SSDs vs. Seagate SSHD (Momentus XT) - a discussion

[TheFluff]

Originally Posted by Ledgem

So what is the bottleneck? Monitoring normal disk read rates outside of benchmarking programs, the XT wouldn't go much over 5 MB/s; I've seen the 840 go to 18 MB/s. I'm not disputing the notion that a program's sloppy coding can create a bottleneck of its own, but faster base resources should result in a speed boost regardless.

If you're not seeing more than 18MB/s read rate in real situations you have a gigantic problem somewhere. Try just copying a file from one folder to another on the same disk; I get well over 40MB/s throughput (that is, 40MB/s read and 40MB/s write) doing this on a perfectly ordinary 2TB 7200rpm HDD. On the SSD (a Crucial M4, 128GB) it's more like 150MB/s.

Originally Posted by Ledgem

I'm referring to program loading, not searching.

Then it's irrelevant. It's not reading your entire 11GB mailbox on startup.

Originally Posted by Ledgem

Interesting, but shouldn't the fact that my benchmark speeds aren't terribly far off from Samsung's advertised speeds indicate that this isn't the case?

I thought you said you were 100MB/s short? That's about a gigabit/s right there.

Originally Posted by Ledgem

I'm reading your posts, and I'm saying again that SSDs over overhyped for the average user and in light of the existence of SSHDs if that's the case.

I think you have a different idea of the average user than I do. Your idea is probably a better fit with the global average, but consider where we are: people who read these forums most definitely aren't average users, and I'd wager quite a bit of money that most of them do a lot more demanding things with their computers than you do.

Originally Posted by Ledgem

Look, a SSD should theoretically theoretically beat out anything with mechanical parts, period.

And they do. There's nothing theoretical about it.

Originally Posted by Ledgem

I'm not just talking about sustained read/write speeds (which seems to be the focus of your points), but "seek times" (random read/write) and overall responsiveness.

Then you've misunderstood me. SSD's don't have seek times, that's pretty much the entire point with them. That one thing is what makes the big user experience difference. The read performance advantage compared to a HDD is mostly irrelevant unless you spend all day unraring scene releases or copying gigantic files around.

Originally Posted by Ledgem

The issue I'm bringing up is that, despite benchmark results and common opinion, it doesn't seem to be that much faster than a SSHD.

Again: I guarantee that if you try some actual demanding use cases, you will see a gigantic difference. The fact that you won't do that doesn't make SSD's slower.

Originally Posted by Ledgem

Even though your choice of wording sounds like disagreement, what you're saying is in agreement with me. SSDs are faster overall? Yep. How about for the average user? You seem to be saying that they're not. That goes against benchmarks and conventional wisdom of flash memory vs. magnetic and mechanical storage, but that's fine. It matches up with what I'm seeing. So then I suppose you'd agree that a SSHD is the better choice for an average user, at least given current costs and performance. This isn't about bashing SSDs (although grounding them to reality is part of it), but about pointing out just how good SSHDs seem to be.

I strongly disagree with this, and the main reason is that the "average user" with your average use cases wouldn't even know what an SSD is, much less considering buying one. In fact, I'd say you are almost unique on this forum in having so utterly boring usage patterns.

[Ledgem]

Originally Posted by TheFluff

I think you have a different idea of the average user than I do. Your idea is probably a better fit with the global average, but consider where we are: people who read these forums most definitely aren't average users, and I'd wager quite a bit of money that most of them do a lot more demanding things with their computers than you do.

Redefining a term? Arguing for the sake of arguing.

Originally Posted by TheFluff

Then you've misunderstood me. SSD's don't have seek times, that's pretty much the entire point with them.

Yes, thanks for stating the most obvious benefit of a SSD which in no way explains what I'm experiencing.

Originally Posted by TheFluff

Again: I guarantee that if you try some actual demanding use cases, you will see a gigantic difference. The fact that you won't do that doesn't make SSD's slower.

Have I been saying that SSDs are slower? Let me answer that for you because you seemed to miss it in practically every other post I made: no, I did not say that SSDs are slower. I suppose there's no point in restating for the third time about how even "non-demanding" tasks should see a benefit from SSDs in theory.

Originally Posted by TheFluff

I strongly disagree with this, and the main reason is that the "average user" with your average use cases wouldn't even know what an SSD is, much less considering buying one. In fact, I'd say you are almost unique on this forum in having so utterly boring usage patterns.

We've been on this forum for a long time, Fluff. You've mellowed out over the years, but you still seem to have that trolling streak in you. Have you used a SSHD? Have you used a SSD? I'll happily hear your user experiences. If you want to write another pretentious post assuming all of my computing uses while spouting off widely known facts about SSDs, please keep it to yourself.

