can high frequency memory, boost CPU performance(if GPU is much more powerful done the CPU)

[vgcs1e2]

ive seen lots of benchmark/trials of HIGH vs LOW FREQUENCIES MEMORY and there is no difference,

BUT IT IS TESTED AT HIGH TIER PROCESSORS(like i7),

how about on the LOW TIER processor(like pentium or i3) where you can see bottlenecking when paired with powerful gpu, will it boost performance?? (like FPS in game)??

 

[genthug]

High frequency memory will only boost performance if the data transfer rate from your RAM to your CPU is causing your CPU to have to wait for that data transfer to complete. For lower clocked processors, this is less of an issue as they will, because they are lower clocked, be asking for information from your memory less frequently. For the moment, often the fastest piece of hardware in the main portion of the computer (not including the GPU) is your RAM, especially where DDR4 is now concerned.

So, TL;DR it really depends on what the specs of what you're looking at are. By and large, it will have pretty minimal impact on what you're running unless you specifically underclock your RAM to cause a data transfer rate bottleneck.

 

[synphul]

If it makes little difference with an i7 then it won't make any difference with a slower cpu like an i5 or i3. Ram speed in terms of memory bus speed or the speed of the ram itself is how fast data moves from ram to the cpu's cache. It won't make the cpu faster, the cpu will only process data as fast as it can. If there's a problem with data being bottlenecked by slow memory it will show up on an i7 same as any lower tier cpu. A better cpu is needed for improved fps in the game (provided it's not a gpu related issue), there's no super fast ram that's a magic fix.

 

[BFG-9000]

High frequency memory mostly helps an IGP like in an APU, which can always use the bandwidth.

Note that even though bandwidth has greatly increased, latency really hasn't changed much since 2-2-2 DDR1-400 15 years ago because as the frequency goes up, timings have got very lax and is now closer to 17-17-17. An analogous situation is RAID0 with HDD (which greatly increases only bandwidth) vs. SSD (which greatly decreases latency). I'm sure most people would prefer using a SSD on SATA-300 or even SATA-150 over a 4-HDD RAID0 setup.

Note that at 3.6GHz processor speed, each nanosecond of latency is equivalent to 3.6 processor clock cycles. So if you test the memory latency and find it is 55ns for example, the CPU wastes 200 clock cycles every time it has to wait on memory (you can look at it like the CPU only runs 1/200 as fast during this time because the RAM can't keep it fed).

Only the cache in the CPU runs at full speed, so more cache at faster CPU speeds are what matters for games, as well as the GPU drivers.