 7 has been great.
So it looks like you have a choice: IMSM with no TRIM or a buggy Microsoft driver with TRIM.
EasterEEL - Monday, October 26, 2009 - link
Once the Intel firmware is updated to v1.4 with trim support does Windows 7 start using trim? i.e I have already installed Windows 7 before dong the firmware update.Does ghosting an image back to the SSD have any impact on trim??
magreen - Monday, October 26, 2009 - link
That Kingston for $85 is looking tempting for sprucing up a circa-2004 laptop of mine that I use daily. But the laptop is ide. I know of others in the same boat.Have you heard of anyone producing an inexpensive quality ssd for ide? The Kingston seems like the perfect candidate since it's anyways not as fast as the regular Intels.
Zoomer - Monday, October 26, 2009 - link
Make sure that the drive interfaces match. If it's from 2004, it might still be PATA. I'd imagine this is SATA only.That said, it'll be interesting to see if soldering more flash chips on = upgrade in space. Shouldn't be too hard to find the flash chips (2 * Kingstons < Intel), but I'll need to find a good soldering jock.
magreen - Thursday, October 29, 2009 - link
Ummmm... maybe I wasn't making myself clear.I wasn't referring to upgrading the drive's capacity.
I was referring to Kingston producing a PATA drive that would work with my PATA laptop. I know PATA doesn't work with SATA. (I loosely referred to PATA as IDE, which people often do...)
thanx
IMFTbestFab - Monday, October 26, 2009 - link
You can't upgrade capacity like this. The drive has a maximum LBA associated with it, presumably set at the factory.Eeqmcsq - Monday, October 26, 2009 - link
Another great SSD article, Anand. Found a problem: Page 3, "Wipe When You Can’t TRIM", when you said "I went into a deep explanation of the relationship between free space and the performance of some SSDs here.", it sounds like there should be a link to a previously written article, but there is no link.ekbond - Monday, October 26, 2009 - link
Any thoughts on TRIM support in OS X?The download page for Intel's SSD utility specifies that it's OS independent, while the user guide says Windows is required.
Apart from formatting the drive, are there any other ways to restore performance (via TRIM or otherwise) currently available to OS X users?
Anand Lal Shimpi - Monday, October 26, 2009 - link
OS X doesn't have TRIM support yet unfortunately. I'm trying to get Apple to at least acknowledge that they will be supporting the feature but haven't had any luck yet. I haven't tried some of the erase options in Disk Utility to see if they do the equivalent of a secure erase. I'll try barking up that tree shortly.Take care,
Anand
chizow - Monday, October 26, 2009 - link
Newegg will just gouge the crap out of prices regardless. $100-130 will be more like $150-180. Remember when you said Intel would be dropping SSD prices with the G2 to $230 and $450 for the 80g and 160g respectively?http://www.newegg.com/Product/ProductList.aspx?Sub...">http://www.newegg.com/Product/ProductLi...scriptio...
They were sub $250 for about 12 hours total since Newegg got them in stock 2 months ago after the initial "recall".
strikeback03 - Tuesday, October 27, 2009 - link
I picked up my 80GB G2 a couple weeks ago from Newegg for $239 shipped, for the retail boxed one; the deal also included a Icy Dock 2.5" to 3.5" adapter and a ThermalTake Element S case for $30 with a $30 MIR.chizow - Wednesday, October 28, 2009 - link
Ya they had them for about 4 hours in an AM Shell Shocker for $240, which is their MSRP. They also had them for that price for about 8 hours at launch and since then its been gouged as high as $400 for a G2. All the other retailers that are selling for around the $240 MSRP are drop-shipping directly from Intel so they typically don't have stock in-hand and are estimating anywhere from 2-8 weeks for delivery....crimson117 - Monday, October 26, 2009 - link
Amazon has decent prices on these, but shipping time ranges from 2 weeks to 8 weeks.80GB G2 for $244: http://www.amazon.com/Intel-Mainstream-Solid-Retai...">http://www.amazon.com/Intel-Mainstream-...-Retail-...
160GB G2 for $467: http://www.amazon.com/Intel-160GB-Mainstream-Retai...">http://www.amazon.com/Intel-160GB-Mains...-Retail-...
