or Connect
AVS › AVS Forum › Video Components › Home Theater Computers › Interesting SSD Reliability Study
New Posts  All Forums:Forum Nav:

Interesting SSD Reliability Study - Page 6

post #151 of 199
It doesn't even have to be HTPC. Even for typical desktop usage you'd be hard pressed to notice a difference between 2 decents SSDs. Mind you, while both the Intel 330 and 520 use Synchronous ONFi NAND, Intel uses the same part numbers for 5,000 P/E and 3,000 P/E cycle NAND. We're still not sure if the Intel 330 uses 3,000 P/E cycle NAND. Meaningless for most users but it could be a pretty significant differentiation for enterprise workloads.
post #152 of 199
Quote:
Originally Posted by ilovejedd View Post

I use MediaBrowser which caches metadata and images in C:\ProgramData. This is the thing that was most noticeable to me. On a PC with a platter drive, there's usually a delay when I'm browsing through thumbnails. With the SSD, loading metadata and poster art is practically instantaneous. It helps make the HTPC feel more like an appliance and less a computer.

+1. here.

Same.

Well explained ilovejedd. Nice work.
Edited by Mfusick - 7/23/12 at 10:44am
post #153 of 199
Quote:
Originally Posted by ilovejedd View Post

Not even then. You'd need to have a fast source to even notice the speed difference. Basically, either transferring from SSD to SSD or RAM disk to SSD. Copying from SSD to HDD or HDD to SSD, you'll be limited by the HDD speed. Blu-ray to SSD and you'll be bottlenecked by your Blu-ray drive. ATSC maxes out at a measly 19Mbps (2.375 MB/s) and QAM maxes out at 38Mbps (4.75 MB/s) so even if you're recording several streams at the same time (course you need to have a big enough SSD to handle all those simultaneous recordings), the SSD wouldn't skip a beat. Really, you'd be hard pressed to notice a difference between an Intel 520 240GB running on SATA III 6Gb/s vs an Intel 330 240GB connected to a SATA II 3Gb/s port.
Tom's Hardware sums it best:

+1. Here.

In my real world experience with over 15 different SSD's you typically can not tell any difference between most SSD's real world- even if the benchmarks are slightly different.

It would take a statistically significant difference in benchmarks - or- a very large difference in SSD performance to tell a difference for an ordinary user in a HTPC application. It's not relevant from one brand to another of the same size.

I can only really tell the difference when going from a low end small drive like a 60GB Agility2 to a MAX IOPS Toggle NAND drive of a much larger capacity. In this scenario your comparing a $50 SSD to a $200 SSD and it's very unlikely a buyer looking at one would also be considering the other.

You do get what you pay for - but even a $50 SSD is light years ahead of a HDD. There is no reason to use a HDD for OS anymore.

On my desktop I was surprised I could tell the difference in speed with a 240GB RAID 0 array but it was pretty obvious it was faster. That was the largest difference I have experienced.

The difference from a 120GB SATAII Vertex2, a 120GB Vertex3 SATA3 and even a 120GB Toggle Nand MAX IOPS SATA3 is pretty much a non issue. As is the difference between most brands of the same size.

You would have to be experienced with SSD performance to notice anything. No one buying an SSD for the first time is going to know any better any small difference in speed.

Price would be a major factor for me. Second would be performance. Third would be reliability. (They are all reliable IMO)
post #154 of 199
Quote:
Originally Posted by ilovejedd View Post

It doesn't even have to be HTPC. Even for typical desktop usage you'd be hard pressed to notice a difference between 2 decents SSDs. Mind you, while both the Intel 330 and 520 use Synchronous ONFi NAND, Intel uses the same part numbers for 5,000 P/E and 3,000 P/E cycle NAND. We're still not sure if the Intel 330 uses 3,000 P/E cycle NAND. Meaningless for most users but it could be a pretty significant differentiation for enterprise workloads.

Yeah, I've read several conflicting articles. One below from Tom's Hardware (while from March) shows the hardware (including NAND) to be identical as what's used in the 520. But they don't specifically mention P/E cycles. Other articles specifically mention the 330 using 3k P/E cycle NAND, with the 520 using 5k P/E cycle NAND. This is more confusing with what you said above: that both 5k and 3k NAND use the same part numbers. Perhaps this explains the "bin" use.

http://www.tomshardware.com/reviews/ssd-330-review-benchmark,3190-2.html

Wish we had official confirmation on which the 330 drives are using. Only other question was whether the 330 would also suffer from the AES-256 bit encryption issue suffered by the 520. Although new or future 520s are supposed to have this issue fixed (an issue with the controller hardware).

