SSD Thread (Benchmarks, Brands, News, and Advice)

Can you show us some results? CDM 3x 100 MB will be enough.

I remember you also said your Intel X25-M is outperforming your Crucial C300. Sounds like something is wrong.

 

Here is the MSAHCI Driver:



Here is the Intel RST Driver 10.6.0.1002:

 

And for a comparison to the OS C300 in my W520:

With MSAHCI Driver:


With the older Intel RST 10.1.0.1008 driver:

 

And the 256 GB C300 data drive in the W520. Yup, it is slower than its 120 GB cousin except in small random read/writes which of course is very important.

Also my back and forth testing has shown that the ultrabay caddy (Lenovo) is actually the faster controller as my OS 128 GB drive is faster in that location.

 

Thanks. 4K random performance seems limited by the chipset. This is the reason MS AHCI and Intel G2 look faster.

To test it you could rerun the tests with CDM 3x 100MB with Intelppm temporarily turned off.

The reason I don't advise AS-SSD is because it creates a lot of wear. Iirc 15 to 20GB on each run.

 

Just got my WD SiliconEdge 256GB that I paid $200 for. Debating whether to keep it or make $100-$125 off it. I don't really NEED the SSD now, but I can't resist using a new one. lol.

 

You will definitely be getting more ROI by keeping it at that purchase price (assuming you have no problems with it...) than by selling it even for $400.

Enjoy it!

Almost every SSD I have tried has been a letdown - but everybody on this forum told me that it was my system, a defective SSD (all of them?!!!) or a combination of that...

However, here is a good thread that shows just how bad the older (SF) SSD's are - failure at 24TB written - (well, I think it was still writing at 6MB/s still - not 'dead', but dead enough).

See:
SSD Write Endurance 25nm Vs 34nm - Page 32


The 25nm nand paired with the Intel controller (320 series) is pretty impressive. The Samsung 470 and Crucial, not so much...

 

so, when's your big "see how hdds are so much faster than ssds" thread coming?

 

Actually, it's not coming - I've finalized my setup and just doing a little extended testing of it before I recommend it for others to follow.

I know, more waiting!!! Sorry.

 

terrible...


you know btw that i don't care about all that that much. i just want you to show a simple repeatable example that proves a hdd can be faster than a good ssd.

 

For anyone waiting for Intel 710 series,

Intel 710 Series Lyndonville SSDs Get Priced

 

i read that they will build them with 2x nm NAND , and the 2nd version with 3x nm NAND; why this change? only the price discount or more?

 

Um, oh boy Did I kill my C300??? Tried to update the firmware to 0006 and now in wont boot.

 

Try booting from another device. For example a Windows 7 installation DVD.

 

You need to be careful on updating the firmware on SSD's. I don't know how many of them do this: The Samsung 470's will wipe itself clean on a firmware update, samsung even states it when you go to check for updates on their software. Most manufacturers even warn of the possiblity of a clean wipe on firmware updates. My Samsung 470 is due for a firmware update, but I'm not going to as according to what I read it dosn't do much and mine is running fine, not enough changes to warrant a disk clone, update firmware, then disk clone back to SSD. Pay attention to all warnings and instructions that come with updating firmwares in SSD's. Most likely this is what happened to you, and you will need to fully re-install everything, unless you cloned the drive beforehand. Goodluck and let us know what happened.

 

Hmmmm! Actually did a reinstall with (Alienware Respawn) and it went through the whole process. But still doesn't detect the drive when booting. About to try to update the firmware with a CD/DVD (USB first time) and see if that helps. Called a Crucial tech and he suggested this. If not....RMA and get another I guess.

 

Well-it's dead. Crucial said RMA it. External won't even recognize it. Sometimes the BIOS will-like 1 outta 3 times. Wierd.......had the same problem about a week ago and it 'fixed' itself. Also, if the computer would lock up and I had to hard restart it did it also (2x). Oh well...got one coming.....

 

Sounds like a bad SSD.

I have updated 3 of my 4 SSDs at least once each without problem. The 4th hasn't needed it yet.

 

Heads up! 8MB bug on Intel 320 Series!

 

I have confidence in Intel to sort this out quickly and efficiently.

See:
http://forum.notebookreview.com/sol...662-intel-320-buggy-firmware.html#post7692410

 

Battery life comparisons: Crucial M4, Intel 320, Samsung 470
http://www.ssdreview.com/review/com...gb-25-inch-4pc10302,54/power_consumption.html

As you probably know it's Sandforce... it will perform like Sandforce. There's a couple of reviews online that confirm it.

Reliability will also be like Sandforce I'm afraid.

Thanks for mentioning. I hadn't seen this yet.

 

Yes, the disappearance of a formatted drive to 8 MB with nothing on the 8 MB is pretty well documented for the Intel 510, 320 and even a few tried and true X25-Ms.