Xentropy - Sunday, November 1, 2009 - link
5 days later now, and Amazon's up to $599 for the 160GB, though they are in stock instead of shipping in several weeks.Exar3342 - Tuesday, October 27, 2009 - link
I bought a G2 80GB yesturday via Amazon from a 3rd party company; was $244.00 (OEM) version. Good times.strikeback03 - Monday, October 26, 2009 - link
With your recommendation to keep at least 20% of a drive as free space, should it be within a partition, or just leave some unpartitioned space on the drive?Anand Lal Shimpi - Monday, October 26, 2009 - link
Either way works, just don't put data on it :)Take care,
Anand
Doormat - Monday, October 26, 2009 - link
Anand: great review.The 40GB 5-channel version has read speeds of 170MB/s. Does this mean that if the 10-channel G2 weren't inhibited by SATA 3Gb/s, it would probably get 340MB/s? Or is there diminishing returns on adding more channels.
Also, any price cuts on the horizon or just stagnating until drives start shipping with SATA 6Gb/s support.
Anand Lal Shimpi - Monday, October 26, 2009 - link
It's possible. All of the high end SSDs are bound by the interface for sequential read speed. I'm not sure how fast the controller could go if it weren't bound by the interface, but over 300MB/s sounds quite likely.Flash prices haven't really gone down lately from what I'm hearing. I don't know of any price cuts on the horizon unfortunately. Next year.
Take care,
Anand
semo - Monday, October 26, 2009 - link
i don't think you will see any price cuts since intel can't make enough G2s.I tend to agree with others regarding SATA 3. If intel is not rushing to enable SATA 3 on their ICH anytime soon, why would they be coming out with SATA 3 devices?
ekerazha - Monday, October 26, 2009 - link
Anand,it's strange to see your
"Is Intel still my overall recommendation? Of course. The random write performance is simply too good to give up and it's only in very specific cases that the 80MB/s sequential write speed hurts you."
of the last review, is now a
"The write speed improvement that the Intel firmware brings to 160GB drives is nice but ultimately highlights a bigger issue: Intel's write speed is unacceptable in today's market."
Anand Lal Shimpi - Monday, October 26, 2009 - link
Personally my pick is still the X25-M, it's had a better track record than any of the other controller makers. The problem for Intel is that Indilinx performance and feature set have both gotten a lot better.You now get TRIM, reliable performance, competitive pricing and generally everything but a super dynamic controller that's constantly trying to maximize performance. Depending on the usage scenario though, the latter may not matter as much (see my Photoshop results from the last review or the heavy trace results from this review).
My position hasn't really changed, just highlighting the importance of the issue so that hopefully Intel's value SSDs next year aren't just rebadged G2s :)
Take care,
Anand
7Enigma - Monday, October 26, 2009 - link
I think his previous review is consistent with this one. His ultimate recommendation between Intel and Indlinx drives has remained unchanged. That is, they are close enough that the size differences should be the decider for a potential buyer.As it stands, in most normal usage scenarios the statement from his last article still applies. There are very few times where you are going to be constantly tapping the 80MB/s barrier. More often you will be limited by the other SSD's lower random write speeds.
What his own benchmark can show is that in uncommon circumstances you can sway the bar back towards parity between the brands. I would argue, however, that the "heavy" usage scenario is not a practical one for an SSD. Just read his description again and look for all of the places the average user would NOT be putting that data onto an SSD:
[quote]
In this test we have Microsoft Security Essentials running in the background with real time virus scanning enabled. We also perform a quick scan in the middle of the test. Firefox, Outlook, Excel, Word and Powerpoint are all used the same as they were in the light test. We add Photoshop CS4 to the mix, opening a bunch of 12MP images, editing them, then saving them as highly compressed JPGs for web publishing. Windows 7’s picture viewer is used to view a bunch of pictures on the hard drive. We use 7-zip to create and extract .7z archives. Downloading is also prominently featured in our heavy test; we download large files from the Internet during portions of the benchmark, as well as use uTorrent to grab a couple of torrents. Some of the applications in use are installed during the benchmark, Windows updates are also installed. Towards the end of the test we launch World of Warcraft, play for a few minutes, then delete the folder. This test also takes into account all of the disk accesses that happen while the OS is booting.
The benchmark is 22 minutes long and it consists of 128,895 read operations and 72,411 write operations.
[quote]
How many people are going to be opening photoshop pictures FROM the SSD and saving them TO the SSD? Same goes with Windows Picture Viewer. Most people will have their pics on a standard large mechanical HD which makes this a moot point. Downloading large files from the net? Not a chance that would be going to my SSD! Bottom line is this is a great HELL test, but nowhere close to real-world usage.
The light-usage probably fits 90% of people 90% of the time (completely made up statistic).
