[TBG] Does RAM Speed Matter? DDR3-1600 vs. 1866, 2133, and 2400 in Games

daglesj said:
What makes me weep is when I see folks that have spent a whole weekend testing and tweaking their RAM timings down as low as they can to get 4578 points extra in a synthetic benchmark but actually only gets them an extra 1FPS (from 112FPS stock settings) in real world situations.
Click to expand...
And, that's why I don't overclock my system, believe in SLI/Crossfire setups, or do extensive watercooling beyond my Cooler Master Seidon 120M, and I only got that because of component clearance verses some of the massive aftermarket cooling solutions. To me, the cost/benefit ratio isn't there, and I want my stuff to last.

If overclocking makes you happy, and extensive treaking for a small incremental increase in performance makes you happy, go for it. I'd rather spend my time running a virtualized web server to enhance my coding ability, learn how to edit videos, and save whatever city or planet from whatever threatens it in the game that I'm playing this week.
 
Its definitely a great way to learn how a computer works and get familiar with the bios ect. I dont overclock anymore but learning how to overclock was a great experience
 
im still on 1066MHz and i noticed no drop in speed, i aim for CL timings, mine are 7-7-7-20 with 12gb of it but my next set is going to be 24gb of CL9 1333MHz so ill report back with any changes :p im expecting not much really
 
tsubasa hanekawa said:
im still on 1066MHz and i noticed no drop in speed, i aim for CL timings, mine are 7-7-7-20 with 12gb of it but my next set is going to be 24gb of CL9 1333MHz so ill report back with any changes :p im expecting not much really
Click to expand...

try to put those cl9 1333hmz sticks at 7-8-7-24 (or 8-8-8-24), it makes a big difference on the overall "feel" of the system.
 
D4rkn3ss said:
try to put those cl9 1333hmz sticks at 7-8-7-24 (or 8-8-8-24), it makes a big difference on the overall "feel" of the system.
Click to expand...


it'd be nice if i could :p the only thing i can change is the actual speed of the ram not the timings, shame really as it'd be nice to tighten up the timings on the CL9 when i get it
 
now running 24gb at 1333MHz (set in bios just to test, all stable :p ) and games are no different so at least with ECC it makes little difference, it makes more of an improvement for me to go back to the 1066 since it has better timings, though the general consensus is that it doesn't past 1600 anyways XD
 
ZLoth said:
And, that's why I don't overclock my system, believe in SLI/Crossfire setups, or do extensive watercooling beyond my Cooler Master Seidon 120M, and I only got that because of component clearance verses some of the massive aftermarket cooling solutions. To me, the cost/benefit ratio isn't there, and I want my stuff to last.

If overclocking makes you happy, and extensive treaking for a small incremental increase in performance makes you happy, go for it. I'd rather spend my time running a virtualized web server to enhance my coding ability, learn how to edit videos, and save whatever city or planet from whatever threatens it in the game that I'm playing this week.
Click to expand...

and you are probably ok with 35 frames per second in your donkey kong game...haha...don't know why you would even peruse this website as you are not [Hard] at all
 
Killahurtz said:
and you are probably ok with 35 frames per second in your donkey kong game...haha...don't know why you would even peruse this website as you are not [Hard] at all
Click to expand...

being on a system that's locked down to only being able to up the ram speed clock (not that it makes a difference) and overclock the slightly bottlenecked GPU i know you don't need an overclocked system to run everything :p
 
ZLoth said:
And, that's why I don't overclock my system, believe in SLI/Crossfire setups, or do extensive watercooling beyond my Cooler Master Seidon 120M, and I only got that because of component clearance verses some of the massive aftermarket cooling solutions. To me, the cost/benefit ratio isn't there, and I want my stuff to last.

If overclocking makes you happy, and extensive treaking for a small incremental increase in performance makes you happy, go for it. I'd rather spend my time running a virtualized web server to enhance my coding ability, learn how to edit videos, and save whatever city or planet from whatever threatens it in the game that I'm playing this week.
Click to expand...

i agree completely, id rather my parts last than have them run a bit faster, i mean of course id overclock given the chance but that's more something i do for fun, i don't go for limits i go for fun and a little bit of free performance but only if i have the proper cooling, though personally being on an intel S5520SC means no overclocking and i don't really miss it if im honest
 
tsubaaaasaaa!! now now tsubasa, dont be trolling us now, youre supposed to be this cute japanese cat girl you know that, we wont accept anything else here. :D

JAT2LRV.png
 
D4rkn3ss said:
tsubaaaasaaa!! now now tsubasa, dont be trolling us now, youre supposed to be this cute japanese cat girl you know that, we wont accept anything else here. :D
Click to expand...

have we met before on here? i remember someone once having a similar reaction :p , meh nyandemonai
 
D4rkn3ss said:
where we come to the realization that youre probably just some random ugly dude from the club of the rejects instead of being "the" tsubasa hanekawa? pretty much yeah :p
Click to expand...


i won't disagree entirely but meh, your views of me are your views and thus mean nothing to me.
 
D4rkn3ss said:
hmm agree, and as for me everything is fine, i wont treat you as tsubasa anymore, i mean cmon, sorry for that ;)
Click to expand...

i can understand where your're coming from though :p all this is besides the point though XD we should probably either get back on topic or be quiet for now XD
 
tsubasa hanekawa said:
i can understand where your're coming from though :p all this is besides the point though XD we should probably either get back on topic or be quiet for now XD
Click to expand...

it is all your fault you know, of all the possible user names available you had to...
 
The whole ram latency thing is pretty much a thing of the past, especially with DDR4. You sorta had to watch it with DDR3-1066/1333 cause there was actually a noticeable boost going to 1600, but gains past that were pretty much muted.

Same idea with DDR4, except DDR4 starts out at 2133MTs, and that's a high bar to start (compared to DDR3). There's so much ram bandwidth being thrown at the CPU with DDR4 that timings just really don't matter anymore baring. You'd literally have to some stupid slow timing like 20-x-x-x to really make a detrimental impact.

This all reminds me of the days tweaking my DDR400 stuff going for those CAS2 / 1T timings...
 
atp1916 said:
The whole ram latency thing is pretty much a thing of the past, especially with DDR4. You sorta had to watch it with DDR3-1066/1333 cause there was actually a noticeable boost going to 1600, but gains past that were pretty much muted.

Same idea with DDR4, except DDR4 starts out at 2133MTs. There's so much ram bandwidth being thrown at the CPU with DDR4 that timings just really don't matter anymore baring. You'd literally have to some stupid slow timing like 20-x-x-x to really make a detrimental impact.
Click to expand...

hmm, you hold a very good point actually, well id have thought like with DDR3 it'll get better and better timings on the more 'standard' speed stuff, though with it being relatively new only really time will show us once it's more 'standardised' and not just on an enthusiast platform only (though if im honest i wouldn't mind a few nice xeons and ddr4 myself :p )
 
ananadtech ddr3 ram benchmarks are useless and as far as there choice of game tests go there ddr4 tests are also useless to as they wont show any change in cpu performance even if there is one with a 4ghz 5960x they are gpu bottlnecked

sleeping dogs gtx770 sli 4570s 2.9ghz within 2fps of 4790k\5960x
bioshock infinite is also similar

tomb raider? well a pentium g2030=92fps while 5960x=96fps 4790k=97fps
http://www.anandtech.com/bench/CPU/1068


actually even worse than being useless they are misleading
 
Yes they even have i7 4790k as slower than lower clocked i5/i7 Haswell parts in some "cpu" gaming benchmarks which should set the alarms to any hardware reviewer.
 
You must log in or register to reply here.
Back
Top