Since i got no reply, here's some 3D diagrams showing off the random read performance of SteveRo's Frankenstien RAID of 4x25-V, some are flashed Kingston V 40GB (therefore frankenstein).
Both sets are 4R0 x25-V ICH10R 16KB stripe, WBC off. First set is stock clocks, second is 119 PCIe and 4,8Ghz HT off on 980x. The OC makes a difference.
Comment: You'll notice 3-4 drops in performance around 8-16KB, this can happen some times, and is the reason this is only a "fairly accurate" representation of performance. For more smooth graphs, i'd use 5 data sets, discard the best and worst numbers for each datapoint, and take the average of the 3 in the middle. This however is a bit too much work for this type of "overlook".
So, here we go:
No OC:
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS no oc WBC off xBSyIOPSzQD.png 
Views:	321 
Size:	130.5 KB 
ID:	103789
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS no oc WBC off xBSyBandwidthzQD.png 
Views:	317 
Size:	127.1 KB 
ID:	103790
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS no oc WBC off xBSyAvgAcczQD.png 
Views:	322 
Size:	122.7 KB 
ID:	103791
OC:
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS OC pcie119 WBC off xBSyIOPSzQD.png 
Views:	321 
Size:	128.4 KB 
ID:	103792
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS OC pcie119 WBC off xBSyBandwidthzQD.png 
Views:	322 
Size:	128.0 KB 
ID:	103793
Click image for larger version. 

Name:	SteveRo 4R0 x25-V ICH10R 16KB SS OC pcie119 WBC off xBSyAvgAcczQD.png 
Views:	322 
Size:	123.0 KB 
ID:	103794

I also have graphs comparing OC to stock, as a spoiler, the OC has a 10-25% advantage at smaller blocks, and 5-15% at larger blocks. The results are consistent, but there is a drop for OC at 16KB QD 8-16 (it happens), but this likely represents an anomaly and not the actual performance at those points.

EDIT: i saw Steve's reply too late. If a moderator sees this, pleese move this post.
For now, i'll post a link there to this post.