I will try to keep it as short as possible, tho there is a lot to get through.
There are a couple of things to know about CPU performance testing for Games, the most fundamental and basic of which is to insure the performance you are getting from the system is bound be the CPU and not the GPU.
If the GPU is doing most or all of the work then you are not seeing how much performance the CPU has to give, make sense, right?
To me a lot of Mainstream reviewers fail to do that.
For the most part their CPU game performance testing they are using a GTX 1080 GPU, 2667Mhz System RAM and the highest game quality settings possible. Toms Hardware use Ultra High quality settings @ 1080P in BF1, for example..... now i have a GTX 1070 OC to run at around stock GTX 1080 performance and i know in campaign mode at those setting the CPU is not being stressed at all.
As a result a CPU with higher single threaded performance, such as a higher clocked CPU like the 7700K will win, predictably that is the conclusion all of these mainstream reviewers arrive at,
A small taster of that....
Pcgamer
Guru3D
That second one, Guru3D actually say Intel is better when CPU bound, yet if you look at their review none of the CPU's are ever CPU bound, what a joke...
Nothing at all wrong with the results they got and their conclusions based on those results.
Its the methodology and reasoning behind it that's flawed.
Most of us keep our CPU's for 3 to 5 years, the thing that we upgrade is the GPU, we do that because they get faster, games get more advanced and use more system recourse as a whole, so what i want to know is how much longevity is in the CPU? how much headroom is left in it?
When you look at reviews that do actually stress the CPU what you see is that often (not always) but significantly; is that the equivalent Intel CPU's are nothing like as powerful as the Ryzen CPU's in a lot of modern games.
I think the problem is an assumption that game do not use more than 3 or 4 threads, that is an out of date idea, its an idea that should have been put to rest when you realised that back in the day a 4 core 2500K was always 20% or so faster than the FX-8350, now its often the other way round. Games these days use way more than 3 or 4 threads.
I can easily illustrate this with what is a well executed CPU review.
We all know Metro Last Light, how CPU Intensive it can be, look at this.....
That is a massive 55% performance advantage to the 6 core Ryzen over the 4 core Intel, a ridiculous performance win for AMD,
why? well this is why a video run-through review is so much more telling that numbers on a slide, look at the thread load on the Intel CPU on the right vs the AMD CPU on the left, all of the Intel's 4 threads are completely saturated, as a result the GPU is over 50% more bottlenecked on it than it is on the Ryzen CPU.
Now, to be even more clear about this,
First image both CPU's at 249 FPS, i caught it just where the Intel CPU was at its limit.
Moving on a little the Intel CPU still at 100% but the game wants more, to the Ryzen chip now pulls ahead, 318 FPS vs 277.
Moving on yet more and the Ryzen CPU is still gathering more pace....
The truth is when actually put to the test AMD's Ryzen CPU is not just faster than Intel's rival, it destroys it....
Here is the Ryzen 1700 @ 3.9Ghz vs the 7700K @ 5Ghz in BF1, yes what you are seeing there is the 7700K's 8 threads completely maxed out.
Back when CPU reviews was Sandy or Ivy Bridge vs Bulldozer or Vishera reviewers did off load all the work on to the CPU, Anand went as far as to benchmark games at 480P to be absolutely sure.
So whats changed now to change the testing methodology to something that does not load the work on to the CPU at all?
IMO the answer, perhaps is in the results once this testing methodology is used.
It makes me sad because not only are the reading public not getting whole truthful information, its also confirming my fears that no matter what AMD do, no matter how good they make their competing products they will always be portrayed as the loser, the one not to go with....
https://www.youtube.com/watch?v=WlOs_McAVZ0
https://www.youtube.com/watch?v=BXVIPo_qbc4
There are a couple of things to know about CPU performance testing for Games, the most fundamental and basic of which is to insure the performance you are getting from the system is bound be the CPU and not the GPU.
If the GPU is doing most or all of the work then you are not seeing how much performance the CPU has to give, make sense, right?

To me a lot of Mainstream reviewers fail to do that.
