Ryzen 5 5600G vs 5600X?

Hello, would you guys recommend that I exchange a Ryzen 5 5600G ($200) that I bought recently for a 5600X ($250)? The main difference is the 5600X’s 32 MB L3 cache vs 16 MB and PCIe 4.0 support in the former although I don’t own any PCIe 4.0 devices atm. Trying to figure out whether that L3 cache would be a substantial improvement for everyday tasks and DCCs. I mainly do sub-d modeling in Blender, texture in Substance Painter, and render using Eevee in Blender.

I could also futureproof for a PCIe 4.0 NVME SSD years later when they become more affordable. I don’t have need for one immediately since I have a SATA SSD as my system drive. I have a dGPU too, an EVGA RTX 2060 KO Ultra (PCIe 3.0) that I don’t plan on upgrading for a while, and I have no need for the iGPU in the 5600G. In terms of benchmarks, I mainly care about single core performance (Blender viewport, sub-d modeling) and there seems to only be a 5% difference or so according to articles like this. Multicore performance difference seems to be between 11 to 15% for rendering and gaming (I don’t play high frame rate esports titles). Mostly DDC uses for this PC instead of gaming. I’d appreciate any thoughts on this and I should have a few days before I have to make a decision.

I should mention that I’ve used my current PC for 9 years and I hope I’ll be able to use this new upgrade for just as long or slightly shorter. I’m gonna couple either CPU with an MSI MAG B550 Tomahawk mobo (highly rated in benchmarks for its VRM) and 2x16 GB G.Skill Ripjaws V DDR4-3600 RAM.

Having a dgpu negates any benefits 5600g would bring. (namely the igpu for those without a dgpu)

2 Likes

The price difference between the two is $50 or 25% though, which is pretty significant. I’m asking if the 16 MB extra L3 cache, PCIe 4.0 support, and very slight single core performance increase is worth the added $50 price tag. :slight_smile:

Let me rephrase that.

How long are you going to use the cpu and is $50 dollar difference worth it?

1 Like

I mentioned at the bottom of my OP that I hope to use this new CPU for 9 years or so and above that I mentioned that I might buy a PCIe 4.0 NVME SSD a few years later. So I’m considering upgrading to the 5600X in order to help futureproof a bit.

I am asking for help in order to determine whether the $50 difference is worth it based on all of the factors that I mentioned, yes. :stuck_out_tongue: My main point of interest is the L3 cache improvement and whether it would affect everyday tasks and CGI-related work to help justify the 5600X’s cost.

If you are talking about 9 years then the difference is negligible. $50 over 9 years is also pennies changes.

blender and substance painter won’t benefits from gen4 ssd.

You will benefits more with gpu upgrades, in that case gen 4 lane might be needed in the future.

That’s true, but I’m also a value-minded buyer and I enjoy getting the most bang for my buck. :slight_smile: Why else would I only choose mid-grade CPUs, motherboards, and RAM? Although $50 is negligible across nearly a decade as you mentioned, it’s a pretty sizeable upfront cost for my budget.

The PCIe 4.0 NVME SSD is just an example of an upgrade I might want to make in the near future in order to justify the 5600X. I know that most DCCs won’t benefit from it unless they get some sequential read/write witchcraft. Random read/write speeds are very similar between NVME and SATA SSDs. I am mainly wondering about the significance of the 5600X’s 32 MB L3 cache over the 5600G’s 16 MB in everyday tasks and DCCs.

If you don’t need the iGPU then this is pretty much a no-brainer. Make the exchange.

5600x also supports pci-e 4.0 tho, no? So long as your mother board supports it too there should be no issue.

But realistically over the course of 9(!!!) years you should be able to find a good deal on a used 5900x or even a 5950x so if I was you I’d likely jump at that anyways.

It’s your motherboard choice that future-proofs you somewhat, not the CPU choice itself.


edit: Ah I misread that earlier. Yeah the extra cache alone is probably worth the $50 and the PCI-e 4.0 support only makes it better, but that is only relevant if you actually get a 4.0 device to take advantage of it within the time-frame of it’s use.

