1866 CL10 Slower than 1600 CL9? :: Hardware and Operating Systems

All Discussions > Steam Forums > Hardware and Operating Systems > Topic Details

[☥] - CJ - 4 Aug @ 5:32am

1866 CL10 Slower than 1600 CL9?

So thats what i see when i do a VS search of the 2. I recently upgraded to Viper 32GB 1866 CL10 from Corsair Vengeance 16GB 1600 CL9

Timings shown here - https://valid.x86.fr/ums7b8

So if its slower simply because of it being CL10, how would i make 1866 be faster than 1600?

The author of this topic has marked a post as the answer to their question.
Click here to jump to that post.

Originally posted by Illusion of Progress:

I think you may be overthinking some things.

The fact that all RAM in a configuration isn't exact in whatever frequency/timings it "maxes at" doesn't matter in an of itself, as long as it's stable. If it's stable, then the only thing that matters is what frequency/timings they are able to run at together.

The difference between 1,600 MHz CL9 and 1,866 MHz CL10 is so close to nil because it's... just not that large of spread. The saying "measurable, not noticeable" comes to mind. The end result will often be margin of error. Older platforms are rarely going to see cost effective performance returns from upgrades (not unless you're sourcing from the used market at "near free" prices anyway).

But since you already have it now, I see no point in changing it because 1,866 MHz CL10 would already be preferable to 1,600 MHz CL9, even if it's not by a lot.

< >

Showing 1-15 of 15 comments

Rumpelcrutchskin 4 Aug @ 5:41am

They are pretty much same, doesn`t really matter if you use one or the other.

hawkeye 4 Aug @ 5:42am

Comparing memory requires checking the exact model id's.
1866 CL10 doesn't mean that much.
Benchmark sites, reviews etc. For example -
https://www.memorybenchmark.net/
The nominal speed and CL values mean little by themselves.

[☥] - CJ - 4 Aug @ 6:03am

The Corsair 1600 i had was 9-9-9-24
The Viper is 10-11-10-30

Rumpelcrutchskin 4 Aug @ 6:10am

Performance difference is so minimal that you won`t be able to tell the difference anyway.

[☥] - CJ - 4 Aug @ 6:11am

Mkay..

_I_ 4 Aug @ 7:58am

1866 / 10 = 186.6
1600 / 9 = 177.7

at 1866 cl10 should be slightly faster, 5% or so
i doubt it would show in many benches if its cpu limited

Illusion of Progress 4 Aug @ 8:57am

It's slightly worse in latency but slightly higher in bandwidth to the point that it more than compensates. Which one is faster will vary, but either way it's small enough of a difference that's it's the "same thing".

Generally, latency matters more for gaming than bandwidth, but DDR3 bandwidths are so low nowadays that the extra bandwidth (if slight) might come out ahead more often than usual.

Overall it is so small of a difference that it is a side grade, but that is to be expected because 1,600 MHz CL9 was already a pretty performant spot to be, and you're probably not going to find new performance DDR3 this late. I imagine you upgraded for the higher capacity, not any added platform performance, and it did give you more capacity without being seriously slower, so it seems like it fulfilled it's purpose.

[☥] - CJ - 4 Aug @ 9:28am

I have 32GB of the Corsair available to me, but the specs and timings are not the same for whatever reason as i bought the kits at different times so that always bugged me, so i just stuck with the newest 2x8GB kit.

I went for the 1866 because of the higher speed which even a little bit of a boost is enough for me, and due to the cheap price i bought another kit for a total of 32GB since oddly enough they are recently manufactured kits with all the same specs.

So with 1600 CL9 SUPPOSEDLY offering better performance, i was wondering if i could tighten the timings or something to lessen the gap between the 2.

But i guess ill just keep it as is.

But yeah, 16GB was beginning to not be enough for specific uses so i figured id go for 32 while also being a tiny bit faster.

Last edited by [☥] - CJ -; 4 Aug @ 9:28am

A&A 4 Aug @ 10:03am

The Viper might be slightly faster but it's close to the margin of error.

[☥] - CJ - 4 Aug @ 10:23am

Should i try overclocking it to 2133?

#10

The author of this thread has indicated that this post answers the original topic.

Illusion of Progress 4 Aug @ 10:51am

#11

[☥] - CJ - 4 Aug @ 10:53am

Thats fine, just wanted to double check

#12

Illusion of Progress 4 Aug @ 10:53am

Originally posted by ☥ - CJ -:
Should i try overclocking it to 2133?

You can try since some RAM will do more than it is rated for, although most RAM is binned by what it can do, and RAM that could do more would likely have been sold for that.

If there's a chance to do it, you're probably going to have to further raise the timings (erasing gains in latency bound scenarios), increase the voltage (adding to heat, shortening lifespan, and further risking instability), or both.

Last edited by Illusion of Progress; 4 Aug @ 10:54am

#13