In either case, like you say it is likely to be meaningless for most home users, but could both be more of a concern for enterprise users. Even if it were using the 3k NAND, having a drive with lots of unused extra space allows the controller and firmware more space to utilize and lowering the number of P/E cycles needed. From my personal situation, considering now my C: partition only takes up 85GB, going to a 240GB SSD and only using this drive for OS and programs, I won't come close to needing the full capacity. Thus having all that extra space available should only mean you'll likely never use up all 3,000 P/E cycles as you'll probably replace the drive before then (either due to a failure, other problem or just to upgrade).
post #155 of 199
Quote:
Originally Posted by bobby2478 View Post

So in the real world, especially so for HTPC use where your library is on a HDD and you use the SSD primarily for OS and programs, a vast majority of the time you would never experience a situation where the speed of your SSD was actually slowing you down. Only in very rare situations would you ever potentially experience this. And in one of these rare situations you did experience it, the difference in speed compared to a "faster" drive is so small you'd need to measure it in fractions of a second.

Again, depends on your workload. Typical client usage, that's true. Not so much for some power users.

For example, copying a 20GB virtual disk file (with NTFS compression) would take the following based on AnandTech's AS-SSD sequential write benchmarks (assuming you're limited by drive write speed):

Intel 330 180GB: 95.4s
Intel 520 240GB: 72.0s
Crucial m4 256GB: 82.9s
Samsung 830 256GB: 51.8s
OCZ Agility 3 240GB: 85.8s
OCZ Vertex 3 240GB: 72.0s

Enterprise workloads with high IOPS requirements (e.g. databases serving millions of users), yeah, the difference between the 2 drives might be noticeable as we really have no idea what Intel did to slow down the 330 series (firmware de-tuning? underclocking?)

Quote:
Originally Posted by bobby2478 View Post

Wish we had official confirmation on which the 330 drives are using. Only other question was whether the 330 would also suffer from the AES-256 bit encryption issue suffered by the 520. Although new or future 520s are supposed to have this issue fixed (an issue with the controller hardware).

In the first place, the Intel 330 series are not marketed as AES-256 capable so that's a moot point. That's one of the things Intel uses to differentiate the 330 and 520 series. The Intel 330 is mostly targeted towards consumers while the Intel 520 is also targeted towards corporate and enterprise where more stringent encryption matters. I don't think Intel was the one who discovered the issue as it appears to be embedded pretty deep inside the SandForce controller and is not something you can change via firmware. I think it was LSI (bought SandForce) who discovered the issue and issued a bulletin to its clients. All SF-2200 drives are affected so manufacturers (well, Intel and Kingston, I believe, not sure about others) who have AES-256 bit encryption marketed as part of their SSD's feature set have issued a recall. Now that SandForce has been acquired by LSI, I'm inclined to trust their controllers a bit more. I don't think LSI's going to want to damage its reputation by releasing unreliable controllers.

There is no fix for the lack of AES-256 bit encryption (it's going to have to come from LSI SandForce). The product literature for the 520 series was changed to specify that it only supports AES-128 bit.
post #156 of 199
Quote:
Originally Posted by ilovejedd View Post

Again, depends on your workload. Typical client usage, that's true. Not so much for some power users.
For example, copying a 20GB virtual disk file (with NTFS compression) would take the following based on AnandTech's AS-SSD sequential write benchmarks (assuming you're limited by drive write speed):
Intel 330 180GB: 95.4s
Intel 520 240GB: 72.0s
Crucial m4 256GB: 82.9s
Samsung 830 256GB: 51.8s
OCZ Agility 3 240GB: 85.8s
OCZ Vertex 3 240GB: 72.0s
Enterprise workloads with high IOPS requirements (e.g. databases serving millions of users), yeah, the difference between the 2 drives might be noticeable as we really have no idea what Intel did to slow down the 330 series (firmware de-tuning? underclocking?)

As I understand it, the 330 realizes slower speeds due to the number of channels it uses. It uses less channels then the 520, hence it isn't as fast.