[Alchemist007]

Got my 840 pro. Cut booting and loading times for everything! Scored about 1000 in AS-SSD. I decided to leave my docs/pics/music on my old drive and link them to the shortcuts (not sure if this is a junction or symbolic link, but it works).

[Ledgem]

Originally Posted by Alchemist007

Got my 840 pro. Cut booting and loading times for everything! Scored about 1000 in AS-SSD. I decided to leave my docs/pics/music on my old drive and link them to the shortcuts (not sure if this is a junction or symbolic link, but it works).

Awesome! What size 840 Pro did you go with, and what were you upgrading from?

[JokerD]

Not sure if anyone has stated this, but if you are not seeing improved load times, try checking your antivirus. During the the windows experence benchmarking, I got a score of 4.3 with AV on and a 7 with AV off.

Also, do not defrag your SSD, fragmentation actually makes the SSD run faster than a normal HDD, and defraging shortens the lifespan of the SSD

[Random32]

I don't think Windows even allows you to defrag an SSD... All you can do is manually force TRIM commands to be sent "Optimize." On a different note, OS X does not support TRIM for SSD's that aren't from Apple, you have to third party software force it to work.

As for fragmentation making it work faster, never heard of that... However, more free space tends to increase SSD performance a fair bit (depending on the controller).

[demonix]

Originally Posted by Random32

I don't think Windows even allows you to defrag an SSD...

Windows 8 does (although they changed the wording from defrag to optimise as it would automatically send the TRIM commands if maintenance is switched on).

[TheFluff]

Originally Posted by Ledgem

We've been on this forum for a long time, Fluff. You've mellowed out over the years, but you still seem to have that trolling streak in you.

Someone telling you you're wrong is not trolling, even if it's become popular to call it trolling these days.

Originally Posted by Ledgem

Yes, thanks for stating the most obvious benefit of a SSD which in no way explains what I'm experiencing.

Since you've continued to show a poor understanding of the benefits of SSD's and a likewise poor understanding of what operations you could expect to see a speedup in, I'd say it's well motivated to try to explain the basics to you.

Originally Posted by Ledgem

Have you used a SSHD? Have you used a SSD? I'll happily hear your user experiences.

I have SSD's in both my desktop and in my Macbook, and the guy who sits next to me at work had a Momentus XT in his Macbook for a while before upgrading to an SSD. For him, it was a big difference, and I can't imagine going back to HDD's either. OS X was basically unusably slow on a 5400rpm HDD, and while the Momentus XT speeded up my coworker's experience somewhat, the jump in performance was not anywhere near as big as the one that those of us that upgraded to SSD's got (most of us also removed the DVD drive from our Macbooks and mounted a HDD there instead, for slower storage). For the record, I work with web programming.

Originally Posted by Ledgem

If you want to write another pretentious post assuming all of my computing uses while spouting off widely known facts about SSDs, please keep it to yourself.

You haven't exactly given me a lot to work with regarding your use cases. You complain about applications not starting as fast as you'd expect them to, and that's about it. The only other use case you've stated was reading video of some unspecified character into some unspecified program.

Your entire argument boils down to "SSD's have really big numbers in benchmarks, why am I not opening Mail.app five times faster"? I've been trying to explain to you that that's not how it works, but apparently that's not relevant or I'm "trolling" or whatever?

[Alchemist007]

Originally Posted by Ledgem

Awesome! What size 840 Pro did you go with, and what were you upgrading from?

256GB, upgraded from a 7200rpm WD.

Originally Posted by JokerD

Also, do not defrag your SSD, fragmentation actually makes the SSD run faster than a normal HDD, and defraging shortens the lifespan of the SSD

I can only think of a few reasons I'm not getting that slightly extra speed I've seen on other benchmarks on AS-SSD, one is that I've made it my OS drive and I've read somewhere that can cause a bit extra slowdown in benchmarking (not too sure how much this is true). Another being that my drive is already 70% full after the cloning (and deleting of unnecessary stuff, I could probably find more). And lastly the way I've connected it is through some sort of bridge, I'll have to look up what it's called exactly but basically the sata and power cables are plugged into the bridge and the bridge is connected to the drive (the bridge is about 1-2 inches in length). I didn't have many options because I only have one other spot to put the drive and it's not physically practical. I'm thinking this bridge might explain it because the slowdown is so small and the .100 latency I'm getting in the benchmark results (vs .03) seems to be an obvious indication.

[JokerD]

Originally Posted by Random32

As for fragmentation making it work faster, never heard of that... However, more free space tends to increase SSD performance a fair bit (depending on the controller).