I have never backed up any systems as much as the new ones I have that rely on SSDs. Luckily I have only needed the backup once and that was after my first SSD install where I was not really up to speed on what part of the Sys and C: partitions to make active and which to make bootable. In fact I couldn't tell you now, but I do know that you need to know that and get it right if you are doing a recovery from backup to install your new SSD.

 

I assume you mean this (for the x25m)

Help! X25-M 80G is changed to intel...: Intel Communities

Based on what is described there, the drive has kind of switched to another mode(thus the change in volume label) and added to the other reported case for the 320 of it is possible to restore to normal, it sounds like there is a hidden diagnostic partition(and test mode) which got turned on due to some oddball PC <-> SSD communication.

So it is very likely that all Intel SSD has this 'feature' which should not be turned on except with a special combination of 'easter egg' triggering mechanism. We actually see this all the time in cell phone, even some LCD monitors.

 

Has anyone tried disabling LPM for performance boost?

The results seems very impressive. A guy with Crucial M4 64GB on SATA II manages to get more AS-SSD points than a Vertex 3 120GB on SATA III.

Results: http://forum.notebookreview.com/sol...e/594298-disabling-lpm-performance-boost.html

 

i did it on a c300, but haven't tested it. i just noticed possible stutter, read it helps, and never noticed stutter afterwards. so i guess it helped.

 

Yes it's a known fix for stutters. And besides that it seems to boost 4k writes. Could work on other drives too.

 

Tried it, didn't change anything on my setup, the only combination that works best for me is to disable intelppm+C1E+EIST with the known side-effects...

 

The OCZ Vertex 3 in 240 GB size is available and on sale at Newegg.

After reading the Newegg reviews I wouldn't take it if it was free.

 

Not just the Newegg reviews either...you should see what is on the OCZ forums, keeping in mind that OCZ likes to randomly delete posts too that bash them. I'm with you.

 

Hi. Can someone please advice me on which would be the ideal SSD for my notebook? I have an Envy 13. I search online and saw that it only support 1.8" hard drives. I would like to upgrade to an SSD since the stock HDD installed on mine performs very poorly.

I am not after benchmarks or anything, I just want a quick boot up and quick loading up of programs.

Looking through Newegg, I see the following 1.8 SSDs:

Crucial RealSSD C300 CTFDDAA256MAG-1G1 1.8" 256GB SATA III MLC Internal Solid State Drive (SSD) - $459.99

Intel X18-M Mainstream SSDSA1MH160G201 1.8" 160GB SATA II MLC Internal Solid State Drive (SSD) - OEM - $427.99


Is there any reason for me to be worried that the 256gb Crucial SSD is priced just the same as the 160gb Intel? I would prefer a drive that doesn't require much user intervention/tweaking (other than software/firmware updates?).

Any of you guys got experience from these two drives? Your input is much appreciated.

 

OCZ Unveils Its SSD Controller: Dual-Core ARM, DDR3, TLC Flash

 

Found an interesting piece of information on Macrumors...

 

The problem with OCZ is that they focus on performance instead of reliability which in my opinion the latter is harder to achieve.

 

How credible does the quote sound to you Stamatisx?

 

It's hard to tell, I don't have any numbers at my disposal to confirm the percentages mentioned there but so far the reports about RMA'd drives are many.
It's really hard to have the fastest and at the same time the most reliable SSD in the market while having it at a competitive price.

 

It is very credible and in fact explains not just them but probably Intel's as well.

The really 'faulty' (which usually means hardware issue) ones are very very rare. All the rest that needs RMA are almost guranteed to be 'locked/panic', IOW bugs caused by corner cases(and why replacement unit usually cannot help).

In Windows terminology, it present itself as BSOD, in linux kernel panic.

and statmatisx is correct as well that when you push for performance, you are usually cutting corners in the software(usually due to complexity). That is the key reason why multithread programs are so hard to write or why kernel drivers aree so hard to write.

 

Interesting data from my Lenovo W520 and X220.

I have tested Intel X25-M, C300 128 GB, C300 256 GB and M4 256 GB and in each case (X25-M in the X220 and the others in my W520) the drives benchmark faster with the older MSAHCI driver than either the newer Intel ...1008 or the newest Intel driver (10.6.0.1002).

And they benchmark faster whether using CDM, AS SSD or PC Mark 7 or PC MarkVantage. Consistently faster and on every test.

I am not saying they will be faster with the older driver in your machine but they are that way in my new Sandy Bridge laptops.

Don't have a clue why that it, it just is.

 

I'm pretty sure it's because the Intel RST driver enables more power saving features (like LPM), which makes synthetic results look worse.

If you really want to know which is faster do some real world benchmarks. Especially a small multi tasking scenario would be good way to measure it.

Another way to check it would be to see what synthetic results you get with Intel RST enabled but LPM disabled.

 

I experienced also that Microsoft MSAHCI driver gives better scores with the various SSD test suites(CDM, AS-SSD, etc), but the Intel Rapid Storage driver was better in real life. Compiling a Winbuilder project the turnaround time dropped by about 30 secs with the Intel driver.

