Driver Verifier BSOD's

Event 41's simply mean that the system is shutting down - and Windows doesn't know why.
This means (IMO) that 3rd party programs and Windows stuff is ruled out - leaving hardware as the cause.

Prime95 failing on the Blend test indicates (to me) that it's a memory problem. Either RAM, CPU cache, or the memory controller. That's why I requested that you run the small FFT's, large FFT's, and Blend test. Running them together keeps the consistency of error reporting more constant (which is essential if we're trying to decide what's at fault).

running small fft's now :) (just a note - 3 different monitor programs give me 3 different temperature readings - core temp,cpuid and everest ultimate - pic attached)
btw. since prime95 allocates a certain amount of ram for its use -leaving maybe 5-10% free ram,could it be possible that if an application requests ram at the same time,a system crash would occur since there is not enough free memory? or it would just slow down until the memory is available?

and also, i am running the ram at 1 speed down - the sticks are pc2-8500 and i am running them at 400mhz-pc2-6400(only one stick at the moment) and at more relaxed timings(5-5-5-18-53) instead of (5-5-5-15-43).
last night,memtest86 concluded 5 passes of this particular stick with no errors.
however if the problem is at the memory controller,and not at the actual ram-stick,would the memtest give out an error?

Memory is a complicated issue but it's not likely that the Windows Memory Manager would allow any program to take too much RAM (although a corrupted program could conceivably do this indirectly). IME the system just slows to a crawl until Prime95 is stopped by the user.

I haven't messed around with RAM timings/settings/voltages since my overclocking days (back around 2000/2001). So I'm not familiar with the specs/performance of modern day RAM. That being said, Memtest is better at testing RAM than Prime95 - but Prime95 is better at stressing RAM than Memtest is. A small difference, but a significant one none the less. It's possible (and happens fairly frequently) that RAM will pass Memtest, but will fail on Prime95.

How soon until the crash? Faster or slower than in Blend?

FYI - your memory controller is contained in the Northbridge on your motherboard.
This means that failing the large FFT's and the Blend test means that it's most likely RAM, but can be the motherboard.

The "test" here is to test the memory sticks individually with Prime 95 in both Blend and Large FFT's - to see how long it takes to error out with each stick (and also trying them in different slots).

blend test failed in 20 mins(much faster than large fft's which crashed after 3 hours or so) - with the following bsod: an attempt was made to write to read only memory
attached is the dump file of the blend test bsod.
also,i attach a dump of a bsod which happened afterwards while watching a movie(using Splayer)

ok i will try each stick to all slots - then move to the next stick

With Blend failing faster it's most likely that one or more of the sticks is at fault (but the slots can be defective also and can give similar results - that's why I asked that you rotate them through the tests also).

What we need here is a stick/slot combo that doesn't spit errors in Prime95. That will confirm that the other, error causing sticks/slots are to blame.

If you don't get a distinct problem stick/slot, then it's uncertain if this is a RAM or mobo problem. As Blend fails faster I'd have to guess RAM, but it's not my wallet that's involved in the next step.

Check to see if you can return your RAM under warranty to OCZ and get a new set. That'll help to rule out the RAM with the least cost.

Good luck!