F5F Stay Refreshed Hardware Desktop Generally, yes—it can be worthwhile if you need more memory for demanding tasks and your system supports the upgrade.

Generally, yes—it can be worthwhile if you need more memory for demanding tasks and your system supports the upgrade.

Generally, yes—it can be worthwhile if you need more memory for demanding tasks and your system supports the upgrade.

Pages (2): Previous 1 2
B
baconman565
Member
207
08-31-2023, 10:19 PM
#11
I think the problem is related to those 24Gb bit microchips being reduced to 16Gb bit versions. I've heard some of their 24Gb chips are cut down to 16Gb and that these also require BIOS updates for compatibility. As for performance, how much faster is a Hi-Nix die compared to Samsung or Micron options in terms of timing? I assume the boards automatically run more efficient subs on the die itself rather than relying on the system chips, which explains why the behavior changes depending on the IC used with my X58A UHD3 where it runs 108 TRFC at 2600 with my Trisil EPIDE STICK but only 224 TRFC with a Samsung 2Gb B Die stick.
B
baconman565
08-31-2023, 10:19 PM #11

I think the problem is related to those 24Gb bit microchips being reduced to 16Gb bit versions. I've heard some of their 24Gb chips are cut down to 16Gb and that these also require BIOS updates for compatibility. As for performance, how much faster is a Hi-Nix die compared to Samsung or Micron options in terms of timing? I assume the boards automatically run more efficient subs on the die itself rather than relying on the system chips, which explains why the behavior changes depending on the IC used with my X58A UHD3 where it runs 108 TRFC at 2600 with my Trisil EPIDE STICK but only 224 TRFC with a Samsung 2Gb B Die stick.

W
WZ_Galaktiik
Senior Member
251
09-01-2023, 12:06 PM
#12
Both the 24Gb and 16Gb models share identical problems. They use the same chip, meaning training needs are comparable. It's noticeably improved over Samsung options, while Micron performs poorly. The performance can vary by board, but I haven't found any that run without manually enabling an "optimize sub-timings" feature. By default, all boards follow JEDEC timings if XMP is turned on, except for certain MSI models where tREFI behavior changes unless XMP isn't active.
W
WZ_Galaktiik
09-01-2023, 12:06 PM #12

Both the 24Gb and 16Gb models share identical problems. They use the same chip, meaning training needs are comparable. It's noticeably improved over Samsung options, while Micron performs poorly. The performance can vary by board, but I haven't found any that run without manually enabling an "optimize sub-timings" feature. By default, all boards follow JEDEC timings if XMP is turned on, except for certain MSI models where tREFI behavior changes unless XMP isn't active.

Pages (2): Previous 1 2