RAID1 Glacially Slow on Aurora Desktop

I'm running an Aurora Desktop with an SSD boot volume for the Win7-x64 OS, and a RAID1 volume for data using two identical drives purchased from Dell/Alienware with the machine, and configured at the factory for RAID.  Recently, performance of the machine has positively tanked.  I've now tracked this issue back to VERY slow access times and high latency on my RAID1 array. 

When I run the microsoft SQLIO utility, I get disk write throughput numbers that are abysmal:

C:\Program Files (x86)\SQLIO>sqlio -kW -s60 -frandom -o256 -b4 -LS -Fparam.txt
sqlio v1.5.SG
using system counter for latency timings, 3132832 counts per second
parameter file used: param.txt
        file d:\testfile.dat with 2 threads (0-1) using mask 0x0 (0)
2 threads writing for 60 secs to file d:\testfile.dat
        using 4KB random IOs
        enabling multiple I/Os per thread with 256 outstanding
size of file d:\testfile.dat needs to be: 104857600 bytes
current file size:      0 bytes
need to expand by:      104857600 bytes
expanding d:\testfile.dat ... done.
using specified size: 100 MB for file: d:\testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec:   254.93
MBs/sec:     0.99
latency metrics:
Min_Latency(ms): 2
Avg_Latency(ms): 1974
Max_Latency(ms): 5846
histogram:
ms: 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%:  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 99

Even the read test is horrible:

C:\Program Files (x86)\SQLIO>sqlio -kR -s60 -frandom -o256 -b4 -LS -Fparam.txt
sqlio v1.5.SG
using system counter for latency timings, 3132832 counts per second
parameter file used: param.txt
        file d:\testfile.dat with 2 threads (0-1) using mask 0x0 (0)
2 threads reading for 60 secs from file d:\testfile.dat
        using 4KB random IOs
        enabling multiple I/Os per thread with 256 outstanding
using specified size: 100 MB for file: d:\testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec:  1049.05
MBs/sec:     4.09
latency metrics:
Min_Latency(ms): 0
Avg_Latency(ms): 485
Max_Latency(ms): 771
histogram:
ms: 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%: 12  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 88

I am at my wits end trying to troubleshoot this.  There is no way that a latency time on disk writes of over 5s can be normal.  I've run the Intel RST Verification test, and it comes back clean.  Currently the system is not failing per se, but I am worried that the slow performance is a harbinger of things to come.

Interestingly, even the SSD isn't stellar in the area of disk performance:

C:\Program Files (x86)\SQLIO>sqlio -kR -s60 -frandom -o256 -b4 -LS -Fparam.txt
sqlio v1.5.SG
using system counter for latency timings, 3132832 counts per second
parameter file used: param.txt
        file c:\testfile.dat with 2 threads (0-1) using mask 0x0 (0)
2 threads reading for 60 secs from file c:\testfile.dat
        using 4KB random IOs
        enabling multiple I/Os per thread with 256 outstanding
using specified size: 100 MB for file: c:\testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec:  6921.06
MBs/sec:    27.03
latency metrics:
Min_Latency(ms): 1
Avg_Latency(ms): 73
Max_Latency(ms): 84
histogram:
ms: 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%:  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 100

This makes me wonder if there is a problem with the entire SATA controller itself.  When I run the test to my RAMDisk, I get GREAT performance, as expected. This tells me that the OS itself is not likely part of the problem:

C:\Program Files (x86)\SQLIO>sqlio -kR -s60 -frandom -o256 -b4 -LS -Fparam.txt
sqlio v1.5.SG
using system counter for latency timings, 3132832 counts per second
parameter file used: param.txt
        file r:\testfile.dat with 2 threads (0-1) using mask 0x0 (0)
2 threads reading for 60 secs from file r:\testfile.dat
        using 4KB random IOs
        enabling multiple I/Os per thread with 256 outstanding
using specified size: 100 MB for file: r:\testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec: 254852.81
MBs/sec:   995.51
latency metrics:
Min_Latency(ms): 0
Avg_Latency(ms): 1
Max_Latency(ms): 9
histogram:
ms: 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%:  0 66 34  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0

I've tried calling Alienware tech support 4 times today.  The first time, I got a guy who put me on hold for over 35 minutes, then the line was disconnected.  When I called back, I got the SAME guy.  Immediately after giving him my Service Tag, he hung up on me immediately.  I called a third time, and got a girl this time.  She put me on hold and I was sitting for about 3 minutes before the line got disconnected again.  The 4th time I got a guy who finally spoke with me and directed me to the wrong support group, who then tried to direct me to the right one, but they were all out for the holiday and I should try back on Tuesday.  Could have saved me 2 hours of phone time if the FIRST guy could have told me that.

Anyway, does anyone have any other suggestions for troubleshooting performance on an Aurora with RAID array?  Can someone else with a similarly configured Aurora desktop actually run the SQLIO tests with the paramaters supplied above and let me know what YOU get for these tests?  Clearly, 27MB/s is under-performing on the SATA drive, and 0.9MB/s is clearly horrible.  However, I just need some appropriate "next steps" in troubleshooting this. 

Thanks in advance.