With all of that said, I wish Intel's 80gig drive had closer sequencial write performance as the Indilinx drives, as it would make the decision a complete no-brainer. Looking at all the data, and planning on buying a ~$250 drive this holiday season has me very likely picking up an 80gig Intel as for me the price is about as steep as I'm willing to go for essentially a boot drive, with performance that for the majority of the time is at the top of the pack. It would be WAYYYY more difficult to decide if both Intel/Indilinx had the same drive sizes but fortunately we don't have to worry about that (at least for now....)
tynopik - Monday, October 26, 2009 - link
on page 7, all the graphs have two lines labelled 'Intel X25-M G2 80GB (MLC) TRIM'7Enigma - Monday, October 26, 2009 - link
Anand,You have listed the 80gig Intel MLC drive twice in each chart. I'm assuming one of them shouldn't say (TRIM)?
Other than that, GREAT article! Those Kingston drives look to be the PERFECT inexpensive replacement for netbooks/cheaper laptops as the performance is vastly superior to cheap mechanical HD's.
Anand Lal Shimpi - Monday, October 26, 2009 - link
That second 80GB MLC drive was actually a second set of results that didn't get pulled out. I've trimmed em out :)Take care,
Anand
fic2 - Monday, October 26, 2009 - link
"The G1 is more resillient (SIC) than the G2 when it comes to performance degradation over time since it doesn't have TRIM."Is it just me or does this sentence make no sense? I thought the whole purpose of TRIM was to make the SSD more resilient to performance degradation. Seems like it should be:
"The G1 is less resilient..."
klil - Monday, October 26, 2009 - link
The G1 is LESS susceptible to degradation than the G2, Intel compensates this by enabling TRIM on the G2.Anand Lal Shimpi - Monday, October 26, 2009 - link
Correct, my apologies for the confusing wording :)Take care,
Anand
edzieba - Monday, October 26, 2009 - link
"None of the other Indilinx manufacturers have Windows 7 TRIM support yet."Crucial has the TRIM firmware (1819) available for download for their m225 series SSDs.
TheWickerMan - Monday, October 26, 2009 - link
I was just going to post that, but you beat me to it. Here's the link for anyone else interested: http://www.crucial.com/support/firmware.aspx">http://www.crucial.com/support/firmware.aspxAnand Lal Shimpi - Monday, October 26, 2009 - link
Thanks guys :)Take care,
Anand
jimhsu - Monday, October 26, 2009 - link
Since I've been wondering about the free space issue, will there be a test where a drive is benchmarked vs. the amount of free space available? My guess is that the graph will be an exponential decay with the "noticeable" regime at 20% free space. Anyone cares to do this (say, at 90%, 80% ... close to zero free space)?Crittias - Monday, October 26, 2009 - link
So, without TRIM, is there anything else G1 owners can do? Can I manually wipe the drive and reinstall my OS every few months? If so, should I?mataichi - Monday, October 26, 2009 - link
I have the G1 and in Anand's previous SSD Anthology article he tells how to get back to 100% performance using HDDERASE.http://www.anandtech.com/storage/showdoc.aspx?i=35...">http://www.anandtech.com/storage/showdoc.aspx?i=35...
This is total crap that Intel is not supporting the G1.
DanH - Monday, October 26, 2009 - link
Buy the G2, clone the G1 to it and sell the gen 1. Really. It is not a big loss of money to do it.lorax1284 - Tuesday, December 15, 2009 - link
Basically, you suggest we pass our problem on to someone else who is less informed. It's called "Ethics". Look into it.Just because Intel has none doesn't mean I'm going to abandon mine.
klil - Monday, October 26, 2009 - link
The only option you, a G1 owner, have to make your drive's performance back to ~100% is to do a full format of the drive. I'm pretty sure it would have to be done very frequently, each time you write 80/160GB of data to the drive... nothing else you can do, sorry =(Intel's wipe ultility (SSD Toolbox) only works with the G2 (for use with Windows XP/Vista) and TRIM only works with the G2 as well (for use with Windows 7-to replace the Toolbox)... =(
Zoomer - Monday, October 26, 2009 - link
One can always try the Indillinx tool, after making a full backup, of course. Heck, 80GB would be less than two blu-rays.klil - Monday, October 26, 2009 - link
Another great article by the one and only Anand Shimpi, king of SSDsSo, basically Intel made a very stupid move by only allowing SSD Toolbox AND TRIM on G2 drives. I would like to know if its due to architectural changes (perhaps?), maybe you know it but can't reveal it to anyone... if I weer you i would edit out the part when you say the G3 are coming next year... or else Intel will kill you, they just lost a custumer for the G2 (me)
But would it make sense for them to release the SATA 6 G3 drives next year if their on-board ICH10 and MCH P55 does not support it? Maybe the P57 and X68 will? Who knows?
I'll still be waiting for those Gulftowns... =D
Bakkone - Monday, October 26, 2009 - link
Don't think its that much of a secret that we will se a new generation next year. We need a version that uses Sata3 (6Gb).slickdoody - Monday, October 26, 2009 - link
Did Intel remove the SSD toolbox software?The link in the article doesn't work and I can't find the software on the site by searching.
Griswold - Monday, October 26, 2009 - link
Here it is again:http://downloadcenter.intel.com/Detail_Desc.aspx?a...">http://downloadcenter.intel.com/Detail_Desc.aspx?a...
slickdoody - Monday, October 26, 2009 - link
thanks!Sunburn74 - Monday, October 26, 2009 - link
Its absolutely ridiculous how intel is screwing those customers who bought their x25m g1s by specifically going out of their way to prevent trim support. At least give a manual wiper tool, you jerks. Geez!lorax1284 - Tuesday, December 15, 2009 - link
I know that Toms Hardware has commented on the fact that Intel is not providing a firmware update for the X25-m first generation drives... but I think TH is in a position to rake Intel over the coals for this terrible decision! Toms Hardware caters to computer enthusiasts, exactly the type of customers who buy first generation hardware like the X25-m G1... so for Intel to NOT issue a TRIM firmware update for X25-m G1 drives should ahve Toms Hardware users up in arms, and certainly questioning whether Intel's support for the G2 drives will stand the test of time!If you think Intel is wrong to abandon the G1 drives by not providing a TRIM firmware update, please consider signing this petition!
http://www.ipetitions.com/petition/intel-x25-m-g1-...">http://www.ipetitions.com/petition/intel-x25-m-g1-...
cbutters - Monday, November 23, 2009 - link
Agreed 100%, I purchased an x-25m G1 at a premium thinking that intel would take care of me once TRIM support was available. Since I heard that intel wouldn't be giving any G1's trim for no reason other than to screw over the consumer into buying an updated drive, I have only purchased Indilinx based SSDs (over 4 now) and I recommend that everyone else do the same!They also promised that the 80gb G2 SSDs would be sold at a price point of $229, but due to the demand they are gouging everyone selling them at price points of 299 and above. (can't blame them, it's economics, but don't PROMISE it at $229 and not fulfill it.)
People buy stuff where the performance is, but I won't buy an intel SSD again unless they release something that performs significantly better than the competitor, but the point is that you should buy indilinx based drives because they are just as good and you won't be supporting Intel's poor decision with regard to the G1 customers.
winterlord - Wednesday, November 18, 2009 - link
great SSD articles you guys have been putting out latly. but it would be great to see a corsair SSD in these benchies. alot of people talk about them and they seem blazing fast even faster then the intels from what iv read around on google but id like to see one of my populer tech websiteas to confirm this. i like this article here but any chance you can throw a corsiar in the lineup :)they have trim right?
chrnochime - Monday, October 26, 2009 - link
You pay the price for being early adapter. I mean even now with the G2 a colleague of mine just bought back in August, the thing still had to be RMA'd.I'd never use something like this for work related use unless it's been tested by many guinea... err I mean other consumers before it's proven to be cheap and reliable.
Griswold - Thursday, November 19, 2009 - link
We got another ADAPTER here! What do you adapt with, if I may ask?WillBach - Monday, October 26, 2009 - link
If you don't have the software to image and restore your drive, get it now! It's an order of magnitude less expensive* than the SSD, and it's the fastest, cleanest, and most reliable way to recover from a failed hard drive.*If you run OS X, you can use Time Machine or Disk Utility. If you run Linux or BSD, you can use dd from the terminal.
DanH - Monday, October 26, 2009 - link
Do what I did. Buy a Gen 2, clone your drive, and sell the Gen 1 on ebay while it's still worth as much as it is. You will easily get close to $200 for the 80 gig version. I lost $30 upgrading, which was very worth it for TRIM support and the newest hardware.masouth - Wednesday, November 18, 2009 - link
I must be misunderstanding you so please let me make sure I have this right...Someone bought your USED G1 for $30 less than you paid for your NEW G2!?!
A sucker born every minute.
mwaschkowski - Monday, October 26, 2009 - link
what cloning software did you use?I guess I couldn't switch to another SSD at the same time due to needing to clone the drive and using the same drivers between both SSDs?
