4C/4T vs 4C/8T vs 6C/6T

[newtekie1]

I'm thinking of redoing these tests, but adding in the 6c/12t, as well as doing a few game tests. I'm thinking GTA:V, Cities Skylines, and FarCry5.

If you throw out the Intel burn test and cinebench, all the 4c/8t and 6c/6t results are within 15% (13% at widest, 6% at narrowest) of each other. We your co-worker was saying has merit as going from 4c/8t to 6c/6t doesn't yield enough benefit to justify it.

In the end it is going to come down to pricing to determine if it was worth it or not. If we only look at K series skus, the 8350k is $185 and the 8600K is $245, that already is only a $60 price difference. If the 4c/8t part was priced somewhere in the middle of those two processors, then it wouldn't be worth it at all, spend the extra ~$30 and get the 8600k(frankly, spend the extra $60 and get the 8600k over the 8350k). Even if the 8350k was released as a 4c/8t part, at the same $185 price point, IMO it is a hard sell to save the $60 over an 8600k.

 

[GorbazTheDragon]

I wonder what intel will do in the future? 8/16 i7s, 6/12 i5s, and 4/8 i3s?

Maybe they will want to have some 8/8 and 6/6 chips at lower cost to be able to sustain sensible production rates for those dies, but considering the market share of the 8700k is substantial compared to the 8600k and 8400, i would not have expected this to be a problem.

 

[SamirD]

I wonder what intel will do in the future? 8/16 i7s, 6/12 i5s, and 4/8 i3s?

Maybe they will want to have some 8/8 and 6/6 chips at lower cost to be able to sustain sensible production rates for those dies, but considering the market share of the 8700k is substantial compared to the 8600k and 8400, i would not have expected this to be a problem.

I would look at the xeon line to see what the roadmap might look like. It's no secret that a lot of the desktop chips share a lot with their server/workstation brethren.

 

[Bill_Bright]

I'm thinking of redoing these tests, but adding in the 6c/12t, as well as doing a few game tests. I'm thinking GTA:V, Cities Skylines, and FarCry5.

The more variety and number of "samples" you include will indeed make the results more realistic. But to be realistic, there should be tests other than games too. Compiling huge data bases, for example.

But frankly, I think you have already proved the point nicely. To be any more useful to someone, your test would have to based specifically on that user's specific system, and the specific tasks (programs) he or she will be running.

I'm going back to what I said earlier. Windows knows how to use HT. And since Windows can complete its own tasks quicker by using HT, that's better for "over all" system performance. So IMO, if the budget allows, that's reason enough to spring for a CPU that supports it.