Conclusion – AMD’s Price / Performance Leader
Conclusion – AMD’s Price / Performance Leader
At the beginning of this review I mentioned the delays which plagued our 1700X and 1700 samples but looking back, that extra time to digest Ryzen has proven to be beneficial. Now that everyone has had time to take a breather and let some of the emotions swirling around this launch subside, it is becoming evident that AMD has done a ton of things right. However, whether or not you will be interesting in the Ryzen 7 series depends largely upon what you want to use it for. That statement holds true for the Ryzen 7 1700X as well.
Based upon the testing in this review, the 1700X provides a much better value proposition than the 1800X ever could. It may not have that sexy flagship status but the actual performance difference between these two processors is virtually unnoticeable. However, what will be noticeable is the $100 saved by choosing the less costly option, money which can be put towards some other system component that will compliment your setup more than a mere 200MHz.
Another benefit of this CPU is its overclocking prowess. I’m using “prowess” in a loose way since thus far Ryzen processors haven’t exactly been overclocking monsters but with a few judicious voltage and multiplier changes it will easily surpass the stock Ryzen 7 1800X in every benchmark. Our sample didn’t hit the same 4.1GHz mark as the 1800X but from plenty of anecdotal evidence it sounds like both chips are able to consistently attain between 3.9GHz and 4GHz within end users’ systems. That’s another point in the 1700X’s “win” column.
Against Intel’s competition it should be more than evident that the Ryzen 7 1700X is an absolute winner, particularly for its intended market: creative professionals and other folks who can befit from the massive 16 threads it produces. It creams the i7-7700K in many productivity tasks and provides tangible benefits over older AMD architectures. However, scheduling tasks over so many threads becomes a bridge too far for some applications, particularly games and programs that rely upon a different component (be it storage, memory or something else) for optimal performance.
A lot of focus has been put upon Ryzen 7’s inability to deliver Intel-beating in-game framerates. If you want to use a processor
exclusively for gaming then the Ryzen 7 1700X (or any 8-core, 16 thread processor for that matter) will be a less than ideal purchase. Much like Intel’s Broadwell-E processors, it just doesn’t attain the necessary clock speeds or IPC rates. However, if you are looking for a CPU that pulls double duty in workstation and gaming situations then look no further. Other claims like improper thread scheduling leading to fixable performance deficiencies are simply red herrings and in their latest community update,
AMD agrees.
Regardless of snake oil peddlers who say otherwise, that situation will likely remain in place for the foreseeable future even as more games transition to DX12. While Microsoft’s latest API has the ability to stream information towards many concurrent CPU threads, it also allows several typically CPU-bound tasks to be easily taken over by the much faster GPU compute engine. By shunting in-game features like particle effects, physics calculations, and other elements towards the graphics cores, x86 processing loads can be effectively reduced. Thus lower core count processors can do more with less while these massively endowed layouts start to look excessively ponderous.
With that being said, extra cores and threads may be beneficial for certain games, especially those which require a large amount of concurrent calculations for AI. Good examples of that would be games like Sega’s Total War series and Ashes of the Singularity, both of which feature massive armies battling across large territories.
How many threads will ultimately prove to be optimal for those situations? There’s no way to tell right now but if I had to gaze into my crystal ball, 8-12 threads on a high speed architecture would likely be beneficial in some cases. This seems to mesh well with the narrative of AMD’s most ardent defenders but it once again puts us into a never-never land of long term hypothetical situations and nebulous execution dates. In plain English that translates to: we’re grasping at straws now.
As we get through these reviews, remember the current crop of Zen-based processors is just the tip of a very extensive iceberg and you can’t judge their performance in certain applications –particularly games- by looking solely at the Ryzen 7 results. There are upcoming Ryzen 5 and Ryzen 3 derivatives that will (hopefully) boast the higher clock speeds which today’s games require for optimal performance.
As for the Ryzen 7 1700X, it is a very well rounded and efficient processor for people who need great multi threaded performance. It also has absolutely no problems subbing in as a very good gaming CPU if the situation calls for it. In my opinion the $100 saved versus AMD’s own 1800X relative to the insignificant performance difference makes it infinitely more appealing. And against Intel? Well, the 1700X just points out once again how overpriced the current Broadwell-E CPU’s really are.
Answering the SMT Question
Right before the Ryzen 7 1800X launched I received an email from AMD saying in-game performance could improve if SMT was turned off. The irony of AMD’s directive wasn’t lost on anyone here at HWC due to the fact that disabling SMT in an effort to boost performance has long been a tool in tweakers’ pockets. This has been the case ever since the advent of Intel’s Hyper-Threading technology. Granted, disabling SMT is completely unrealistic for most end users and it effectively turns off one of the primary selling points of Ryzen 7 but it also brought up a good question: would Ryzen benefit more than Intel’s processors if their version of SMT was turned off?
The answer to that question is “sort of”. With SMT off almost every single game in this review (other than Battlefield 1) saw framerate boosts which ranged from infinitesimal to moderate. In no circumstance would those increases actually be noticeable to the naked eye though. On the flip side of the coin many real world benchmarks naturally benefit from SMT, though to varying degrees.
It needs to be mentioned that every single benefit Ryzen 7 has without SMT was mirrored by the i7-6900K with its own Hyper-Threading disabled. One interesting result of this testing was Ryzen 7’s SMT scaling beating the Broadwell-E; turning on Simultaneous Multithreading typically led to larger gains than Hyper-Threading. It looks like AMD’s version is simply more efficient than Intel's older architecture. How does it fare against the improvements baked into Skylake and Kaby Lake? That's a question best left to another article.