Ryzen vs. Kaby Lake/Skylake

Hi, Nolo here. Recently I've been spending a lot of time browsing the forums, and I've become somewhat confused. People on these forums seem to do nothing but push Ryzen these days, and that confuses me.

The Ryzen 7 1800x is AMD's current flagship processor. The Core i5-6600k is Intel's previous-generation mid-high range enthusiast processor.

http://cpu.userbenchmark.com/Compare/Intel-Core-i5-6600K-vs-AMD-Ryzen-7-1800X/3503vs3916

The 1800x only sits one spot above the 6600k, with an effective speed rated at 1% more. It costs 100% more.

You can quote the 1800x's multi-core performance, which obviously blows the 6600k out of the water. But overall, how can a processor with a 1% greater effective speed be worth double the dollars?

Obviously, I am an Intel user. I tried to write this post with no bias. Please, do the same in your repsonses.

 

you're looking at user ratings placing it 1 spot higher.


Single core performance, yes the i5 wins with ryzen being 18% slower. This scales to the 4 thread limited test too. However. The i5 only has 4 logical threads, most games these days are designed around 6, some of the DX12 stuff will happily utilise more. The 1800X has 16 logical threads, 8 physical. It doesnt compete with an i5, the ryzen 3 parts are the i5 competitor, also having 4 cores without hyperthreading.

So your 6600k costs £215.48 for current price on my usual retailer. The equivelant to it in AMDs current lineup is the ryzen 1300X, which does have the caveat of only hitting 3.7ghz to the i5's 3.9, however, it will do 3.7 on all 4 cores simultaneous whereas the i5 drops to 3.5 in that situation. It has a lower TDP of 65 watts versus the i5 91W. It is overclockable like the 6600k, and with that 65 watt TDP has a huge headroom to do so with very little in the way of serious cooling gear. It costs £124.99. Motherboards for it are also competitively priced.

This is why ryzen is so popular. When you compare the like for like chips and take account for the workloads they will perform. They offer similar performance for much lower cost.

Even the 1500X, which is the ryzen 5 hyperthreaded quad core, 3.7ghz on boost, designed to go up against intels 6700k rather than your 6600k, only costs £169.49 and is more suited to workloads we will encounter in modern games.

Going further still. The Ryzen 1600 (not X). Slightly lower clock speed. But its a 6 core hyperthreaded chip. 12 threads, each more than capable of handling a T series in game. Costs a mere £190.49, much more power than the 6600k, and still costs less than a 6600k. The 1800X is more powerful than some parts on intels x99 platform which cost far more.


Additionally, anyone that does video editing or streaming, really benefits from having more cores. And the game i am working on myself, early demos I have done as max stress tests can completely max a 6700, although this wont be experienced in actual play.

The only way intel currently wins is on pure single thread performance, but that is simply not a metric which regularly limits people unless we are talking about AMD FX with its absurdly low single thread performance, and even then it had other things that made a bigger difference to its performance in gaming than single threads (main reason *was not* single thread at all).

 

Correct me if I'm wrong, but I thought UserBenchmark was based on the user benchmarks of the hardware, not a user-submitted rating. This is why my question came up; if UserBenchmark was opinion based then it would be of no real comparison value.

Thank you for sharing about the target platforms. I was comparing R3 to i3, R5 to i5, etc. Now I understand a little better the value side of things, at least.

 

No, it's based on user scores only, which is stupid, passmark and CPU benchmark both do based on CPU actual performance. Passmark is the quickest to grab reference from but it only ranks by one metric which is their own multithreaded workload