DanH - Monday, October 26, 2009 - link
I used Carbon Copy Cloner on the mac, but Acronis TrueImage would work fine on PC.chizow - Monday, October 26, 2009 - link
Acronis True Image is probably the easiest for Windows, especially if you're cloning to identical sized drives. With the trial you can do this for free, I believe Seagate and WD have simplified free versions if you have one of their drives installed.Don't mess with any of the backup/restore options, go to the Utilities and choose Clone drive, then select destination and source drives. Takes about 10-20 minutes depending how much data you have and is effortless.
mwaschkowski - Tuesday, October 27, 2009 - link
OK, Thanks!!pcfxer - Tuesday, October 27, 2009 - link
sudo dd if=/dev/da0 of=/dev/da1Boot up into Knoppix and it is legitimately free.
Concillian - Wednesday, October 28, 2009 - link
No kidding. I clone my drive every once in a while on my fileserver with dd. Why it's made so difficult in every other OS is beyond me. I really shouldn't need to buy software in order to make a copy. It's a pretty basic function, really.Griswold - Thursday, November 19, 2009 - link
Just that the aforementioned software isnt primarily a cloning software but a backup and image software. Cloning a disk is just a minor function of it. And no, cloning a disk isnt a worthwhile backup strategy for most people.That said, I backup my system and data on a daily basis, but I only clone a drive once maybe twice a year. I wouldnt mind if it was built into windows, but its far from a necessity.
Taft12 - Monday, October 26, 2009 - link
This is nothing new, it has been well known for many years that there is a price to pay by being an early adopter.The0ne - Monday, October 26, 2009 - link
No kidding. Early adopters really shouldn't complain, that's what you get for being the "elite." LMAO.As I've said before, just wait til early next to summer and the tech will be more mature. Yea, go ahead and make your argument that you can wait til next year but by then they'll be new products. That's cool...but I think ppl saying this have no clue about what's really happening around the technology.
Regardless, sounds nice to be saying it so woohooo be the first!
klil - Monday, October 26, 2009 - link
I dont think that you should lose the wiper and trim support for being an early adapter, it does not make senseThe only early adapter lesson i know of was that one from a intel processor that had a bug... pentiums from 1994: http://en.wikipedia.org/wiki/Pentium_FDIV_bug">http://en.wikipedia.org/wiki/Pentium_FDIV_bug
If they were not able to put it in due to hardware problems its one thing... but those drives cost 665 dollars in the beggining. I bet that if you had one you would not be happy :|
pcfxer - Tuesday, October 27, 2009 - link
ADOPTER, A-D-O-P-T-E-R.I can't even fathom how you could read what someone posts, "early adopter this....early adopter thaaat" and you go and type ADAPTER.
Adapter -> Adaptation between two or more objects.
Did you not have coffee yet man?
chizow - Monday, October 26, 2009 - link
Agreed, Intel will certainly meet more resistance in the future for their emergent technologies and undoubtedly lost the support of many of their most dedicated supporters with this decision.If there were some technical reason or incompatibility behind the lack of TRIM support on G1 that'd be one thing, but from what has been published by AT and others, its nothing more than indifference on Intel's part preventing support.
Blur - Tuesday, October 27, 2009 - link
Eloquently stated, i bought two of the G1 drives. I won't be suckered into supporting any of intels new tech early again and I'll be taking a much closer look at their competitors products when purchasing more established tech.Zingam - Monday, October 26, 2009 - link
Is there a change that a newer better technology will replace Flash so that we can finally have real SSDs at affordable prices?Zingam - Wednesday, October 28, 2009 - link
Sorry about the spelling/grammar errors above!therealnickdanger - Monday, October 26, 2009 - link
Intel was working on some sort of phase-change technology, not sure where it is now. Probably still too expensive for mainstream:http://www.dailytech.com/searchresults.aspx?keywor...">http://www.dailytech.com/searchresults.aspx?keywor...
MS - Monday, October 26, 2009 - link
Nanochip went belly-up and the IP was auctioned off. Ovonyx turned into Numonyx and PCM is still an emerging alternative to other technologies. The biggest issues relate to the limitatios in process technology, i.e, at about 20 nm geometry, the required write currents are expected to actually melt the metal interconnects. There are other technologies such as resistive memory as developed by 4-d-s and organic memories such as the ones from Zettacore that may become very attractive.Souka - Monday, October 26, 2009 - link
phase-change technology?Sounds like vapor-ware to me! :)
Donkey2008 - Monday, October 26, 2009 - link
(Rimshot)erikejw - Thursday, October 29, 2009 - link
When can we expect the first benches and reviews with the JMF612 controller?