Sure, for power users or enerprise users who are processing a 20GB file with NTFS compression, this may be more noticeable. I was referring to most "average" users who are making a HTPC to serve as their home PC as well as a media server. People who use it to play media (movies and music), as well as basic every day computer use (surfing the internet, MS Office, photo editing, etc). So for those "average" users, the difference is so small it doesn't compute. It would be helpful when review sites publish specs and benchmarks to put those into some perspective. As one site put it, the fastest SSDs are about 87% faster then a HDD for most tasks, while the SSDs at the very bottom end of the performance spectrum still outperform HDDs by 84%. That difference of 3% won't be meaningful to most "average" users.
Quote:
Originally Posted by ilovejedd View Post

In the first place, the Intel 330 series are not marketed as AES-256 capable so that's a moot point. That's one of the things Intel uses to differentiate the 330 and 520 series. The Intel 330 is mostly targeted towards consumers while the Intel 520 is also targeted towards corporate and enterprise where more stringent encryption matters. I don't think Intel was the one who discovered the issue as it appears to be embedded pretty deep inside the SandForce controller and is not something you can change via firmware. I think it was LSI (bought SandForce) who discovered the issue and issued a bulletin to its clients. All SF-2200 drives are affected so manufacturers (well, Intel and Kingston, I believe, not sure about others) who have AES-256 bit encryption marketed as part of their SSD's feature set have issued a recall. Now that SandForce has been acquired by LSI, I'm inclined to trust their controllers a bit more. I don't think LSI's going to want to damage its reputation by releasing unreliable controllers.
There is no fix for the lack of AES-256 bit encryption (it's going to have to come from LSI SandForce). The product literature for the 520 series was changed to specify that it only supports AES-128 bit.

That would make sense as another delineating factor. I was aware it was a problem with the controller itself, meaning the only way to "fix" the issue would be to build a drive with a different controller. There was an article on Anandtech that said after getting a refund, any Cherryville owners who wanted AES-256 encryption would need to "wait a few months for a new spin of the controller", indicating that Intel might start producing a different revision of the 520 which may support AES-256. Regardless, for "average" home users this is meaningless as personally I don't have a need to encrypt my hard drive.
post #157 of 199
Quote:
Originally Posted by bobby2478 View Post

As I understand it, the 330 realizes slower speeds due to the number of channels it uses. It uses less channels then the 520, hence it isn't as fast.

I think that's speculation from some websites. Others have speculated that they reduced the clock speed or done other things in firmware. I don't think anyone really knows and Intel's not talking.
post #158 of 199
Quote:
Originally Posted by Zon2020 View Post

I think that's speculation from some websites. Others have speculated that they reduced the clock speed or done other things in firmware. I don't think anyone really knows and Intel's not talking.

If it was firmware wouldn't some have figured out a hack by now to get the 520 Firmware into a 330 drive?
post #159 of 199
Quote:
Originally Posted by bobby2478 View Post

As I understand it, the 330 realizes slower speeds due to the number of channels it uses. It uses less channels then the 520, hence it isn't as fast.

Link? Because from the PCB pics I've seen so far, NAND and channel arrangement appears to be the same for both the Intel 330 and Intel 520 at same capacities. Both appear to use the same 16GB NAND (P/N: 29F16B08CCME2). Right now, the reason for the Intel 330's lower speed is all speculation.
post #160 of 199
Quote:
Originally Posted by ilovejedd View Post

Again, depends on your workload. Typical client usage, that's true. Not so much for some power users.
For example, copying a 20GB virtual disk file (with NTFS compression) would take the following based on AnandTech's AS-SSD sequential write benchmarks (assuming you're limited by drive write speed):
Intel 330 180GB: 95.4s
Intel 520 240GB: 72.0s
Crucial m4 256GB: 82.9s
Samsung 830 256GB: 51.8s
OCZ Agility 3 240GB: 85.8s
OCZ Vertex 3 240GB: 72.0s
Enterprise workloads with high IOPS requirements (e.g. databases serving millions of users), yeah, the difference between the 2 drives might be noticeable as we really have no idea what Intel did to slow down the 330 series (firmware de-tuning? underclocking?)
In the first place, the Intel 330 series are not marketed as AES-256 capable so that's a moot point. That's one of the things Intel uses to differentiate the 330 and 520 series. The Intel 330 is mostly targeted towards consumers while the Intel 520 is also targeted towards corporate and enterprise where more stringent encryption matters. I don't think Intel was the one who discovered the issue as it appears to be embedded pretty deep inside the SandForce controller and is not something you can change via firmware. I think it was LSI (bought SandForce) who discovered the issue and issued a bulletin to its clients. All SF-2200 drives are affected so manufacturers (well, Intel and Kingston, I believe, not sure about others) who have AES-256 bit encryption marketed as part of their SSD's feature set have issued a recall. Now that SandForce has been acquired by LSI, I'm inclined to trust their controllers a bit more. I don't think LSI's going to want to damage its reputation by releasing unreliable controllers.
There is no fix for the lack of AES-256 bit encryption (it's going to have to come from LSI SandForce). The product literature for the 520 series was changed to specify that it only supports AES-128 bit.

