Download

5960X Owners - How is the performance?

Hi all,

I’m putting together a new PC and can’t decide between the 5820K, 5930K or 5960X. The 5960X performs really well in benchmarks but, as reviews seem to only test games and offline renderers, I don’t know if it will provide enough of a real world performance boost to justify the cost (over 2x the 5820K :eek:).

I’m upgrading from a 2500K, and would be overclocking whichever I get. It’ll be paired with 32GB DDR4, Samsung 850 SSD and a 390X once released.

Problem is my game’s getting to the point where it needs some engine modifications, and reports of 45min+ compile times are worrying to say the least. Plus compiling shaders, importing large assets and hot reload are already pretty slow.

So is the 5960X a worthwhile investment?

Thanks,
Nathan

It depends on your budget. I wouldn’t get it, personally. I’d use the money saved to buy some Allegorithmic/Foundry/Autodesk software. Or some assets.

There again, if you are waiting for the 390X to be released, Intel may have released some more CPU SKUs in the meantime. :slight_smile:

Is this for gaming / normal modding. Or building the UnrealEngine from github?

I get about 30 min compile time when I build the whole UnrealEngine from scratch. This is with a 4790k clocked at 4.6GHz and a Samsung 850 PRO.

PS: Once you have compiled you only need to compile the modified files. It takes 20 seconds to compile when I change something in the engine.

Oh I didn’t realise it would only take so long for the first compile (I’m still learning c++ :o). That makes the choice much easier. Plus as Jezcentral says I could put the difference towards some assets. Thanks guys :).

Now I guess the question is, 5820K or 5930K? Am I right in thinking the 5930K is only worthwhile for multi GPU which UE4 doesn’t support? So 5820K is the better choice as overclocked performance should be identical?

Edit: @hallatore this is for building the engine from github.

To me it looks like they are the same CPU except from speed and PCIe lanes.
Personally I would go for something faster but 4 cores. But that’s just because 6 cores doesn’t really help you other than when compiling.

PS: I’m a bit of a C++ noob myself. The source only fully compiles when it thinks it need to. But for some reason it seems that after a couple of hours it always thinks it needs to compile everything. (Might be a setting I don’t know about).
But after that initial “compile of the day” everything is pretty fast after that.

Thanks, TBH it’s mainly compiling that’s slow, not just c++ but also shaders and blueprints, so I figured the 6 core would be the better option. Plus it can always be overclocked, whereas you can’t add more cores to a quad core.

If you’re hellbent on going X99 out of the two I would say the 5820K.

Personally, I would go for a Z97 + 4790K, invest the money you’d blow on the X99 stuff into either software, or towards a Titan X.

I’d go 5820K, supposedly it easily OC’s to 4Ghz to evolve into a 6 core 4790K.
4930K seems like a waste unless you want 4x SLI/xFire.

Compiling shaders/source isn’t the only thing that will benefit from more cores either, a good example being world machine or baking lots of large textures in Xnormal.

It’s also not really more expensive then a Z97 build (going by Newegg prices).