R7 1700 or R7 1700X?

Sovereign936

Reputable
Nov 8, 2015
2
0
4,510
Just wanting to see which one you think would be a better buy for the price. I am replacing my i7-4790k with one of the two, and plan to use a Noctua NH-U9S as a cooler. The only difference between the two that I can see is the different clock speeds in addition to a feature on the 1700x known as "Extended Frequency Range". Since both these processors can be overclocked, which I plan to overclock to about 4.0 or so, would this feature justify the additional $70?
 
Solution
XFR = Auto OCing basically, it looks like

As someone who dared to overclock since his Pentium 4, I honestly see no problem in doing it the same old way. Especially since I still haven't found a single auto tuner that doesn't optimize voltages.
XFR = Auto OCing basically, it looks like

As someone who dared to overclock since his Pentium 4, I honestly see no problem in doing it the same old way. Especially since I still haven't found a single auto tuner that doesn't optimize voltages.
 
Solution
The R7-1700 has a 65W TDP, so it looks like it's the same as the R7-1700X architecturally but with slightly limited overclock potential.

I'm not sure exactly how this limitation is done. Probably the motherboard detects the CPU and won't allow voltages above a specific point.

*In case I'm not clear, the XFR is an automatic feature. You can overclock FURTHER AFAIK within the limitations of what the motherboard, temperature etc allows.

In theory it could make more than 10% difference in performance (i.e. 4.7GHz vs 4.1GHz) but we don't know how these chips overclock yet.

*In fact, if you GAME the i7-4790K should match (due to lack of CPU bottleneck) and may even be slightly BETTER in some scenarios where SINGLE THREAD is more important.

AMD did demo a CPU plus recording (recording with CPU, not GPU acceleration) to demonstrate the CPU benefit with more cores but that's not commonly done.

(Unless you edit video or have some other demanding task aside from gaming I would not spend the money on RYZEN when you have such a great CPU already. Wait for the BENCHMARKS, but I can already tell you there are very LIMITED situations where the extra cores will help.)
 
http://noctua.at/en/noctua-offers-upgrade-kits-for-amd-s-ryzen-platform-free-of-charge

Noctua needs a free adapter kit, however it's not clear how LONG it would take for you to get one. You might want to look into that ASAP.

Other:
Another reason to WAIT, apart from seeing if you really would benefit, is for companies to iron out issues with the BIOS and drivers for the motherboards. It's a good idea to wait at least THREE MONTHS after a motherboard has come to market. There should also be some tweaking with the silicon fabrication as they evaluate YIELDS.

OTHER:
SINGLE THREAD performance is usually the most important thing for gamers. More than FOUR CORES (especially with hyperthreading/SMT) rarely benefits. This goes back to what I said above about seeing if the i7-4790K was as good or not for gaming.

There are going to be lots of articles comparing the pros and cons of RYZEN vs INTEL in the next few months. Some ones to look to include GAMERS NEXUS and PCPER.

DX12/Vulkan?:
Okay, you've heard that more cores is becoming a thing. Great. However, you can always upgrade your computer LATER so do you need to do it now? It's also important to note that DX12 and Vulkan games are slowly trickling in. Once these new API's are properly implemented (they are mostly just tacked on now with limited benefit) we'll see two main benefits to CPU usage:

1) more efficient use of CPU (draw calls) which may save 30% or so. Nobody is certain yet.
2) more THREADS of the CPU to be utilized (many games can only really use about 2.5 cores due to the main game thread being bottlenecked on one core)

There are exceptions as some games can use close to 100% of a quad-core CPU but that's not common, and the reduced draw call benefit will still help.

Let's assume an i7-4790K might be the bottleneck in some situations despite only 60% of its processing power being used. Let's also assume we can use 100%, AND that the code is more efficient with DX12/Vulkan.

60/100*1.3 = .46

That means that the SAME GAME can process almost 50% faster. Add in the fact that most situations are not CPU bound and this starts to eliminate much more of the CPU-bound scenarios.

Of course games could start to use more of the CPU if they want, but I don't think that will be common as game developers try to keep the game running on as much hardware as possible. The amount of CPU processing that the XBOX ONE (base model) has is a strong factor in cross-platform games.