Did you ever try running it for a long time instead?Originally Posted by CNCoRIDDoR
Yeah I did it lets say about 2-3 hours on 2000 9 9 9 24 2t 1.5v. Which is weird right? No errors during that time.
Maxxmem as a stability test? I've found that if I can boot into windows, I can run Maxxmem with no issues even if I get errors the second I hit start on any other program. Hyperpi is good for initial testing, but I'd still do a long P95 run after Hyperpi has finished because 10-15 minutes of stressing the memory isn't particularly thorough.Originally Posted by Jahocowi
Memtest 86+ is primarily used for seeing if you have bad sticks. If you want to test a RAM OC try programs like hyper pi and MaxxMEM2. You can also try out Prime95 and IBT/LINX, but the other two are more memory intensive.
Maxxmem as a quick speed test. Overclocking RAM and timings can be extremely time consuming. Would you run Prime95 for 4 hours after lowering TRc one digit from stock? No,you tweak with maxxmem to get highest stable speed and lowest timings,then run your Prime/memtest or whatever and adjust from there.Originally Posted by Vonnis
If you don't know how to use it properly, you probably didn't use it right. Assuming you're using the free version, you need to open one instance of the program for every CPU core you have. Then take your available RAM (or 90% of all RAM), divide that by the number of CPU cores, and enter that number in each instance of HCI memtest. Let that run up to 1000%, if that produces no errors move on to another program (or just call it a day if you just need basic stability for gaming).
Maxxmem as a stability test? I've found that if I can boot into windows, I can run Maxxmem with no issues even if I get errors the second I hit start on any other program. Hyperpi is good for initial testing, but I'd still do a long P95 run after Hyperpi has finished because 10-15 minutes of stressing the memory isn't particularly thorough.
Nha, I use Hyperpi 32M for quick testing. I've never seen Maxxmem have issues with an overclock, even if it turns out to be unstable as hell when I use any other program, so although I do run it every time I've tweaked something, I never even considered it as anything but a speed test.Originally Posted by Redwoodz
Maxxmem as a quick speed test. Overclocking RAM and timings can be extremely time consuming. Would you run Prime95 for 4 hours after lowering TRc one digit from stock? No,you tweak with maxxmem to get highest stable speed and lowest timings,then run your Prime/memtest or whatever and adjust from there.
I usually can tell if RAM is somewhat stable just by the way Maxxmem runs. If you are close to the edge of benching stability,it will bug or hang-up,or your scores will drop. Another good one to use is SiSoft's Sandra bandwidth and latency tests.
Not only that, but it's limited to around 2-4 GB per instance of the program, so it wouldn't do much good any way.Originally Posted by Imprezzion
This way the RAM will be filled up 10)% and tested fully and it'll go nice and fast since all threads can work on it. HCI is single threaded so using all RAM on 1 HCI is slow as hell.