Also, if you can - I’d see if a good deal on a used 5600x can be found now instead of getting a brand spanking new one. Might make the calculation a ‘do I get this or that for 200’ instead. Making the 5600x even more of a no-brainer.

A no-brainer based on which of the points I mentioned? >> I know that is what most gamers recommend in various PC building communities like Linus Tech Tips, Tom’s Hardware, and some subreddits. But they are basing their recommendation of the 5600X on game benchmarks that use multiple cores, which result in an 11 to 15% performance increase. CPU-bound games in particular have very skewed benchmarks when comparing the 5600X to the 5600G, between 30 to 40% performance increases. Single core performance is only like a 5% performance increase.

The 5600G only supports PCIe 3.0 and the 5600X supports PCIe 4.0.

I’m very hesitant about buying any used PC components since I have no idea whether the previous owner overclocked too hard and possibly damaged the chips. I’d rather buy new (not even refurbished) parts straight from the factory in hopes of winning the silicon lottery. It’s worked for me this past decade and I won’t change this habit anytime soon. :stuck_out_tongue:

When you say that the extra cache alone is probably worth $50, could you please explain why that is? As I mentioned before, everyone keeps going on about the extra 16 MB L3 cache without explaining what advantage it brings and they can’t back it up with benchmarks.

Edit: I should mention that I appreciate you guys offering your suggestions and I hope I don’t sound ungrateful. :grinning_face_with_smiling_eyes: Much appreciated to be receiving replies.

12600k is twice faster than 5600x

The extra cache allows more data to be physically closer to the CPU itself. So it is quicker to fetch the data and improves performance of the CPU. Think of the extra cache like having a bigger table to store books on vs having to go to the bookshelf in another location to get more books.

For $50 and your plan it use it for 9 years the 5600x is a better choice.

Please link the benchmark you saw. I’m mainly interested in single core performance and the i5-12600K is definitely not twice as fast as the 5600X, maybe 21% according to benchmarks. Multicore performance difference is higher at about 38%, but that’s not double. I’m on a budget and Intel motherboards are currently very expensive compared to similar AMD offerings. The i5-12600K also lacks a cooler.

@Chuk_Chuk I definitely understand the analogy and I’ve read similar ones online. I’m interested in the performance implications though and most of the benchmarks I’ve found are like 7 years old comparing 2 MB to 8 MB L3 cache. I assume this is the reason for the 6500X’s 30 to 40% frame rate increase in CPU bound games like Overwatch, LoL, CS:GO, etc.? I’m just wondering how double the L3 cache affects everyday tasks and DCCs like Blender though.

I think 9 years is quite unrealistic. Things will change very fast now that there is new CPU competition, plus memory will improve with new bus, the new Ryzen with 3D memory are around the corner, then by late next year Ryzen 4 family.

It should translate to a straight increase in performance for an application that can take advantage of it. The only application that could not take advantage of it would be an application or calculations that would fit within the smaller cache value (16MB).

If you want to know how much of a performance increases you get, you will need to find comparisons between the CPUs with the exact software you are interested

Thank you guys for all the suggestions. I’ve decided to go with the 5600X in order to futureproof my setup a bit. :slight_smile:

It’s not even so much about ‘added performance’. More cache is actually all about avoiding cache misses, which causes a trip to fetch the data from RAM, which is slooooooooow. Not as slow as having to fetch that data from swap/virtual memory stored on a HDD tho. So it’s more about avoiding slowdowns rather than gaining speed.

Might seem similar and they are in that they are opposites but lets say one is more catastrophic than the other.

Yeah, I’ve googled what the different cache levels do and it makes sense. A reddit post that I found also explained just how fast the L3 cache is compared to RAM. I was mainly looking for real world benchmarks that compare 16 MB to 32 MB L3 cache in everyday tasks or single core, CPU-bound tasks like Blender’s viewport while sub-d modeling. Nothing as specific as that from what I could find, lol. It’s just that when you said the 16 MB extra L3 cache is worth $50, it sounded awfully vague to me due to the lack of supporting evidence or benchmarks. :S