Is there a big difference between a 2666mhz and 3600mhz?

[Stealth3si]

I know higher speed number is always better but I want to know how much of a noticeable difference is there between 2666mhz ram and 3600mhz ram on an i9 9900k and Gigabyte Z390 Aorus Ultra. I will be gaming and general productivity. FYI, the difference in price between the two "identical" sets I'm looking at getting is $150. One set is 18-18-18-43 1.20V while the other set is 18-22-22-42 1.35V. How much of a difference in performance speed would I be missing out in games and general productivity if I went with the cheaper set? Thank you for your help!

[BiotechBen]

11 minutes ago, Stealth3si said:

I know higher speed number is always better but I want to know how much of a noticeable difference is there between 2666mhz ram and 3600mhz ram on an i9 9900k and Gigabyte Z390 Aorus Ultra. I will be gaming and general productivity. FYI, the difference in price between the two "identical" sets I'm looking at getting is $150. One set is 18-18-18-43 1.20V while the other set is 18-22-22-42 1.35V. How much of a difference in performance speed would I be missing out in games and general productivity if I went with the cheaper set? Thank you for your help!

The 18-18-18-43 set is going to give you slightly better performance, unlikely that you would see much of a difference outside of benchmarks 

[Bad5ector]

If you were talking a Ryzen system, yeah there would be a measurable difference. But I doubt it with your 9900k, outside of benchmarks like @BiotechBenmentioned. 

[Stealth3si]

4 minutes ago, BiotechBen said:

The 18-18-18-43 set is going to give you slightly better performance, unlikely that you would see much of a difference outside of benchmarks 

mb forgot to mention, the 18-18-18-43 set is the 2666mhz kit, would that change your answer?

[BiotechBen]

6 minutes ago, Stealth3si said:

mb forgot to mention, the 18-18-18-43 set is the 2666mhz kit, would that change your answer?

Cl43 on a 2666 kit?! That sounds like Sodimm laptop memory

 

And yes, that dramatically changes the answer. 

The clocks per latency of the 2666 is vastly lower than that of the 3600, but that outcome is also heavily dependent on if your 9900K and mobo can run the 3600 kit, no guarantees.

Both of those kits are terrible kits given the clocks per latency

2666 kit: 26

3600 kit: 84 

 

A half decent 3200 kit: 200

[Stealth3si]

10 minutes ago, BiotechBen said:

Cl43 on a 2666 kit?! That sounds like Sodimm laptop memory

 

And yes, that dramatically changes the answer. 

The clocks per latency of the 2666 is vastly lower than that of the 3600, but that outcome is also heavily dependent on if your 9900K and mobo can run the 3600 kit, no guarantees.

Both of those kits are terrible kits given the clocks per latency

2666 kit: 26

3600 kit: 84 

 

A half decent 3200 kit: 200

I see. Where do you get those numbers and what do they mean? The 26, 84 and 200

[svmlegacy]

10 minutes ago, BiotechBen said:

Both of those kits are terrible kits given the clocks per latency

2666 kit: 26

3600 kit: 84 

Latency doesn't work like that... I'm much more worried about the actual cas latency of the 2666 kit: 13.5 ns, while the 3600 is 10ns. 35% is massive.

[BiotechBen]

6 minutes ago, svmlegacy said:

Latency doesn't work like that... I'm much more worried about the actual cas latency of the 2666 kit: 13.5 ns, while the 3600 is 10ns. 35% is massive.

Agreed, the latencies on the 2666 kit are awful. 

 

Oh dear, I just looked at what I wrote, did I have a momentary stroke?

[BiotechBen]

11 minutes ago, Stealth3si said:

I see. Where do you get those numbers and what do they mean? The 26, 84 and 200

Disregard, momentary brain shutoff.

[Stealth3si]

5 hours ago, BiotechBen said:

Disregard, momentary brain shutoff.

So ignore the entire post and just get the 2666mhz kit?

[BiotechBen]

1 hour ago, Stealth3si said:

So ignore the entire post and just get the 2666mhz kit?

You are much more likely to have the 2666 kit run without issues vs the 3600 kit on your 9900K.

[xg32]

14 hours ago, Stealth3si said:

.

it's been awhile since i did the test on hitman 2, but with 5.1 all core, the difference from 3200 to 3600 cl14 was 138 to 141 fps, going to 4000 unstable yielded 144fps, i guess going from 3200 to 2666 would be about 130-132fps average with lower dips vs 138 avg at 3600, maybe 5, or 10% at most. That's only one game and it just happened to be memory bottlenecked

[Stealth3si]

update report so I got the 2666 set and overclocked it to 3067mhz at 16 18 18 38 stable

 

I tried out a set of owl oloy 3200 16 20 20 38 and it was unstable at advertised settings