Intel PM55 chipset and X25-E SSD Very low 4K random reads/writes Alienware M17x -R2

[idata]

I am trying to figure out if there is something wrong with the PM55 chipset from Intel

This is a screenshot taken from my Alienware M17x - R1 with an NVIDIA chipset (open the pictures on a new tab or window to see them)

http://img843.imageshack.us/i/4kcrystalformated.png/

This is from my Alienware M17x - R2 with Intel Chipset using Dell's drivers chipset drivers and the Microsoft's default drivers for storage.

http://img299.imageshack.us/i/slow.png/

This is after a clean Windows installation with only the Dell chipset drivers and for storage the latest RST drivers from Intel

http://img819.imageshack.us/i/latesty.png/

This is in safe mode

http://img413.imageshack.us/i/safemodee.png/

I have tested the SSD and it's fine and I have also performed a secure erase to the disk before each clean Windows installation to make sure the SSD's performance is set back to factory defaults (8 secure erases and 8 clean Windows installations so far).

I have also tried every possible driver combination (Matrix Storage drivers from both Intel and Dell, RST, chipset drivers from both Intel and Dell) and I haven't managed to get the performance close to half of what it was with my R1.

Note, the mode is set to AHCI from inside BIOS, I get the exact same performance when Windows are installed with the RAID option instead of AHCI.

The same behavior is noticed among other manufacturers like Asus, Dell, Lenovo, HP

http://forum.notebookreview.com/6399977-post2096.html

http://forum.notebookreview.com/5519041-post2424.html

http://www.ocztechnologyforum.com/forum/showthread.php?69573-OCZ-Vertex-120GB-Slow-4K-Read-Results-on-Asus-G51J-Notebook

http://thenextbench.net/t5/HP-Envy-Notebooks/Poor-SSD-Performance-in-Envy-4k-Read-Write-Times/m-p/55680

Is this by any chance the reason for this problem?

http://hardforum.com/showthread.php?p=1035010918# post1035010918

[idata]

In device manager for the drive in properties > policies tab does checking the "Turn off windows write-cache buffer flushing on the device" help?

[idata]

Depending on the driver, after trying all of them with or without write-cache, it only makes a 2 MB/s difference, either to the better or to the worse. It's not that. I have already eliminated that factor from the very beginning . Thanks for suggesting though.

[idata]

Users of the G73JH system have started posting about it also. Since we don't have many options in the BIOS we have seen that to get full performance of the SSD we have to go in to save mode.

The issue has been brought to the attention of an Asus manager so I hope that Intel responds quickly to this issue. There is definetaly something wrong which needs fixing because when you are on external power you shouldn't have to have processor load to have maximum SSD performance.

[idata]

Yeah, I've read the thread at the notebookreview forum, I have reported that to Dell as well so now that Asus is involved too I hope that this issue will catch their attention and do something about it.

[idata]

I find it weird we get no answer from Intel about this. The news of this issue has been widespread now and people on Notebookreview.com are looking into workarnounds doing the least possible amount of harm to the power management.

From the silence of Intel I can gather that they know the problem but don't have a solution for it. I seriously think that the HM55 and PM55 are flawed on a hardware level which can't be fixed. Or in tech speak; fubarred.

I find it weird that a company making both chipsets and SSD's didn't check this before releasing this flawed shipset to market.

[idata]

I would like you to take a look at this as well and see how the real world performance is affected due to this problem

http://www.storagereview.com/how_improve_low_ssd_performance_intel_series_5_chipset_environments

[idata]

my asus n61ja with intel hm55 chipset has this issue also. Ive 15-25 mb/s r/w. But with processor idle mode disable Ive 29-55 mb/s r/w. We should inform intel about this issue. what can we do?