For the most part their CPU game performance testing they are using a GTX 1080 GPU, 2667Mhz System RAM and the highest game quality settings possible. Toms Hardware use Ultra High quality settings @ 1080P in BF1, for example..... now i have a GTX 1070 OC to run at around stock GTX 1080 performance and i know in campaign mode at those setting the CPU is not being stressed at all.
As a result a CPU with higher single threaded performance, such as a higher clocked CPU like the 7700K will win, predictably that is the conclusion all of these mainstream reviewers arrive at,
A small taster of that....
Pcgamer
For pure gaming performance, I'd still point people to Intel, but if you're looking for something of a Jack-of-all-trades processor that won't break the bank, the new Ryzen 5 chips warrant serious consideration.
Guru3D
The biggest discussion at Ryzen's launch was 1080p gaming performance. This problem is still here, but not as big as some state it is. Ryzen is a truly great processor series, but it lacks a little in 1080p gaming situations where you are more CPU bound (if you have a fast enough graphics card). There has been much debate on the cause of it, memory latency, latency in-between the CCX modules on the processor, driver issues, Windows 10, game optimizations, benchmarking with a GeForce card over an AMD one, thread schedulers and so on. The reality is simple, the results are what they are. Ryzen 5 and 7 lack a good 10-20% in performance with super fast graphics cards in a lower resolution compared to the fastest clocked Intel SKUs.
That second one, Guru3D actually say Intel is better when CPU bound, yet if you look at their review none of the CPU's are ever CPU bound, what a joke...
Nothing at all wrong with the results they got and their conclusions based on those results.
Its the methodology and reasoning behind it that's flawed.
Most of us keep our CPU's for 3 to 5 years, the thing that we upgrade is the GPU, we do that because they get faster, games get more advanced and use more system recourse as a whole, so what i want to know is how much longevity is in the CPU? how much headroom is left in it?
When you look at reviews that do actually stress the CPU what you see is that often (not always) but significantly; is that the equivalent Intel CPU's are nothing like as powerful as the Ryzen CPU's in a lot of modern games.
I think the problem is an assumption that game do not use more than 3 or 4 threads, that is an out of date idea, its an idea that should have been put to rest when you realised that back in the day a 4 core 2500K was always 20% or so faster than the FX-8350, now its often the other way round. Games these days use way more than 3 or 4 threads.
I can easily illustrate this with what is a well executed CPU review.
We all know Metro Last Light, how CPU Intensive it can be, look at this.....
That is a massive 55% performance advantage to the 6 core Ryzen over the 4 core Intel, a ridiculous performance win for AMD,
why? well this is why a video run-through review is so much more telling that numbers on a slide, look at the thread load on the Intel CPU on the right vs the AMD CPU on the left, all of the Intel's 4 threads are completely saturated, as a result the GPU is over 50% more bottlenecked on it than it is on the Ryzen CPU.
Now, to be even more clear about this,
First image both CPU's at 249 FPS, i caught it just where the Intel CPU was at its limit.
Moving on a little the Intel CPU still at 100% but the game wants more, to the Ryzen chip now pulls ahead, 318 FPS vs 277.
Moving on yet more and the Ryzen CPU is still gathering more pace....
The truth is when actually put to the test AMD's Ryzen CPU is not just faster than Intel's rival, it destroys it....
Here is the Ryzen 1700 @ 3.9Ghz vs the 7700K @ 5Ghz in BF1, yes what you are seeing there is the 7700K's 8 threads completely maxed out.
Back when CPU reviews was Sandy or Ivy Bridge vs Bulldozer or Vishera reviewers did off load all the work on to the CPU, Anand went as far as to benchmark games at 480P to be absolutely sure.
So whats changed now to change the testing methodology to something that does not load the work on to the CPU at all?
IMO the answer, perhaps is in the results once this testing methodology is used.
It makes me sad because not only are the reading public not getting whole truthful information, its also confirming my fears that no matter what AMD do, no matter how good they make their competing products they will always be portrayed as the loser, the one not to go with....
https://www.youtube.com/watch?v=WlOs_McAVZ0
https://www.youtube.com/watch?v=BXVIPo_qbc4
Last edited: