![]()
Uploaded with ImageShack.us
I was wondering how it would effect a SSD if this was implemented.
Not that I need to but was just curious.TBH I have never used it![]()
-
I was hoping to see an answer to this, because I was thinking of doing the same.
Seems like it would cut down on writes of compressible data if the SSD controller did not have built in compression. (if the SSD controller did its own compression, like Sandforce does, I think enabling Windows compression would be a bad idea) I don't think the Barefoot controller in my Nova 64 does its own compression.
Unless the CPU is slow, I'd think it would speed things up too. Not to mention, it would stretch the useful capacity of the drive a bit.
Am I missing anything? -
I would assume it would slow things down...
-
It slowed down our regular hard drive at work when one of our folks "thought it would save some space...". Caused the system to lag terribly when accessing those files. (granted that was a spinner, but doubt would help much on a SSD either).
-
Try it and let us know how it goes. Not much to loose I guess, you can always switch back.
-
-
I have used NTFS compression on my gaming latops for years but I have only ever used mechanical drives. Many people assume it will slow down their systems so they don't use it, and in the process miss out on saving a lot of space. I notice no performance loss at all and it feels exactly the same if disabled.
Of course this may be a different result for SSD drives so I will let some one else comment on that one. -
No real tests from me as I want to switch this option off quickly.
CPU is a P8600.
Before: (not my screenshot but similar enough)
After: (LOL - I'm certain that this is a flawed representation of results)
-
It looks like the files uncompress very fast, but it takes a while to compress them with 1600 CPUmarks? Since a 64GB Barefoot is slower than the 256GB Barefoot (half the memory bandwidth) and my i7-2630qm gets 6300 CPU marks, it might be very interesting for me to try this.
When/if my new HPs arrive, I'll set up the one with the SSD, run some disk benchmarks, then compress the SSD and run the benchmarks again. Is AS SSD the best software for this? -
This is a common issue with NTFS compression. You get huge speed boost on read but write speed suffers. IOW, it is good for c:\windows and c:\program files but bad for user data.
That said, real life data is not that compressible. Mine show about 20%, i.e. using 80% of the original size. Reason speed increment at no cost. Just don't do it on the whole drive. -
I'm very curious as to what's going on with the "after" results. Doesn't AS SSD use incompressible data and therefore should be seeing neither gains nor losses from compression? -
-
I've also no idea why AS SSD is giving such high read speeds.
Either,
1. The data is being compressed (just because it's not easily compressed, does not mean it cannot be compressed - I'm not sure what data AS SSD actually uses)
Or,
2. There's an issue with AS SSD
( or 3. - someone's got a hugh RAM cache??? )
@namaiki - have you contacted AS SSD ? -
@namaiki: Thanks for the clarification. Given how old NTFS compression is, I'd never have guessed that a relatively modern CPU like the P8600 would struggle with it. I guess this is one of the very few areas where a fast CPU could be justified for users who only handle typical office workloads
@EnglishCoder: I'm starting to wonder if the "incompressible" data used by AS SSD is truly incompressible, or merely written in such a way as to defeat Sandforce's particular compression algorithm. -
Any modern CPU should barely be taxed with the compression / decompression on the hard drive or SSD. In actuality it should improve performance because the speed of (de)compression should realistically be much faster than the time it takes to read or write that extra amount of data were it not compressed.
This is curious, I might give it a try and post results. -
Yes try it, I would like to see the results.
-
I think I will try on both my ULV netbook (with SU4100 CPU) and my Sager notebook (i7-2720QM) to see if the lower end CPU can manage the task.
-
-
Have it enabled on my laptop, no perceptible speed loss (but haven't run benchmarks), but I gained a few gigabytes of space back.
-
I just used NTFS Compression on my Kingston 96GB and SU4100 CPU netbook. Went from 52GB used to 42GB used. No noticeable performance depreciation. I let the system idle for an hour logged off Windows to let GC do its thing before testing again.
Here's CDM results:
No Compression:
NTFS Compression:
-
Any ideas why your before/after comparison numbers are so similar? From earlier posts, I'd have expected a huge increase on Read I/O, while the Write I/O might be more dependent on processor speeds.
-
I was surprised too. Not sure. But I was wondering if I ran it with compressible data (all 0's) before and after it might have improved, I dunno, but not a fair assessment.
