[SOLVED] what is the purpose of four (4) RAM slots in motherboards with daisy-chain routing?

[dabidoy]

aside from overclocking stability, what is the purpose of daisy-chaining in motherboards with four (4) RAM slots? is it cheaper to make the traces that way? or more RAM is still better in most use cases rather than stability?

newbie in PC building here. thanks

 

[SkyNetRising]

what is the purpose of daisy-chaining in motherboards with four (4) RAM slots? is it cheaper to make the traces that way? or more RAM is still better in most use cases rather than stability?

Why does that matter to you?
Generally it isn't important.

 

[dabidoy]

if i want to OC in the future? if it will affect may cost-benefit analysis when buying boards?

newbie in PC building here. thanks

 

[kanewolf]

aside from overclocking stability, what is the purpose of daisy-chaining in motherboards with four (4) RAM slots? is it cheaper to make the traces that way? or more RAM is still better in most use cases rather than stability?

newbie in PC building here. thanks

Capacity over performance. Why are there pickup trucks and sports cars sold. Capacity over performance.

 

[hotaru.hino]

T-topology may be slightly more expensive to design and implement, but at this point, it doesn't really matter. And in general, one topology isn't necessarily better than the other, it depends on how the board was laid out:

In any case, overclocking memory isn't really useful for performance in most cases. Especially if the memory controller has to start doing funky ratios that aren't 1:1 or close to it. For example, going beyond DDR4-3733 for Ryzen isn't really useful because beyond that point, the IMC goes into a 2:1 ratio and now there's extra latency when accessing RAM. I forget the exact number, but you have to go to something like DDR4-4800 before the latency goes back down to where it was at DDR4-3733.

And for a lot of use cases, memory access tends to be random and in small chunks where latency is the primary spec to worry about. The only two cases I've seen (but not the only ones) where increased memory speed consistently correlates to increased performance is when it's used in an iGPU setup or for file archiving (7zip, RAR, etc).

 

[dabidoy]

T-topology may be slightly more expensive to design and implement, but at this point, it doesn't really matter. And in general, one topology isn't necessarily better than the other, it depends on how the board was laid out:

In any case, overclocking memory isn't really useful for performance in most cases. Especially if the memory controller has to start doing funky ratios that aren't 1:1 or close to it. For example, going beyond DDR4-3733 for Ryzen isn't really useful because beyond that point, the IMC goes into a 2:1 ratio and now there's extra latency when accessing RAM. I forget the exact number, but you have to go to something like DDR4-4800 before the latency goes back down to where it was at DDR4-3733.

And for a lot of use cases, memory access tends to be random and in small chunks where latency is the primary spec to worry about. The only two cases I've seen (but not the only ones) where increased memory speed consistently correlates to increased performance is when it's used in an iGPU setup or for file archiving (7zip, RAR, etc).

wow thanks so much!