W520 support for SATA III or 6GB with Vertex 3 240GB SSD

[macmac]

can i ask if the w520 supports sata iii (despite lenovo not mentioning it )

- do you think the g770 also would?

as it as the HM65 board which offers 2 6gb/s ports.

[doubleoerik]

Just came across this thread and wanted to try to draw upon a larger community to help with my issues...

See my original post here: http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/Please-help-Major-problems-performance-and-loc...

Quick summary - I have a brand new W520 (2 weeks old) that was purchased with the standard 500GB spindle HDD. I have removed the drive and swapped it with an Intel Series 520 256GB SSD and relocated the original drive to the optical bay. Long story short, I cannot get this SSD to run at SATA III speeds despite trying absolutely everything. For those that might ask - it has the latest firmware, my BIOS has been updated (running version 8BET57WW v1.37), AHCI was set in the BIOS before installation of Windows 7, and I've tried installing windows using the msahci driver and the latest and greatest iaStor driver. I've also perused TONS of online forums and tried a number of other tweaks and fixes but nothing has fixed this issue. I have been periodically using AS SSD to check read/write performance (does not verify SATA II or III directly but it can be inferred from the speeds) and just found a cool program called Speccy (by Piriform) that actually reports the transfer mode of the drive (and it is reporting SATA II, see the link above for a screenshot).

Any and all help would be GREATLY appreciated...I'm tearing my hair out.

[Chatbox]

Problem seems to be coming from how Speccy is reporting the information.

I use AIDA64 (and I have the Intel 520 drive, 240GB).  All good here.

[doubleoerik]

What's your system configuration and what driver are you using? Are you using windows 7? Can you post your performance numbers? I would be shocked if you're actually getting SATA III rates...

I highly doubt Speccy is wrong since AS SSD (widely used to compare performance between configurations - not measure real world performance) is reporting the exact rates I should get with SATA II. Plus, I've found a number of other forums that claim this drive can not be run in SATA III mode on this machine currently....

[Chatbox]

AS SSD:

And ATTO:

Intel Chipset 9.2.0.1016

Intel RST 10.8.0.1003

BIOS 1.36

 Windows 7 Pro x64 with SP1.

Personally, I prefer ATTO as it gives me a better drive performance charactistic (IMO), and I can specify the type of data to use for the benchmark (random, in this case).

And here with just 00000000:

So, shocked?

[doubleoerik]

Your read speeds look ok but your sequential write speeds are way low. You should be getting 350-400 using AS SSD. I'm getting 280-290 (which is significantly higher than your 220) using the outdated msahci driver. Actually, I'm getting 480 read and 280 write using the msahci driver.

Something is still amiss...

Would you at all entertain running Speccy to see if it says your in SATA II or III?

Also, look online at other baslines with the intel 520...your numbers are lower than many other baselines I've seen.

[Chatbox]

AS SSD and the Intel 520 gives around 750.  The write data that AS SSD uses/generate for the write test is highly voletile with regard to how compressible it is, and so Sandforce performance is questionable as well for the duration of the write test.  (The score in AS SSD doesn't mean much to me, it's only provided here for comparison reason.  Also, a score of 30 different off from 750 is less than 5% error...I can tolerate that for a screenshot).  Plus I currently have a VM running, and torrents downloading.

I can give you a screenshot with higher write speed...but that's shouldn't be the concern here.  Point is, it's operating in SATA III speed with the raw 00000000 read and write in ATTO.

Your tests should be done with highly controlled data samples.  If you (or me) had a Samsung 830 SSD, and we get highly fluctuating performance result even when using controlled data, then that would be a problem.

Speccy gives me "SATA type: SATA-II 3.0Gb/s" and "Transfer Mode: SATA III"...go figure.  I'm not giving the app much credit at all when it only shows "Manufacturer: Unknown manufacturer".  Sorry, but I think Speccy has issues.

Here's another Intel 520 240GB baseline for seq write: http://www.tomshardware.com/charts/ssd-charts-2011/AS-SSD-Sequential-Write,Marque_fbrandx14,2783.htm...

IMO, AS SSD gives you rougher than rough performance readings.

For me, it was a tough choice between the Samsung 830 (Marvell chipset) and the Intel 520 (Sandforce chipset).  But in the end, I picked the Intel drive for its 5-year warranty...over the Marvell chipset.  From a pure tech performance standpoint, I would have picked the Samsung drive is warranty consideration didn't come into play.

[Chatbox]

Also, if the number games is really troubling you, you can run AS SSD's "Compression Benchmark" from the Tools menu.  You'll get above 400MB/s for write with highly compressible data.

[doubleoerik]

I completely agree - the AS SSD numbers don't mean much in terms of "real world" performance but they are useful in comparing the relative performance difference between systems/configurations (as you mentioned). I also agree with you the Speccy application information should be used as "for reference only" but I appreciate you humoring me anyway.

All of this aside, the fact still remains that I'm not getting anywhere near the numbers you are quoting despite having a very similar system configuration and the same drive. I know the issue is not the drive since I have two and have tried them both. I'm also doubtful that the issue lies with the chipset (at least for a non-expert amateur like me) because if that was the case I would expect many other errors and not just performance issues with different drivers. It's super strange (again, at least for me) that I'd be getting significantly better performance with the msahci driver as opposed to the intel driver. Let me download one of the other performance assessment programs you're using so I have more than one data point.

Thanks for all your help so far!.

[doubleoerik]

I think you're misunderstanding what I'm saying to some degree - I agree that the numbers mean virtually nothing. What I am saying, however, is that assuming all things the same we should be getting the same numbers. I'm getting nowhere near what you are reporting or what others have posted online...