-
How about a real world test? Maybe a file copy with mixed files.
-
-
I just ran a few tests both with and without compression. I ran each test twice to confirm the numbers, and all results were consistent.
The tests were run on Vertex2 240GB
without compression:
with compression:
So my conclusion is that it gives just a little bit of more space, but at the same time it increases the performance!
Do you guys have any requests for tests? -
What do your AS SSD Benchmark results look like?
-
Thanks for those tests gnarr. What CPU are you using? I think that will dictate performance somewhat too considering it has to compress and decompress on the fly.
-
AS SSD shows that the read speed is dramatically increased and the write speed dramatically decreased.
But I don't really believe much in synthetic tests like this. But I guess PC Mark gives a hint to how it will work in real world scenarios.
But as we know, read speed matters way more than write speed when executing programs. This would mean that installs would take much longer, but loading times in games and apps would be much quicker.
Without compression:
With compression:
The system I'm doing the tests on is a Dell XPS15 L502x. Those are the specs:
Intel Core i7-2630QM (quad-core 2.00-2.90GHz, 45W)
Intel HM67
2x4GB DDR-1333 (CL9)
NVIDIA GeForce GT 540M 2GB DDR3
15.6" B+GR LED Glossy 16:9 1080p (1920x1080)
OCZ Vertex2 240GB
750GB 7200RPM Western Digital Scorpio Black HDD
Gigabit Ethernet(Realtek RTL8168/8111)
802.11n WiFi + Bluetooth 3.0 (Intel Advanced-N 6230)
2.1 JBL Speakers + Waves Audio
6-cell, 11.1V, ~5.0Ah, 56Wh
It seems to me that this is a "hard to write easy to read" compression algorithm, and I would guess that the CPU won't matter alot.
I would love to see a similar test run on a system with a faster SSD and a slower CPU. Then we can see if it's always a good thing to keep the compression on. -
compression on for mostly read only folders like windows, program files, off on user folders
-
I cannot even believe that they have that checkbox still there. It is vestigial from the days of "Driver Doublers" software. YUCK!
-
If you guys are interested in reading a small article I just put together on this:
Google Þýða
Ps. It's a google translate Icelandic -> English. -
And been using it with my netbook 96GB SSD and saved 10GB already, and hasn't affected performance or battery life. -
Me personally, I will not be checking that box for any reason in this lifetime. Just do not need any extra space, I already have way more than I need.
-
On my external it's a boon because you can't get any slim USB powered hard drive > 1TB at the moment, so it helps me. I use it to store all my Steam games. -
I just did one "real world" test on Portal 2.
Install time was 18 minutes and 56 seconds on an uncompressed drive and 20 minutes and 9 seconds on a compressed drive.
Loading times decreased with compression, so maps loaded quicker.
Dark purple is compressed and light purple is uncompressed. Lower is better:
-
I have always been under the impression that one of the reasons not to do compression is that it makes data more difficult to recover if you develope a bad sector on a spindle drive.
I wonder how that would affect an SSD?
btw I have to believe that with SSD space at such a huge premium if compression was a good idea it would be generally known and used. -
bad sector is a thing of the past, the HDD has its internal way of handling that.
-
-
I don't think compression would be beneficial with Sandforce because it's already compressed. But other drives that don't use compression it works. It doesn't offer much space savings usually, I've seen about 10% in general, but for a 120GB drive, 12GB extra is always welcome. -
I recommend against using it. I have just seen it mess up too many times in the past. OK, they were older computers, but still. Many of the issues come when you later UNCHECK the box for some reason.
-
-
If you want to save even more space, Drive Press 2.0 (included with MagicRAR 6.3) saves even more space:
Simon King`s MagicRAR: Drive Press 2.0
I have really saved twice as much space on my Windows 7 x64 SSD, compared to what Windows had saved me previously.
They seem to use NTFS as the underlying storage, so I don't know how they manage to compress better...
...but the results speak for themselves, especially when the compression is much faster too! (It's multi-core so you can max out the CPU when running it on fast SSDs).
What if I checked the box "Compress this drive to save disk space"....
Discussion in 'Hardware Components and Aftermarket Upgrades' started by OneCool, Apr 16, 2011.