1333MHz Vs 1600/1866MHz DDR3 Speeds (Gaming)?

[Chaos Water]

I currently have 16GB (2x8GB) of Kingston HyperX Red 1333MHz RAM installed and was wondering if there was a difference in terms of performance between 1333MHz and 1600MHz.

Would it be worth upgrading to 1600MHz or even 1866MHz?

Games I mainly play are GTA V, Battlefield 1 & 4 etc..

I'm also running a GTX 780 STRIX 6GB (eventually upgrading to GTX 1070) and i7 4790 (non-K due to H81M PLUS MoBo)

 

[Sakkura]

There is a difference, but it's not huge. The maximum memory speed your motherboard supports is 1600, so faster memory would just be limited to that anyway.

 

[faalin]

For gaming timings matter more then speed, this is why people can get way with 1333Mhz CL7 vs 1600mhz CL9. If you doing autocad, video rendering, or things of that nature then raw speed is better then timings.

 

[Sakkura]

For gaming timings matter more then speed, this is why people can get way with 1333Mhz CL7 vs 1600mhz CL9. If you doing autocad, video rendering, or things of that nature then raw speed is better then timings.

Nah, clock frequency matters more than timings. Because actual latency is based on both timings and clock frequency, while bandwidth is only based on clock frequency.

 

[faalin]

when comparing latencies at different clock speeds, latencies must be translated into absolute times to make a fair comparison

If you take a 1333mhz cl7 and compare it to a 1600mhz cl9 the 1333mhz cl7 is just a hair faster

when reading the first, fourth, and eight word

1333mhz 10.50ns, 12.75ns, 15.75ns
1600mhz 11.25ns, 13.13ns 15.63ns

You would then have to drop down to a 1600mhz cl8 to be faster then a the 1333mhz cl7. When gaming is short random calculations the lowest CL# is better.

When doing long calculation like autocad or rendering videos you what faster MHz to chew threw the data faster.

 

[Sakkura]

Benchmarks show clocks matter more than timings for gaming.

http://www.anandtech.com/show/7364/memory-scaling-on-haswell/8

 

[faalin]

if you look at a single gpu card, not the triple card page that you linked (who runs triple cards), they are within 1 fps. The best 1600 cl9 could do is .65 fps in bioshock

Your theory of faster Mhz is better doesn't hold water ether since I run 2133 cl9 (which I paid out the nose for) on my computer and all 3 are within 1fps of each other in the tests. also look between 1333 cl7 vs 3000 cl12 other then bioshock those two are between .5 fps in the other 3 games.

1333mhz ram was also cheaper back in the ddr3 days then any other ram as well. Now they are all about the same price, companies just want to dump stock to move people to ddr4

 

[Sakkura]

There's no point looking at the single GPU page where all the games are GPU bottlenecked. The point of the triple-Xfire page is to shift the bottleneck to CPU/memory to better illustrate memory scaling. Where we see frequency matters more than timings.