Teach me about RAM timings and what the numbers mean

Teach me about RAM timings and what the numbers mean

Attached: RAMurmum.jpg (1590x534, 237K)

Other urls found in this thread:

computer.howstuffworks.com/ram.htm
users.ece.cmu.edu/~omutlu/pub/onur-ACACES2013-Topic1-dram-basics-and-scaling.pdf
hardwaresecrets.com/understanding-ram-timings/
twitter.com/NSFWRedditGif

higher is better

*higher is better

For the average consumer it's meaningless today.

15 years ago better timed RAM could mean 10 % more performance though.

clock speeds of ram are negligible as an average user. Dont spend too much time thinking about it

it's a counter. Every time you use your ram, the number on the sticker goes up by 1

*higher IS better

*better is HIGHER

Attached: 1515351548319.png (848x1492, 487K)

>For the average consumer it's meaningless today.
Wrong. It was meaningless but has recently become pretty significant.

>ADOLF
???

>tfw you got your VLB rig stable with 0 wait states

Hahah. Now I can't unsee.

>t. Owns 24 24 46 DDR4

My cpu and mobo support up to 2600. I ahve two 2333mhz sticks in but it runs at 1333mhz, why?

you have a shit motherboard and/or haven't turned on xmp profile

I am currently shopping for used 20$ 8gb ddr3L's for muh 7 years old dell precision to pump it up to 32gigs.
any tips?

>xmp profile
>extreme memory profile profile

Attached: Cylindrical.png (309x309, 266K)

i do have a shit mobo but it is supported anyways, how do i turn that on?

computer.howstuffworks.com/ram.htm

learn from one of the elder gods of DRAM
users.ece.cmu.edu/~omutlu/pub/onur-ACACES2013-Topic1-dram-basics-and-scaling.pdf
see slides 96, and 116-117 for the high level idea
basics start on slide 40
i'm sure you can find some lectures on youtube if you prefer.

once you have a clue how dram works then it's easy to google and understand pages like this
hardwaresecrets.com/understanding-ram-timings/

Attached: memory_hierarchy.jpg (664x522, 67K)

thanks from op

*slide 94, not slide 96

Basically latency of RAM doing stuff. The higher the latency the longer it will take to get things done. In a lot of vidya and specific workloads this has a direct correlation with higher or lower FPS/performance especially on AMD's Zen processors. Pic related.

Attached: 67215981_latencyvsspeed.png (767x441, 35K)

Attached: computer_hardware_chart_2_0_by_sonic840_d9pyk3j.jpg (3600x4800, 3.99M)

needs an update

from the bios, check your mobo manual.

Attached: witcher3_avg[1].png (500x810, 40K)

BUY HIGH SELL LOW

>that resolution
so you're supporting his argument? you need to be really reaching for those numbers

Attached: 1553479573425.jpg (680x813, 56K)

the performance gain in the image you posted is practically negligible even at 720p. you are a retard with no argument

I fucking keked at the captions

the difference between constant 144hz and having dips. your casualness is showing, pleb

its the difference between 50 and 50.5 at a real resolution

Tfw quad channel so high bandwidth, high latency
Tfw the latency is probably raping me a bit

At least RAM bottleneck tends to come in after GPU bottleneck (unless you do shit like 720p medium there) and my GPU is definitely the bottleneck of my system for gaming

>real resolution
>anything over 1024 and less than 120 fps/hz
confirmed casual pleb

>He doesn't game at 4k144hz
>1998 + 21

yeah cus im not a casual pleb like you. look up what res the pros use, shitter

seething zoomer

>benchmarking RAM while being GPU bound

Attached: brain_damage.jpg (600x604, 52K)

the point is "for the average consumer its meaningless" which is absolutely true unless you can demonstrate a significant performance difference under appropriate conditions

>not being RAM bound because clan uses Discord

>implying average consumers do nothing demanding besides playing vidya

Attached: cat.jpg (600x450, 117K)

so you can surely provide some examples then

This image is fucking sexy.

Oh I'm sorry. Do benchmarks not count unless they're 640x480 minimum settings like Intel ones are? Fuck off, 720p med is just fine for RAM benches.

This user gets it.

Attached: tomb_raider_min[1].png (500x810, 40K)

>rendering video

average consumers watch

>more 720p

Attached: 1549491746914.jpg (500x700, 148K)

Yes, let's test infinity fabric bottlenecks at high resolutions.

Attached: 1513308792562.png (679x769, 48K)

Testing with 720p is like testing GPUs with fucking Quake 3. You get a result but it has no meaning in todays tech world.

keep posting self portraits

>not benchmarking GPUs with glgears

Attached: brainlet_black_hole.jpg (1024x595, 74K)

Reminder to pair your RAM with a fitting fan color.

Im stealing this shut nigga

So do you idiots benchmark GPUs with AIDA 64? There's no way people on Jow Forums can be this retarded.

Divide the Hz by the cla

To get the real speed

Rate timings are inversely proportion to the propensity of you having a gf. Rate my ram:
> 8.00GB Dual-Channel DDR3 @ 802MHz (11-11-11-28)

>>me married

3000Hz 16GB cl16 ddr4

being married is worse than being single lmao