95 seconds vs 51 seconds is noticeable for sure.


BTW- 256GB SAMSUNG SSD on sale NewEgg for $199 today eek.gif

http://www.newegg.com/Product/Product.aspx?Item=N82E16820147164&Tpk=20-147-164&nm_mc=EMC-GD072312&cm_mmc=EMC-GD072312-_-index-_-Item-_-20-147-164
post #161 of 199
Quote:
Originally Posted by JackDiesel14 View Post

If it was firmware wouldn't some have figured out a hack by now to get the 520 Firmware into a 330 drive?

It's definitely firmware. Those who have opened the drives have all reported the hardware is identical in the two drives. Same controller. Same NAND.
post #162 of 199
Quote:
Originally Posted by ilovejedd View Post

Right now, the reason for the Intel 330's lower speed is all speculation.

+1

Intel knows, but it's not talking.
post #163 of 199
Quote:
Originally Posted by Mfusick View Post

95 seconds vs 51 seconds is noticeable for sure.
BTW- 256GB SAMSUNG SSD on sale NewEgg for $199 today eek.gif
http://www.newegg.com/Product/Product.aspx?Item=N82E16820147164&Tpk=20-147-164&nm_mc=EMC-GD072312&cm_mmc=EMC-GD072312-_-index-_-Item-_-20-147-164

Sure, that specific test would be noticeable. However those results are based on copying a 20GB virtual disk file (with NTFS compression), which isn't a scenario that you're "average" user is ever going to experience. Now if the test was booting the OS and there was a 41 second difference that would be noticeable to your "average" user.
post #164 of 199
Quote:
Originally Posted by bobby2478 View Post

Sure, that specific test would be noticeable. However those results are based on copying a 20GB virtual disk file (with NTFS compression), which isn't a scenario that you're "average" user is ever going to experience. Now if the test was booting the OS and there was a 41 second difference that would be noticeable to your "average" user.

I think most folks in this thread are already aware that for normal usage, there's practically no difference with different SSD models. I believe Zon2020, Mfusick and myself have already pointed that out in earlier posts. Just saying that there is a tangible difference depending on specific workloads since one of your earlier replies imply otherwise. It's just not something the average user is likely to encounter.
post #165 of 199
Quote:
Originally Posted by ilovejedd View Post

Link? Because from the PCB pics I've seen so far, NAND and channel arrangement appears to be the same for both the Intel 330 and Intel 520 at same capacities. Both appear to use the same 16GB NAND (P/N: 29F16B08CCME2). Right now, the reason for the Intel 330's lower speed is all speculation.

Here is one: http://www.storagereview.com/intel_ssd_330_review
Quote:
Originally Posted by Quote from article 
Pros
•Highly underrated 4K random read and write speeds
•Very competitive performance to price value
•Same NAND as found in the Intel SSD 520, just fewer channels

Here is another: http://www.hardwareheaven.com/reviews/1528/pg2/intel-330-series-ssd-120gb-review-the-drive.html
Quote:
Originally Posted by Quote from article 
Also worth noting is that these are synchronous NAND modules, not asyncronous as some other manufacturers use on their lower specification drives. We get the same spec parts as the 520 series, just on fewer channels.

I agree, too many conflicting details in various reviews. Nothing more then speculation as to the reason why the 330 is slower then the 520, as well as whether any hardware differences drive the lower cost.
post #166 of 199
Quote:
Originally Posted by bobby2478 View Post

Here is one: http://www.storagereview.com/intel_ssd_330_review
Quote:
Pros
•Highly underrated 4K random read and write speeds
•Very competitive performance to price value
•Same NAND as found in the Intel SSD 520, just fewer channels

Here is another: http://www.hardwareheaven.com/reviews/1528/pg2/intel-330-series-ssd-120gb-review-the-drive.html
Quote:
Also worth noting is that these are synchronous NAND modules, not asyncronous as some other manufacturers use on their lower specification drives. We get the same spec parts as the 520 series, just on fewer channels.
I agree, too many conflicting details in various reviews. Nothing more then speculation as to the reason why the 330 is slower then the 520, as well as whether any hardware differences drive the lower cost.

