• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD's Infinity Fabric

opethdisciple said:
Faster memory improves Ryzens gaming performance drastically up to 3200MHz after which you get diminishing returns.
Click to expand...

Yeah I'd agree with that if you have a nvidia 1080 or above... see here https://www.youtube.com/watch?v=U4k_ErEg-FU

For us in the 3D Mark Timespy results I linked to, we both have a PowerColor Radeon RX 580 RED Dragon V2 Graphic Card 8192 MB so the memory speed doesn't really have any affect. My reference? that youtube vid I just linked to.
 
Rossi~ said:
Impossible to see the difference unless you're using the same CPU, that comparison is just telling you that a better CPU is better...
Click to expand...

I could try and compare my machine with another with the same CPU but different graphics cards overclock to different speeds hence why I compared two machines with the same graphics card and the same GPU overclock, also I am not overclocking my CPU so I'd have to find another who also isn't overclocking.

Finding someone else with same graphics cards and GPU overclocks with the same CPU at same speed can be frustratingly difficult.
 
nlel1975 said:
I could try and compare my machine with another with the same CPU but different graphics cards overclock to different speeds hence why I compared two machines with the same graphics card and the same GPU overclock, also I am not overclocking my CPU so I'd have to find another who also isn't overclocking.

Finding someone else with same graphics cards and GPU overclocks with the same CPU at same speed can be frustratingly difficult.
Click to expand...

Why not then overclock your memory and see what happens? rather than comparing apples to pairs, compare your machine with slower ram to your machine with faster (overclocked) ram. Even your 2133 might hit like 2400 or more with some tweaking.
 
Vince said:
Why not then overclock your memory and see what happens? rather than comparing apples to pairs, compare your machine with slower ram to your machine with faster (overclocked) ram. Even your 2133 might hit like 2400 or more with some tweaking.
Click to expand...

I could and it probably would overclock a bit but I can't be arsed to mess around with it, the difference between 2133 and 2400 is less than 2133 to 3000 obviously. There maybe small gains purely down to RAM speed but it wouldn't be in the hundreds.
 
nlel1975 said:
I could and it probably would overclock a bit but I can't be arsed to mess around with it, the difference between 2133 and 2400 is less than 2133 to 3000 obviously. There maybe small gains purely down to RAM speed but it wouldn't be in the hundreds.
Click to expand...

Probably not. I think the take away here is just because you can't see it in a single bench does not mean that it isn't there. We can't see the black hole at the center of our milky way galaxy for example but we know it is there, physics tells us it is there. In much the same way AMD tell us in many whitepaper that IF scales off of the memory speed, so we know it is there, AMD tell us it is there. People have tested the theories and confirmed them, unlike the black hole at the center of our galaxy you don't need to be a rocket scientist to test this theory.
 
Last edited:
single_channel_vs_dual_channel.jpg


https://forums.overclockers.co.uk/t...l-vs-dual-channel-best-ram-settings.18831204/
 
Yes but that table is using a 1080ti, it's proven that RAM speed has an effect on FPS with a 1080 and above. See the YouTube video I linked to above.
 
nlel1975 said:
Yes but that table is using a 1080ti, it's proven that RAM speed has an effect on FPS with a 1080 and above. See the YouTube video I linked to above.
Click to expand...

Err, you can't just say it only does it with 1080 and above... It doesn't, even if you only notice it here in fps stats, perhaps at 1080p high fps when the bottleneck isn't a cheap gpu it is more obvious, but listen, take the gpu out of it entirely and run blender or another productivity suite or something... I bet you see the cpu with the higher speed ram performing better. As I said about my black hole analogy above, the fact you cannot see it in a given scenario doesn't mean that it isn't there.
 
If you're bottlenecked by the GPU, then you'll not really see much of a gain at all (if any) from the CPU/RAM.
 
Vince said:
Err, you can't just say it only does it with 1080 and above... It doesn't, even if you only notice it here in fps stats, perhaps at 1080p high fps when the bottleneck isn't a cheap gpu it is more obvious, but listen, take the gpu out of it entirely and run blender or another productivity suite or something... I bet you see the cpu with the higher speed ram performing better. As I said about my black hole analogy above, the fact you cannot see it in a given scenario doesn't mean that it isn't there.
Click to expand...

I meant nvidia 1080 graphics card not 1080p.
 
nlel1975 said:
I meant nvidia 1080 graphics card not 1080p.
Click to expand...

I know exactly what you mean, I elaborated by saying using a 1080 @ 1080p (where you are likely to see the biggest difference between gpu's and cpu's) as the res goes up the gaps close up.
 
What about this:
https://www.reddit.com/r/Amd/comments/6nwudy/i_compared_3200mhz_ram_vs_2133_in_12_games/

Granted its a 1080ti but still it does no favours to pretend that for AMD Ryzen CPUs that the memory speed is unimportant. If the extra cost is not worth the percentage then thats definitely your opinion.

But for the fact that you can see 20-40% improvement in gaming with a good CPU and Fast ram vs slower ram (speeds your ram runs at) does definitely show that if you were not getting the FPS you wanted on your system, the cheapest upgrade currently to a Ryzen system would be faster RAM. If you were running slow stuff.

RAM speed may not make a CPU core any faster, but normally the CPU core is choked by the fact the data isn't coming in fast enough, especially when tasks move across CPU cores to each other they are limited by the Infinity Fabric speeds. If you speed up the RAM you enable the data to move between cores faster, which will improve performance by reducing the bottleneck on the Infinity Fabric. I like how all the technical data I sent you previously regarding how much of a difference it made you ignored and tried to find other people to agree with your mistaken point.
 
What i dont get is why you opted for 32GB of slow ram when for the same price you could have gotten 16GB of really high end Ram? is your workload extremely ram hungry?
 
In other news, I'm going to argue that the sky is bronze and no amount of empirical evidence will convince me otherwise...

...although the sky was indeed bronze in ancient Greece.
 
SiDeards73 said:
What i dont get is why you opted for 32GB of slow ram when for the same price you could have gotten 16GB of really high end Ram? is your workload extremely ram hungry?
Click to expand...

Well I had 16GB of DDR3 1333 in my old machine so wanted double the RAM, the DDR4 was faster at 2133 anyway. I know I had an Intel chip before which didn't have IF but I would rather have 32GB of RAM I won't use then 10% to 15% faster RAM I won't notice.
 
You must log in or register to reply here.
Back
Top Bottom