F5F Stay Refreshed Hardware Desktop Comparing DDR5 6400MHz C38 and DDR5 6600MHz C32 models highlights differences in speed and compatibility.

Comparing DDR5 6400MHz C38 and DDR5 6600MHz C32 models highlights differences in speed and compatibility.

Comparing DDR5 6400MHz C38 and DDR5 6600MHz C32 models highlights differences in speed and compatibility.

V
vincentnele
Member
223
08-09-2021, 12:37 AM
#1
The two options differ mainly in speed and compatibility. The 6400MHz model runs faster, while the 6600MHz version supports higher memory capacities. Differences in performance and future-proofing may exist between them.
V
vincentnele
08-09-2021, 12:37 AM #1

The two options differ mainly in speed and compatibility. The 6400MHz model runs faster, while the 6600MHz version supports higher memory capacities. Differences in performance and future-proofing may exist between them.

A
augustb19907
Senior Member
456
08-09-2021, 08:22 PM
#2
Depends heavily on CPU and motherboard capabilities. The 6600 C32 might be faster in theory, but consider first word latency. A C32 at 28ns versus a C38 at 12ns could make the latter better for everyday use.
A
augustb19907
08-09-2021, 08:22 PM #2

Depends heavily on CPU and motherboard capabilities. The 6600 C32 might be faster in theory, but consider first word latency. A C32 at 28ns versus a C38 at 12ns could make the latter better for everyday use.

M
Marcustheduke
Senior Member
679
08-31-2021, 08:01 AM
#3
If you're only going to turn on XMP, expect little change in performance. Getting the 6600 MT/s CL32 kit should work for this purpose. For RAM overclocking, 6400 CL38 and 6600 CL32 are essentially the same memory chip (Hynix M die) and can handle each other's settings. The M die is fairly reliable, even OEM units can reach 6800MT/s CL34 when unlocked, with differences mainly in timing, maximum speed, and tRCD/tRP values—those timing details don't affect XMP profiles. Maximum speeds typically sit between 6800 and 7000MT/s, so both kits should synchronize properly. The extra tick on tRCD won’t stand out unless you're using SuperPi 32M. In the end, choose the more affordable option; the 6400MT/s kit is usually the best value. It’s unlikely to be worth chasing a kit faster than 5600MT/s CL40, though that’s also Hynix-only if you plan to overclock. At XMP, any real difference should be around 5% at most, not measured by specific benchmarks.
M
Marcustheduke
08-31-2021, 08:01 AM #3

If you're only going to turn on XMP, expect little change in performance. Getting the 6600 MT/s CL32 kit should work for this purpose. For RAM overclocking, 6400 CL38 and 6600 CL32 are essentially the same memory chip (Hynix M die) and can handle each other's settings. The M die is fairly reliable, even OEM units can reach 6800MT/s CL34 when unlocked, with differences mainly in timing, maximum speed, and tRCD/tRP values—those timing details don't affect XMP profiles. Maximum speeds typically sit between 6800 and 7000MT/s, so both kits should synchronize properly. The extra tick on tRCD won’t stand out unless you're using SuperPi 32M. In the end, choose the more affordable option; the 6400MT/s kit is usually the best value. It’s unlikely to be worth chasing a kit faster than 5600MT/s CL40, though that’s also Hynix-only if you plan to overclock. At XMP, any real difference should be around 5% at most, not measured by specific benchmarks.