Both review sites above were comparing Intel 330 120GB and Intel 520 240GB. Of course the 120GB model would have fewer populated channels. If you check out the following reviews, you'll see that at the same capacity, the Intel 330 and Intel 520 both use the same NAND and channel configuration:

http://www.rwlabs.com/article.php?cat=&id=617&pagenumber=4
http://www.rwlabs.com/article.php?cat=&id=625&pagenumber=4
post #167 of 199
Thread Starter 
Quote:
Originally Posted by Mfusick View Post

+1. Here.
In my real world experience with over 15 different SSD's you typically can not tell any difference between most SSD's real world- even if the benchmarks are slightly different.
It would take a statistically significant difference in benchmarks - or- a very large difference in SSD performance to tell a difference for an ordinary user in a HTPC application. It's not relevant from one brand to another of the same size.
Quote:
Originally Posted by Mfusick View Post

No one buying an SSD for the first time is going to know any better any small difference in speed. Price would be a major factor for me. Second would be performance. Third would be reliability. (They are all reliable IMO)

Hmmm.

So performance is the same according to you for HTPC (I agree completely).

And in your opinion reliability is the same (you state that "they are all reliable")

So we should just choose the cheapest drive? And they are all exactly the same?

I respect your opinion MFusick but I just don't agree on this one. There are differences that set these drives apart. Either you are confused or you are starting to see the arguments/points that many others are making in this thread.
post #168 of 199
Quote:
Originally Posted by ilovejedd View Post

I think most folks in this thread are already aware that for normal usage, there's practically no difference with different SSD models. I believe Zon2020, Mfusick and myself have already pointed that out in earlier posts. Just saying that there is a tangible difference depending on specific workloads since one of your earlier replies imply otherwise. It's just not something the average user is likely to encounter.

My bad, I realize my earlier post was from my vantage point as an "average" user, and not a power user who may very well notice a difference in performance. While I've heard several times that for normal usage there would practically be no difference, it was hard for me to quantify exactly in which situation I would encounter a difference in performance and what this difference would mean. The last thing I want was to get the drive and think "gosh, if I had just gotten the faster drive it would have been money better spent."

The test result you posted earlier was very helpful for me to gain this understanding, thank you for that. I know now that I will never encounter a situation like the one used in the test, and as such that variance isn't something I would ever encounter. That definitely DOESN'T mean that someone who is a power user won't run across this situation and experience that lag in performance. While I'm not a computer novice, I'm no longer a programmer and plan on using my HTPC as my primary personal PC (for common every day tasks) along with playing media. I'm definitely NOT a power user who will be doing any hard core gaming or other extremely resource intensive tasks on my machine, so I would classify myself as the "average" user.

So I was trying to talk out loud if you will that I had the correct understanding, not just so I understood correctly, but in the event there were other people planning to use their machine in a similar fashion as I am, they would know not to be concerned about the differences. I suppose it's my fault for being analytical (comes from being a programmer). Someone says don't worry it won't be a problem, I always want to know more detail around why it won't be a problem, which helps me to understand better.
post #169 of 199
Quote:
Originally Posted by bobby2478 View Post

While I've heard several times that for normal usage there would practically be no difference, it was hard for me to quantify exactly in which situation I would encounter a difference in performance and what this difference would mean. The last thing I want was to get the drive and think "gosh, if I had just gotten the faster drive it would have been money better spent."

If you want to quantify results, I think the following illustrates it best:
http://www.tomshardware.com/reviews/sata-6gbps-performance-sata-3gbps,3110-7.html





post #170 of 199
Quote:
Originally Posted by ilovejedd View Post

Both review sites above were comparing Intel 330 120GB and Intel 520 240GB. Of course the 120GB model would have fewer populated channels. If you check out the following reviews, you'll see that at the same capacity, the Intel 330 and Intel 520 both use the same NAND and channel configuration:
http://www.rwlabs.com/article.php?cat=&id=617&pagenumber=4
http://www.rwlabs.com/article.php?cat=&id=625&pagenumber=4

