You can try 2 instances of IBT, manually adjusting the memory amount.Quote:
Originally Posted by Leeghoofd
Anyway, I am deeply convinced from my tests that IBT is very poor to find NB/RAM instability.
Printable View
You can try 2 instances of IBT, manually adjusting the memory amount.Quote:
Originally Posted by Leeghoofd
Anyway, I am deeply convinced from my tests that IBT is very poor to find NB/RAM instability.
what is actually THE BEST program to test RAM with in windows, i can barely think prime blend do that so well as it really could be?
Think that will be HCI memtest Infa...
Memtest 86+ v2.01 (not in Windows) if you are truly looking to isolate RAM
Why are you looking to test in windows?
Quote:
Originally Posted by Razmatazz
Because you run your rig in Windows, most of the time, no?:shrug:
I use all three tests.
Razmatazz
Windows MemTest 3.7 is much more harder for RAM than the DOS one.
Quote:
Originally Posted by zlojack
Anything that test your ram in windows cannot stress your ram in it's entirety.Quote:
Originally Posted by WaterFlex
Some of that ram has been isolated for use running windows.
The ram used to keep windows operational cannot be tested by the system, so windows based stress tests provide incomplete results.
Memtest 86+ remains the best, and only complete method of testing Ram stability.
Hence why I use all three tests, plus Prime95 and OCCT and others :D
ThaLoveDocta
I used to pass MemTest86+ with NO ERRORS during the night but I had errors in Win MemTest 3.7. So, I had to encrease Vdimm a little, like +0,07.
Not stability but integrity.Quote:
Originally Posted by ThaLoveDocta
Under DOS the memories are tested by memtest with only one thread (DOS limitation) and it is not the normal Windows environment.
Normally I can reach at least 10% more MHz in memtest than is Windows.
IMHO the best tests for memory stability in Windows are OCCT RAM test, S&M ram test and memtest for Windows.
Brama
You are absolutely right, mate. But MemTest 3.7 is the best RAM killer :D. I use it always.
thanx for the programm.
IntelBurnTest was made for testing the CPU, and motherboard chipset primarily...
Quote:
Originally Posted by WaterFlex
Quote:
Originally Posted by Brama
Quote:
Originally Posted by WaterFlex
Very interesting to hear that fron another user, i've never received these results from any of my setups.
appreciate the input, will double test in the future on rigs which require absolute stability. :)
Any ideas why this is?
Last i checked, Mem test ran a series of stacked, random and sequential test which should cover all ram nodes. why is windows returning errors?
Intersting. How long do you run it?Quote:
Originally Posted by WaterFlex
300% here, once i runned 1600% by mistake when i forgot my keyes inside and locked myself out of my house :DQuote:
Originally Posted by davidk21770
Got 4 instance running then, one problem in 400% others have runned to 1600% :ROTF:
Updated to v1.8!
- Updated Linpack binaries to October 2008, and patched for AMD compatibility
- Fixed crash triggered by using more than 2 GB of memory in 32-bit Windows
- Fixed a bug triggered by using more than the available memory in the system
- Added automatic compensation for memory selection for 32-bit Windows
- Added smart-allocation to disallow allocating more memory than possible, and
set it to the maximum possible when needed
- Added smarter error detection algorithm
Still getting v1.7, did you update the X64 link?
Updated 64-bit Linpack to v1.8
- Same updates (except those for 32-bit) as the hybrid executable
The x64 executable is really for WinPE. For anything else, use the main link (hybrid executable). But I just updated the native 64-bit one too (which again, is for WinPE).Quote:
Originally Posted by stealthbomber
Cheers running it now!
Good luck! Let us know how it went :D
so whats good Xtimes to run for 24/7 settings
10 to 20 would be good, using at least 3 GB of memory for testing (from my own experience).
well i only got 2GB of ram
Well then you want to use 20 runs :DQuote:
Originally Posted by Dom7184