Ryzen 5600G Bottlenecks High-End Graphics Cards? (RTX 3070)

preview_player
Показать описание
"Will the Ryzen 5600G bottleneck a high-end graphics card?" is too vague; it's not a question that has a clear answer.

A bottleneck does not generally apply to a CPU-GPU combination in the broader sense. "Bottlenecking" is something that exists regarding a specific use case. In another case, the roles may be reversed. Let me give you two examples.

"Will the Ryzen 5600G bottleneck an RTX 3070 in CS: GO at 720p competitive settings?". And the answer is yes; the frame rate will be CPU limited. There is a CPU bottleneck in that scenario; however meaningless it may be because the system is pushing 100s of frames already, it exists.

And two: "Will the Ryzen 5600G bottleneck an RTX 3070 in Cyberpunk 2077 at 4k using max settings?". And the answer is no; the 5600G will not hold the 3070 back in that scenario because the 3070 can't even push 30FPS in the first place.

These are extreme cases, but I wanted to highlight that It depends on the application. CPU performance matters less at higher resolutions and settings because it will max out the GPU so hard that the GPU caps the frame rate, and the CPU will have an easier time keeping it fed.

System
CPU: AMD Ryzen 5 5600G
GPU: RTX 3070 (Gigabyte Vision OC)
RAM: G.skill 16GB (2x8) DDR4 3200C14 DDR4 4733C20 (single rank)
MB: Gigabyte Aorus B550i PRO AX
Cooler: Asetek 280mm
PSU: Corsair SFX 450W Gold
OS: Windows 10
Рекомендации по теме
Комментарии
Автор

Your work is so underrated, ty for your tests

Aminsx_
Автор

I know that the 5600x allows PCIe 4 speeds whereas the 5600g is limited to PCIe 3. But I was not expecting these results! Thank you very much for making this video 😃😃

ionamygdalon
Автор

the difference between 5600g - 5600x is 16mb of cache and pcie version. These results proove this.

vrekman
Автор

Thanks man. This was really informative and we'll put together.

Apelgrandeur
Автор

I'm using my 6750 XT with my 5600G and getting 98-100% GPU utilisation, they work great together

iitzfizz
Автор

Love how much work you put into these! I get my 5600G this week, will be disabling the iGPU for a 6800XT and playing at 4K on my TV. By disabling my iGPU will it make it easier to OC the CPU as well as still be able to increase the RAM speeds higher? I have C18 3600Mhz Corsair and wanting to push those harder but worried disabling the iGPU will gimp ram performance?

whatsthatthere
Автор

Man thank you so much. I thought i messed up by buying the 5600g, because i'm using it until i had enough money for a good gpu, but i regret of that when i saw that 5600g will perform worse than others cpus. But by seeing your video, i don't regret anymore and i'm going to buy the GPU (maybe rtx 4060 or its AMD's / Intel's contrapart) and a water cooler to overclock my 5600g and keep it at least for this console generation. Thx again.

lechuga
Автор

Man, I love this video and I always come back to see if you upload that "100 games tested on rtx 3070 & 5600g" video. 🤧🤧

NehuelXD-te
Автор

Intel's initial iGPU bet was to offload some tasks to the integrated GPU if possible. You can actually see gpu usage in some AVX tasks on the Intel cpus that do take advantage of that. However in reality it seems that outside of video encoding/decoding, this "feature" never got used. Looking at 11th gen and up benchmarks, it seems that the feature is no more (or maybe we can't detect it?)

It's interesting that in the end AMD's approach worked out better, initially I really thought that Intel's approach will edge out. However I think we might end up going the same way anyway with performance and efficiency cores. If I had to bet, if none of the usual assumptions about silicon change, a few generations down the line the CPU/GPU concepts will finally fuse and we'll have a handful of perf cores, many (10s to 100s) of efficiency cores and other core "tiers", an integrated FPGA, and many many many (1000s) of tiny GPU-like cores, all on the same (multi-chip) silicon, likely with on-chip HBM memory, like larger and faster eDRAM. We might get (even more?) tiering in memory too. Imagine having fast "DDR6" memory close to the CPU acting as a sort of L4/L5 cache, and a second tier of slower but much larger DDR5 memory right next to it

MenkoDany
Автор

Hey, great video! I'm glad someone looked into this as I bought a 5600G last year because of the GPU shortage and was planning on picking up a 3070 now that the prices have normalised a bit. How do I turn off the integrated graphics once I install the 3070? Is there a BIOS toggle? Tried googling it but I am unable to find an answer. Thanks!

AhaanM
Автор

I just bought Rtx 3070 to pair with R5 5600g(stock) but my performance is not same as yours on 1080p high settings 🥲, The GPU barely uses 60% of it. do you have any advice? Thank you so much.

akanade_chi
Автор

Hi Reto, just stumbled on your video because I still have 5600G, and planning to go for 3070. However, I am wondering if I OCed my CPU to its boost clock of 4.4Ghz, will I get the same result as your 4.7Ghz OCed CPU? Thanks in advance for replying if you do. 😁

xBBBYx
Автор

Thank you so much!!! This cleared up my ideas about getting this processor hell yeah

somedude
Автор

running Ryzen 5 5600G, gigabyte 3070 gaming and zotac 3060 on Asrock A320M HDV-r4.0 with cooler master PSU 650. still rockin

wigunad
Автор

amazing christmas present. announcement of more than 100 games to be tested on r5 5600g + rtx 3070 :D

sesshomarutr
Автор

3:37 Ray tracing on or of ? sorry I'm very bad at English

Михаил-зом
Автор

just copped 5600g for my 3080 since it was on sale and the bottleneck shouldn't matter cause i'll be mainly using it for 4k

gabe
Автор

Did i make a mistake by buying 5600G ? I only use it for 1080p60fps locked, with Rtx 2070.

lil_per_vert
Автор

I play Apex Legends @1080p low setting, my target is geting 220-240fps avg. I use 3060 ti. what will be better 5600 or 5600G? the 5600G is very cheap for about USD 112, while 5600 is around USD 145. Will the 5600G enough?

andreastanu
Автор

How can you seek to compare these configurations whilst using two slightly different spec'd GPUs (including an overclock on the 3070 that was used with the 5600g which definitely works in its favor. You used a better kit of ram for the "stock" test of the 5600g with the "overclocked" test getting an even better kit of ram (or a higher overclock on the same better kit). Please redo your tests with matching ram configuration and the same graphics card.

Rastaaaaaa
visit shbcf.ru