 

Although loaded with more video and audio encoding than I normally use I think PCMarkVantage is very much a real world benchmark (at least for those that game, and encode video and audio files).

AS SSD and CDM are not.

 

maximinimaus, pkincy

try it with intel RST installed and LPM disabled please

I think you'll get even better scores. Like the guy here: http://forum.notebookreview.com/sol...e/594298-disabling-lpm-performance-boost.html

As far as I know it's close but it's not real world. It plays back captured traces.

But besides that, it was made for Windows Vista, I'm not surprised it can't take advantage of intel RST.

 

Sorry I don't get better results. The turnaround time for my Winbuilder project is:
Project time: 1 minute, 55 seconds and 347 miliseconds
Project time: 1 minute, 54 seconds and 894 miliseconds (LPM disabled)

That's negligible!

I have the newest Intel Rapid Storage software 10.6.0.1002 installed.

I also know the cause. LPM can be disabled by the BIOS.
I checked the capabilities of my BIOS with Intels RaidCfg.exe(can only run in an DOS environment) and got following information for my SATA ports:

4.7 Port Implemented Bit Set for Port 0.......................ENABLED
4.8 PI Bit Matches PE Bit for Port 0.............................PASS
4.9 Aggressive Link Power Management on Port 0...............DISABLED
4.10 Hot Plug Support on Port 0..............................DISABLED
4.11 Interlock Switch Attached to Port 0..........................N/A
4.12 ALPE and HPCP Cannot Both be Set on Port 0..................PASS
4.13 ISP Requires HPCP on Port 0.................................PASS
4.14 Port 0 Command Register Locked..............................PASS
4.15 External Port Bit Set for Port 0........................DISABLED

4.7 Port Implemented Bit Set for Port 1.......................ENABLED
4.8 PI Bit Matches PE Bit for Port 1.............................PASS
4.9 Aggressive Link Power Management on Port 1...............DISABLED
4.10 Hot Plug Support on Port 1..............................DISABLED
4.11 Interlock Switch Attached to Port 1..........................N/A
4.12 ALPE and HPCP Cannot Both be Set on Port 1..................PASS
4.13 ISP Requires HPCP on Port 1.................................PASS
4.14 Port 1 Command Register Locked..............................PASS
4.15 External Port Bit Set for Port 1........................DISABLED

The suggested improvement of LPM off depends also on the implementation in the BIOS.
Deactivating LPM has no effect for me as the BIOS overrules the registry.

 

Thanks maximinimaus.

And it's good to know the Intel RST driver performs better in real world.

Sofar I've only seen 'the disable LPM tweak' boosts Crucial SSDs performance.

 

Hi, I have a question regarding Toshiba 1,8" SSDs that Lenovo sells/sold. Does anybody have one and/or could provide me with benchmarks?

Why am I asking this: My company uses Safeboot (McAfee Endpoint Encryption) with all laptops. What I want to know is the performance impact of it!
These numbers were taken on an X201 i7 with such an 1.8" SSD in a 2,5" adapter on Win7 using Safeboot (McAfee EE with a 256bit AES - FIPS, not AES-NI) CPU utalization wasn't limiting during the tests.




Either these Toshiba SSDs are VERY slow compared to others or safeboot just cripples the performance here!
Can anybody put these numbers into perspective? Can't find much about them.
In order to get a higher performance I ordered me an Intel 320 with hardware AES. I want to be able to provide my IT department with some hard facts (numbers) in order to convince them to let me use hardware encryption over McAfee EE. That's why it would be a real help if I could find some performance numbers on the Toshiba 1.8" drives!

thanks

 

According to this webpage: Maximum read speed 240MB / s; maximum write speed 200MB / s

So it seems fair to conclude that something is seriously crippling performance.

Perhaps the 2.5" adapter isn't helping either...

 

The adapter shouldn't do anything.
It has to be this software full disk encryption. It is running with the standard FIPS cert AES256 algorithm, it could do AES-NI (Intel) but as I said, cpu-utalization wasn't significant.

 

randfee,

cpu utilization doesn't have to be significant for a software based full disk encryption program to be crippling performance like that. McAfee is simply very inefficient, obviously.


Can't you simply uninstall this to confirm if it is junk or not? Rather than buying another SSD?

Like A/V, FDE takes it's toll with a storage subsystem's performance - even if cpu utilization remains near/at zero.

 

Hi I was wondering why am I getting such low scores for my write speed for my crucial c300? read speeds are fine, but the write is low?

 

I'm assuming you haven't applied any tweaks (if you did I suggest to reverse them).

You could try how write caching settings influences write performance.

PS. 3 x100 MB will be fine for testing and reduces wear.

 

If your saying to turn on my writing cache in the ssd. Its on. It just drastically dropped when i finished with all the updates for windows when i did a fresh install. I'm only getting around 85 write speed but my read is good (331)