That's helpful, thanks! All the reviews I read (I admit I didn't spend a lot of time searching for reviews, I googled "Intel 330 reviews") always compared the 330 to the 520 across different capacities. It would be better to compare apples to apples like the articles you linked to, and not apples to oranges like in the reviews I read.
post #171 of 199
Quote:
Originally Posted by ilovejedd View Post

If you want to quantify results, I think the following illustrates it best

I saw those and those tests make sense, however don't mention the 330 vs 520 comparison. Some reviews comparing the 330 and 520 don't test based on common scenarios like the ones you posted, they use much more complex tests that don't make any sense if you're not a power user. Most reviews are naturally geared towards more technically advanced readers, so they don't always connect with less technical or "average" users. Replicating a bunch of very complex resource intensive tasks is fine for running benchmarks, but if your average user will never encounter the scenarios being tested, then none of it will really matter.

If a user is not very technical and never uses their machine in advanced scenarios like the ones being tested, then those numbers by themselves don't provide the necessary context to really understand what they represent.
Edited by bobby2478 - 7/23/12 at 3:05pm
post #172 of 199
Quote:
Originally Posted by bobby2478 View Post

I saw those, however unless you are very technical in nature (of which I'm not) those benchmarks don't mean very much. So while I see all sorts of numbers on the different spec tests, for me they need to be put in context. Some reviews do a better job of others when framing up context around the tests so you can see which usage scenario their test is meant to replicate. Replicating a bunch of very complex resource intensive tasks is fine for running benchmarks, but if you're average user will never encounter the scenarios being tested, then none of it will really matter.
If a user is not very technical and never uses their machine in advanced scenarios like the ones being tested, then those numbers by themselves don't provide the necessary context to really understand what they represent.

Huh? Those images I posted show the time it takes (in minutes and seconds) to do a 16GB File Copy (various file sizes and types), Steam Backup and Windows Boot Times respectively comparing various SSDs (on both SATA 3 and SATA 2) and an HDD. No technical knowledge required unless you don't know how to count minutes and seconds. I don't know how much simpler you can get.
post #173 of 199
Quote:
Originally Posted by ilovejedd View Post

Huh? Those images I posted show the time it takes (in minutes and seconds) to do a 16GB File Copy (various file sizes and types), Steam Backup and Windows Boot Times respectively comparing various SSDs (on both SATA 3 and SATA 2) and an HDD. No technical knowledge required unless you don't know how to count minutes and seconds. I don't know how much simpler you can get.

No, I said the metrics you posted DO make sense, they just don't show the 330. The problem is OTHER reviews I've read specific to the 330 compare things that sound greek to me. For example:
  • 4kb Random Performance
  • 128kb Sequential Performance
  • Incompressible Performance
  • Storage Suite v1 and PCMark 7
  • 2mb Sequential Transfer read/write
  • 2mb Random Transfer read/write
  • 4k aligned read/write
  • 4k write latency

All of those things sound like gibberish to me as a non-technical user. "OS Boot Time" on the other hand is a very meaningful benchmark I can comprehend.
post #174 of 199
Quote:
Originally Posted by assassin View Post

Hmmm.
So performance is the same according to you for HTPC (I agree completely).
And in your opinion reliability is the same (you state that "they are all reliable")
So we should just choose the cheapest drive? And they are all exactly the same?
I respect your opinion MFusick but I just don't agree on this one. There are differences that set these drives apart. Either you are confused or you are starting to see the arguments/points that many others are making in this thread.

No You got it right.

Reliability is about the same for all.
Performance is about the same for all.
Price is not always the same.

Therefore- Price is most important.

Performance is second most important. (to me)

reliability is not certain prediction so any decent 3 year warranty drive it's really a non issue.

I am not scared at all and so reliability continues to be the last thing I look at.

I value performance and price ratio much more than the false idea one drive is better than another.

I been stating this for months.

If one drive is cheaper I would get that.

if they were the same price I would get the one that has better performance.

If they were the same on both performance and price I would buy the one that is not a Crucial M4.

I would choose a Sandforce drive over a Marvel drive, but I would choose a Samsung first assuming the price was the same.

I will never buy a Crucial drive because I think they are over rated.

