Overclock.net banner

LinX - Memory Burn Test?

6.9K views 8 replies 6 participants last post by  TranscendentalProtagonist  
#1 ·
While LinX is generally used to test CPU stabiliy, would it be a good/fast/quick way to determine memory stability as well? That is, when the CPU stability is completely eliminated by setting it to stock with a few notches Vcore above the default. Or would you say, stick to MemTest86 (which I would use overnight).

I am redoing my 920 OC and first want to test the uncore and memory for stability/max OC by playing with various BCLK and uncore/mem multies.
 
#3 ·
Yeah I have the bootable CD. I just need something semi reliable for the initial overclock. I will let MemTest86 run overnight, but it's not ideal to do your initial OC, takes so long to even do the first 6 tests of the loop with 6GB lol.
 
#5 ·
HCI Memtest would be the best test for memory stability, particularly on systems with more RAM.
You want to run as many instances of HCI as you have cores (or virtual cores, in the case of i7s) set to utilize nearly all your available RAM.

Memtest86+ is fine for checking for memory errors, but not for memory stability really.
You can pass Memtest86+ with unstable RAM/NB settings that will error in HCI or LinX/IBT max RAM testing.

I suggest HCI & LinX or IBT @ maximum RAM stress (very hard on systems with more RAM as the calculation size is then larger).

Obviously i do run Memtest86+ for initial basic testing, but you'll find it cannot be relied on once you really start pushing things on systems with more RAM.

P95 Large ffts is good for testing motherboard/NB/FSB testing, & okay if you have say < 4 GB RAM or less.
Small ffts is just a CPU test.
 
#6 ·
Quote:

Originally Posted by -n7- View Post
HCI Memtest would be the best test for memory stability, particularly on systems with more RAM.
You want to run as many instances of HCI as you have cores (or virtual cores, in the case of i7s) set to utilize nearly all your available RAM.

Memtest86+ is fine for checking for memory errors, but not for memory stability really.
You can pass Memtest86+ with unstable RAM/NB settings that will error in HCI or LinX/IBT max RAM testing.

I suggest HCI & LinX or IBT @ maximum RAM stress (very hard on systems with more RAM as the calculation size is then larger).

Obviously i do run Memtest86+ for initial basic testing, but you'll find it cannot be relied on once you really start pushing things on systems with more RAM.

P95 Large ffts is good for testing motherboard/NB/FSB testing, & okay if you have say < 4 GB RAM or less.
Small ffts is just a CPU test.
I am running HCI Memtest right now

I am wondering with an i7 920 and HT disabled, so 4 cores, I am supposed to run 4 tests?

I have 6gb of ram

So i am running about 99% of the RAM over the 4 Tests....does this sound correct?
 
#7 ·
Yeah do four instances.

You don't want it to utilize more than you have available for RAM.

I usually aim for about 75%-90% of available memory.

I'd say for your 6 GB i'd be doing something like 4x1200 MB.
 
#8 ·
Another vote for multiple instances of HCI Memtest for windows.

Quote:

Originally Posted by chronicbucks View Post
So i am running about 99% of the RAM over the 4 Tests....does this sound correct?
I'd run a bit less than that, if any of HCIs memory needs to be swapped it will ruin the usefulness of the test.