From someone who ships software that runs on hundreds of millions of computers: please consider running a memory tester once in a while, particularly if you have any unexplained crashes. The days of being able to assume "it's not the hardware, stupid" are past. Bit flips can and do happen. We've made everything small, fast, and cheap. Those are great attributes, but they come with a cost: we've also made things hot and unstable and removed all the safety margins that were in the way.
But also note that a memory test passing doesn't mean the ram (or CPU) is fine.
I overclocked my system for fun and for ram you'd run 2-3 different tests to be able to call it stable with some confidence, as they all have different access patterns and r/w ratios.
Memtest86/+ are usually seen as worse compared to the other tests like say or karhu ram test or testme5, but those also need an OS to run and in the rare chance only some region of memory is faulty could miss it
Testmem5 is in a league of its own yet its unclear how. Maybe the AWE table is prone to corruption. Theres nothing special about the internal workings. It doesn't use fences or clflushes.
Passmark bought MemTest86 after MemTest86+ was forked. It does offer a freeware version but it's limited.
Since the UEFI-capable MemTest86+ came out a few years ago, there have been multiple new releases and it seems to be advancing quite significantly. PassMark seems to have given up trying to keep up.
From someone who ships software that runs on hundreds of millions of computers: please consider running a memory tester once in a while, particularly if you have any unexplained crashes. The days of being able to assume "it's not the hardware, stupid" are past. Bit flips can and do happen. We've made everything small, fast, and cheap. Those are great attributes, but they come with a cost: we've also made things hot and unstable and removed all the safety margins that were in the way.
But also note that a memory test passing doesn't mean the ram (or CPU) is fine.
I overclocked my system for fun and for ram you'd run 2-3 different tests to be able to call it stable with some confidence, as they all have different access patterns and r/w ratios.
Memtest86/+ are usually seen as worse compared to the other tests like say or karhu ram test or testme5, but those also need an OS to run and in the rare chance only some region of memory is faulty could miss it
Testmem5 is in a league of its own yet its unclear how. Maybe the AWE table is prone to corruption. Theres nothing special about the internal workings. It doesn't use fences or clflushes.
Nice to see Memtest86+ being maintained.
I use it on random machines as one of the burn-in tests of hardware. Not only of RAM, but it also lightly exercises CPU, and helps test cooling.
Early on, I used it on a bootable CD-ROM. In recent years, I've been putting it in the Grub boot menu (on systems not afflicted with UEFI).
>I've been putting it in the Grub boot menu (on systems not afflicted with UEFI).
Modern Memtest86+, versions 6 and higher, support systems that boot using UEFI.
Yes, if a machine is afflicted with UEFI, and they no longer permit UEFI to be disabled, you can use that as a workaround.
Not to be confused with MemTest86, a similar program by PassMark.
Quite.
I tried to explain the relationship here, when the previous major release appeared:
https://www.theregister.com/2024/01/11/memtest86_70_released...
Passmark bought MemTest86 after MemTest86+ was forked. It does offer a freeware version but it's limited.
Since the UEFI-capable MemTest86+ came out a few years ago, there have been multiple new releases and it seems to be advancing quite significantly. PassMark seems to have given up trying to keep up.