Oops, I should correct my statement. Defrag does make a HDD go faster due to it eliminating the need for seek times on the disk. However, this does not help for SSD which do not have seek times. For large sequentially stored (non-fragmented) files, HDD might actually have the advantage in speed. However, when it come to fragmented files, SSD are faster.

Originally Posted by Alchemist007

256GB, upgraded from a 7200rpm WD.


I can only think of a few reasons I'm not getting that slightly extra speed I've seen on other benchmarks on AS-SSD, one is that I've made it my OS drive and I've read somewhere that can cause a bit extra slowdown in benchmarking (not too sure how much this is true). Another being that my drive is already 70% full after the cloning (and deleting of unnecessary stuff, I could probably find more). And lastly the way I've connected it is through some sort of bridge, I'll have to look up what it's called exactly but basically the sata and power cables are plugged into the bridge and the bridge is connected to the drive (the bridge is about 1-2 inches in length). I didn't have many options because I only have one other spot to put the drive and it's not physically practical. I'm thinking this bridge might explain it because the slowdown is so small and the .100 latency I'm getting in the benchmark results (vs .03) seems to be an obvious indication.

You might want to check the settings that your OS has the services taking advantage of SSD's speed turned on, like TRIM or whatever, since your OS is cloned and not (presumably) install fresh with the SSD

[Alchemist007]

I did do many things in this guide (including that): http://www.tomshardware.com/reviews/...weak,2911.html , and it improved the random read/write speeds by a bit, but it's still falling a bit short. It's not too big a deal since the differences are so small but I at least want to find the source of the problem. If it's the bridge, I'm going to just test it without it, I'll just have to find a platform in the case to hold the drive...if at least only for this test.

Edit: Wasn't the bridge, got similar speeds.

[Random32]

Various ideas about why you don't get as good of performance:
-SSDs have less consistent and slower performance when they are fuller (still miles ahead of HDD though) as the controller has a lot less space to work with and optimize and stuff. 70% is enough to start seeing the effects of this.
-If you cloned from an HDD, there might be a performance hit if the partitions aren't aligned how an SSD likes them.

The first one is pretty much a fact of life. Either buy an SSD that prioritizes having good performance when close to capacity, don't fill the SSD up, or live with it since its still blazing fast anyways.

The second one, you can realign the partitions if they are off.

[Ledgem]

Originally Posted by TheFluff

Someone telling you you're wrong is not trolling, even if it's become popular to call it trolling these days.

You took part of at least two posts to call my computer usage "boring." It could have been construed as a cute joke the first time, but when repeated and even expanded over multiple posts I have to ask what you think it contributes to the discussion, other than to come off as anywhere from passive aggressive to mildly insulting. You either knew what you were doing - slipping in little insults here and there that prompt responses and frustrate, also known as trolling - or you are socially inept and didn't realize. You've been at this for a long time, so I'm pretty sure you know what you're doing. I'll give you the benefit of the doubt one last time.

Originally Posted by TheFluff

Since you've continued to show a poor understanding of the benefits of SSD's and a likewise poor understanding of what operations you could expect to see a speedup in, I'd say it's well motivated to try to explain the basics to you.

In terms of performance SSDs have faster random and sustained read/write performance compared with HDDs. The latency that was owed to the mechanical parts with HDDs is virtually non-existent with SSDs, and fragmentation is a non-issue. In other words, practically every aspect of a SSD's operation should be faster than a HDD, and you should see a benefit in any operation that relies on read/write operations from the disk, with the difference in performance being more noticeable when operations are performed in parallel or when larger data sets are required.

Do you have a misconception to correct? Can you tell me something that I don't know? Or is this another example of trolling with a condescending attitude?

Originally Posted by TheFluff

I have SSD's in both my desktop and in my Macbook, and the guy who sits next to me at work had a Momentus XT in his Macbook for a while before upgrading to an SSD. For him, it was a big difference

What SSD did he upgrade to, and how did he transfer his data over to the SSD?

Originally Posted by TheFluff

Your entire argument boils down to "SSD's have really big numbers in benchmarks, why am I not opening Mail.app five times faster"?

I did not make an argument. I made the observation that my SSHD wasn't that much slower than the SSD I'm currently using. Your response was that SSDs only really make a noticeable impact in certain usage scenarios. That goes against what one would conclude given the benchmark numbers and mechanical differences between HDDs and SSDs, as well as "common knowledge" formed from heaps of anecdotal stories. I am not suggesting that you don't know what you're talking about, but what I'm trying to get from you is an explanation for why there's that discrepancy between what makes sense given even a basic understanding of the differences between HDDs and SSDs, what the rest of the world claims and thinks, and what you're saying.