I think you know all this.
post #175 of 199
Quote:
Originally Posted by bobby2478 View Post

No, I said the metrics you posted DO make sense, they just don't show the 330. The problem is OTHER reviews I've read specific to the 330 compare things that sound greek to me. For example:
  • 4kb Random Performance
  • 128kb Sequential Performance
  • Incompressible Performance
  • Storage Suite v1 and PCMark 7
  • 2mb Sequential Transfer read/write
  • 2mb Random Transfer read/write
  • 4k aligned read/write
  • 4k write latency

That was after you edited your post. tongue.gif

Anyway, it doesn't really matter that the Intel 330 isn't included in that line-up. You can just follow the trends - which is basically it doesn't matter which SSD you get or if you're using SATA 2 or SATA 3, they all perform around the same with normal usage. The reason why review sites don't bother posting more "practical" benchmarks is because there's practically no difference between various SSDs with normal usage. Speed of SSDs far exceed most single users' ability to tax them.

As for the various metrics, sequential read/write basically tells you how fast the SSD can do large file transfers (e.g. ISO files, etc). Random 4K performance and latency is primarily what affects operating system and application performance (of course, it also depends on the specific application).

To put things in perspective, I'll use Storage Reviews' "Real-World" Benchmarks from their Intel 330 review.
Quote:
For the average user, trying to translate random 4K write speeds into an everyday situation is pretty difficult. It helps when comparing drives in every setting possible, but it doesn't exactly work out into faster everyday usage or better game loading times. For this reason we turned to our StorageMark 2010 traces, which include HTPC, Productivity, and Gaming traces to help readers find out how a drive might rank under their conditions.

The first real-life test is our HTPC scenario. In this test we include: playing one 720P HD movie in Media Player Classic, one 480P SD movie playing in VLC, three movies downloading simultaneously through iTunes, and one 1080i HDTV stream being recorded through Windows Media Center over a 15 minute period. Higher IOps and MB/s rates with lower latency times are preferred. In this trace we recorded 2,986MB being written to the drive and 1,924MB being read.


2,986MB writes and 1,924MB reads gives us a total of 4,910MB over a 15 minute span. Assuming you're able to fast forward through time, then drives are able to finish each task in the following amount of time:
Intel 520 240GB: 10.91 seconds
Intel 330 120GB: 13.34 seconds
Intel 320 300GB: 32.30 seconds

Unfortunately, unless you're in possession of super powers, this task will still take 15 minutes regardless of which drive you use. In fact, for this usage scenario, I reckon even an HDD will suffice.

Quote:
Our second real-life test covers disk activity in a productivity scenario. For all intents and purposes this test shows drive performance under normal daily activity for most users. This test includes: a three hour period operating in an office productivity environment with 32-bit Vista running Outlook 2007 connected to an Exchange server, web browsing using Chrome and IE8, editing files within Office 2007, viewing PDFs in Adobe Reader, and an hour of local music playback with two hours of additional online music via Pandora. In this trace we recorded 4,830MB being written to the drive and 2,758MB being read.


4,830MB writes and 2,758MB reads for a total of 7,588MB over a 3 hour span. We'll do the same calculations as above.
Intel 520 240GB: 22.12 seconds
Intel 330 120GB: 32.99 seconds
Intel 320 300GB: 50.25 seconds

Good news if you happen to be The Flash or Superman, the above means you can finish what you're doing in as little as 22.12 seconds if you get the Intel 520 240GB. If you're among us mere mortals, again, does it matter? At most, you'll be shaving a second here, a couple of seconds there. If you get an SSD, all it means is the computer will be waiting for you instead of you waiting for the computer.
post #176 of 199
Thread Starter 
Quote:
Originally Posted by Mfusick View Post

No You got it right.
Reliability is about the same for all.
Performance is about the same for all.
Price is not always the same.
Therefore- Price is most important.
Performance is second most important. (to me)
reliability is not certain prediction so any decent 3 year warranty drive it's really a non issue.
I am not scared at all and so reliability continues to be the last thing I look at.
I value performance and price ratio much more than the false idea one drive is better than another.
I been stating this for months.
If one drive is cheaper I would get that.
if they were the same price I would get the one that has better performance.
If they were the same on both performance and price I would buy the one that is not a Crucial M4.
I would choose a Sandforce drive over a Marvel drive, but I would choose a Samsung first assuming the price was the same.
I will never buy a Crucial drive because I think they are over rated.
I think you know all this.

But that's not what...

Ah, just forget it.
post #177 of 199
Thanks for the more specific examples IloveJedd, that helps. I updated my initial post right after I initially posted as I realized it wasn't as clear as I'd hoped, however this was before you posted your reply.

