No memtest failures and why this worries me

[SunshineOutsideRainInside]

First off, my issue is on a laptop Intel 7500U with 16GB Crucial RAM, laptop is 2 years old and issues have been intermittantly getting worse over ~3 months.

I've a java developer, using an IDE called Eclipse. I was (and am) having issues whereby both the programs I write and Eclipse itself (which runs on Java) are crashing with crash dumps showing EXCEPTION_ACCESS_VIOLATION (0xc0000005) - though the actual 'stack' varies quite alot even though the error is always 0xc...05. I tried many things to resolve this including various virus scans, windows updates, java updates, driver updates, Windows RAM test, SSD drive checks and Windows formats. Didn't help, so I assumed hardware problem. Also worth noting that the 'Windows October' update refused to install, repeatedly also showing a 0xc0000005 error - suspiciously the same code as the JVM). RAM is the obvious culprit so I ran memtest86.

Memtest passed after 4 passes over about 4 hours. No errors.

So I booted back into Windows and ran a common CPU load tester, I used 'small FFT' mode (CPU intensive, minimal RAM) for about 2 hours without error. I then ran it in 'Blend' mode (lots of RAM tested) for about 6 hours without error.

This is where I formatted Windows, but upon starting work I had crashes again without a few hours.

At a bit of a loss, I tried a 3rd party Windows based RAM test (not the inbuild Windows RAM test). I expected nothing as its Windows based and thus cannot test as much RAM as memtest86 due to Windows obviously using a fair chunk of it. Within 5 minutes Windows threw a BSOD 'irql_not_less_or_equal'. I rebooted 3 times with each reboot Windows BSOD'd to 'whea uncorrectable error' after about 2 minutes (the CPU Fans whirling to max just prior each time).

I boot BACK into memtest and low and behold everything passes, 4 passes over about 4 hours. I then ran it again and it all passed again.

I format Windows AGAIN and now its booting but I still get the java crashes.

So, is it fair to assume that as memtest86 is refusing to show up any errors, that its more likely a CPU problem, if so, why is the Windows based RAM test BSODing the thing (conceivably a software fault).

Appreciate the help!

[David (PassMark)]

EXCEPTION_ACCESS_VIOLATION (0xc0000005)

This is one of the most common crash causes for software. It means that that the software accessed some RAM addresses that were not allocated to that software. But unfortunately it says nothing about the true cause of the fault. Plain old vanilla software bugs can cause it, as can a multitude of hardware faults.

It it totally possible there is a bug you are provoking in the Java JVM and it isn't a hardware fault.

In MemTest86 were you running in Multithreaded mode? This generates more heat and more faults than single thread testing.

Windows threw a BSOD 'irql_not_less_or_equal'.

Typically this is also a software bug, but in a Windows device driver, not application software. But hardware faults can also cause it. You might be able to look up the details of the blue screen and get the name of the device driver at fault.

So hard to give a conclusive answer. If this was a desktop PC, I would start updating drivers the JVM, then start swapping out parts, but modern laptops are nearly unrepairable.