Is i5 still king for gaming or will i7 show an advantage with DX12 ?

[whitedragon101]

For ages it seems that a k series i5 is the choice for gaming and i7 shows no real improvement. But with DX12 and the new VR gaming will an i7 be worth the extra cost for gaming ?

[mardon85]

I couldn't answer the but one thing that did surprise me was the i7 with hyperththreading performed worse than an i5 in GTA V in the linus video.

[shaithis]

Well, at the moment there are cases where the i7 will show gains....going forward with DX12 I think the number of cases where the i7 is better will diminish.

[Ttaskmaster]

I certainly hope i5 is king... I'm about to go from Red to Blue!!

[Mister]

The i7s are pulling ahead slightly. If games start using hyper-threading (not sure how likely this is yet) then the i5s will be left way behind.

[evs]

Cost wise, the i5 is better for gaming as it's hard to justify the i7 pricing if all you do is games and never take advantage of the features. If you do anything else for a fair chunk of time using HT, then the i7 maybe better for you. It's certainly not a case of i7 chips tanking during games!

[watercooled]

Games don't 'use hyperthreading' as such. Without getting into the details, multiple threads are made available by the processor to the operating system which assigns them to running threads on the system.

If a game were to efficiently use more than four threads, and the workload was one well-suited to SMT (not all workloads will benefit equally from SMT because it's based around sharing resources between threads), and the performance with i5 wasn't sufficient in the first place, then an i7 may give improved performance.

One of D3D 12's improvements is better multithreading, however it also reduces CPU load overall, so as shaithis said, we could see even less reliance on CPU, not more.

[KeyboardDemon]

It's a good question and an interesting thread.

Based on what shaithis has said, and watercooled has added it does look like the i5 is still in the sweet spot for CPUs, but of course their arguments are based upon known factors, ie. what we already know about DX12 and what we know about current comparative information on the i5 vs i7. If you can wait until the end of July, it might be worth holding on so you can see actual results in order to find out for sure, but if you have to buy now, I don't think you would that the i5 would be lacking in 2 months time.

[mardon85]

I think he'd probably have to wait longer than July, just because dx12 will be out we still have to wait for the games to mature.

I went for an i7 from the previous generation and haven't looked back. The only really difference between the two is the internal GPU which we don't use.

[KeyboardDemon]

I think he'd probably have to wait longer than July, just because dx12 will be out we still have to wait for the games to mature.

Good point, though it wouldn't surprise me if there were a few games and at least a handful of tech demos, not to mention a DX12 version of 3dMark to keep our interest levels up to begin with.

[Phage]

For SLI/CF an i7 does better due to the extra pcie lanes

[watercooled]

The primary difference between i5 and i7 (socket 1150 desktop that is) is whether SMT is enabled. Other differences are increased cache and a token clock speed bump. WRT the cache, we have i3 CPUs with both cache assignments (1.5MB or 2MB per core) and the difference in desktop performance tends to be negligible. They are produced from the same die and hence no architectural differences; performance differences arise from one or a combination of the above.

i5 and i7 on socket 1150 have the same number of PCIe lanes available. Socket 2011 CPUs offer more but themselves are generally less well-suited to gaming than workstation use because of a greater number of lower-clocked cores. Also, even 16 PCIe 3.0 lanes is very rarely a bottleneck even with dual-GPU setups i.e. 2x8 lanes.

On the subject of DX12, see what I said above - there's very little reason to think the i5 will start struggling since the aim is to reduce CPU overhead, with the addition of improved threading. If something other than DX overhead is causing a bottleneck on the CPU, that has nothing to do with DX so no different to now.

Some benchmarks of CPU scaling here, bearing in mind they're using synthetic tests designed to stress the DX layer (i.e. real games should have far less overhead): http://www.anandtech.com/show/8962/t...a-star-swarm/4

[mardon85]

Let's hope the take up of dx12 is faster than dx10. It really does seem a great Api.

[abaxas]

If dx12 is windows 10 only, it'll take years before it is used.

Why would a developer target something with such a small marketshare? Might as well develop for the ps2

[mardon85]

I think that's one of the reasons they're pushing the free upgrade so hard. It is a very similar situation as vista and windows 10. The only possitive difference is xbox one this time. Xbox only supported dx9.

[DanceswithUnix]

If dx12 is windows 10 only, it'll take years before it is used.

Why would a developer target something with such a small marketshare? Might as well develop for the ps2

Vulkan should be out shortly after DX12, and as it is based on Mantle I would expect writing for Mantle would make for an easy conversion later. That should be cross platform then, allowing good performance on Win 7 and Linux as well as Win 10. The new APIs *will* happen, the performance improvement is too obvious to ignore. Gamers will upgrade to 10 for this, it isn't like the meh performance improvements of Vista that everyone ignored.

Back to the original question though, when I bought my current setup a couple of years ago I was choosing between i7 (fast) and Fx8350 (good value) even back then. If you want the king of gaming, then you want an i7. If you are on a tight budget like most of us, then the i5 allows you to spend more on the graphics card which is usually a win overall.