In the end I absolutely understand why they compare drives based on those less practical measurements because that's all they have to go by to really demonstrate what's different between devices since they're all so much faster then HDDs.

As a typical consumer that likes to understand more about what they buy, naturally I look at specs and try to buy the "best" option available that will suit my needs and is in the price range I'm looking for. So naturally, when I see some specs much higher then others, I mistakenly assume that by getting the slower device I might regret not getting something faster. My home PC is a prime example, I bought from Dell several yrs ago and while it seemed fine at the time, there are times where it's painfully slow. So now I wish I had gotten something with better performance at the time instead of what I got. I want to avoid the same situation with the new system I will be building.

The review sites a lot of times show bar graphs like the ones you illustrated in more detail, but unlike your very detailed explanation, they don't always translate those metrics into actual performance. i.e. this task would take you 10 seconds on Drive A, and 13 seconds on Drive B. In this context, it all comes together and you see that isn't a huge deal. You can't always translate MB/s, IOps or Average Latency into a performance value that resonates.

You on the other hand did that very well, which is a great help. It's unfortunate that most of the reviews I read on PC components (not all, but a lot of them) don't always go the extra step to translate their findings into something an audience without advanced hardware knowledge could actually understand. It's fine for their primary audience, who I'm sure understand very well. It just makes the job of a a non-pc entusiast or less-technical user, shopping for components when building their own system, a little more difficult.

Which is why forums like this are valuable: they provide a vehicle for people of all knowledge levels and skill sets to ask questions, and for more knowledgeable people to answer those questions and share their opinions. A lot of variety: some people are experts on PC Hardware and building their own systems. Others may know more basic things about PC hardware and are novices at building their own systems, but want to learn more about what they are doing.

So I find these types of discussions very helpful because I'm learning more about the various components that ultimately will end up in my new machine. The primary objective of me joining this forum was so that I could ask questions to people who know, and can help me to learn. So I greatly appreciate all the helpful posts and observations.
post #178 of 199
Quote:
Originally Posted by bobby2478 View Post

My home PC is a prime example, I bought from Dell several yrs ago and while it seemed fine at the time, there are times where it's painfully slow. So now I wish I had gotten something with better performance at the time instead of what I got. I want to avoid the same situation with the new system I will be building.

I highly doubt buying a faster PC (for that generation anyway) would have helped. My cousin bought a top of the line Pentium IV PC (at the time, his CPU was just one step below the Extreme Edition) and right now, it feels darned slow compared to my systems with entry-level Celeron E3300 Wolfdale (running Windows 7 from a 5400RPM drive).
post #179 of 199
Quote:
Originally Posted by ilovejedd View Post

I highly doubt buying a faster PC (for that generation anyway) would have helped. My cousin bought a top of the line Pentium IV PC (at the time, his CPU was just one step below the Extreme Edition) and right now, it feels darned slow compared to my systems with entry-level Celeron E3300 Wolfdale (running Windows 7 from a 5400RPM drive).

Modern pc just feel faster.

I have a high end E8500 core2 duo system that when built was top of line.
It has 120GB sata2 SSD and over clocked 4GB ddr2 with a 3870X2 raedon card.

It was expensive back when I built it and cutting edge.

Today my $300 HTPC with a $60 g630 CPU seems light years faster.

While CPU specs are just slightly lower (the core2 is over clocked at 4.0ghz) it's obvious the modern pc is superior.

It's because of faster SSD drives and SATA3 ports today with much faster DDR3 1600mhz memory.

You don't need a cutting edge i7 today to see major performance improvements because most times the CPU is not the bottle neck.
post #180 of 199
Quote:
Originally Posted by Mfusick View Post

It's because of faster SSD drives and SATA3 ports today with much faster DDR3 1600mhz memory.
You don't need a cutting edge i7 today to see major performance improvements because most times the CPU is not the bottle neck.

In my case it was my FSB that was slow. I got my Dell about 5-6 yrs ago, but I got a refurbished one as it was a great deal. So even when I got it, it was already a couple yrs old at that point, but upgrade from my old machine was dramatic (old machine was about 7 yrs old). The CPU isn't that bad, 2GHZ AMD Athlon dual core, but my FSB is only 533MHZ. So stepping up to a machine with a much faster FSB (1600MHZ) in and of itself will make a huge difference.

Not to mention I'm still rocking the "Vista Virus"...
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Home Theater Computers
AVS › AVS Forum › Video Components › Home Theater Computers › Interesting SSD Reliability Study