AMD RYZEN 7 1700x and 1700 CPU Review

After much anticipation, the Ryzen 7 1700X and Ryzen 7 1700 CPUs are finally here. The internet has been buzzing since launch and not everything has been in-favor of the new architecture. My personal experience so far has been great, really, and I feel AMD has delivered this time. Going into development they had a fairly lofty goal of a 40% IPC (instructions per clock) gain over “Excavator,” the last generation in their K15 architecture line. So AMD gathered together a formidable team of engineers and went to work.

Designing a new CPU from the ground up is no small task and does take some time. The ground work was laid in 2012 when AMD brought Jim Keller back on board to lead the design team. Things seemed very quiet for a long time in the AMD camp, and for three years all we had were internet rumors and whisperings until the day this hit the web.

That was some announcement. So many firsts for AMD and all of them squeezed into a whole new design. This is huge. AMD has taken the plunge more than once and it’s paid off for them: the on die memory controller (IMC) which greatly improved memory handling and the introduction of AMD64 which added 64 bit computing capabilities to existing x86 architecture. So let’s fast forward to 2017. Did AMD’s gamble pay off?

In my opinion, it most certainly did. The performance boost over my FX systems is outstanding. Their implementation of SMT (Simultaneous Multi-threading) is highly efficient and on top of that they have dropped the TDP significantly. According to AMD they have beaten their goal of a 40% IPC gain by an additional 12%.

Specification and Features

So lets have a closer look at the two CPUs I have up today, the Ryzen 7 1700X and the Ryzen 7 1700. I’ll quickly go through the specifications of each and how they differ from one another. The Ryzen 7 1700X and Ryzen 7 1700 are identical to their big brother the Ryzen 7 1800X in every way aside from base clocks and boost speeds. The Ryzen 7 1700 is the only one of the three that doesn’t have the XFR (eXtended Frequency Range) technology which will raise the base and boost clocks by 100 MHz as long as your cooling is sufficient. Being Ryzen 7 octo-core CPUs they all share the same architecture, cache allotments and PCIe lanes. For a more detailed set of features and their breakdown jump over to ATMINSIDE’s review of the Ryzen 7 1800X.

Benchmarks

OK, let’s get into the meat of things and see what these two CPUs can do. I have an array of benchmarks lined up to compare the two Ryzen 7 CPUs to each other, the Intel i7 7700k, the Intel i7 5960x, and also the big 10 core 20 thread i7 6950x just to see how they compare. The first set of tests is stock out of the box with only the memory set to DDR4 3000 CL15-15-15-35. Now because of BIOS restrictions and the lack of a DDR4 3000 divider, memory on the AMD set-ups are done at 2933 CL15-15-15-35. That being said, let’s have a look at the hardware involved in our test set-ups.

The test System:

CPU Tests

• AIDA64 Engineer CPU, FPU, and Memory Tests
• Cinebench R11.5 and R15
• x265 1080p Benchmark (HWBOT)
• POVRay
• SuperPi 1M/32M
• WPrime 32M/1024M
• 7Zip

All CPU tests were run at their default settings unless otherwise noted.

AIDA64 Tests

Just a quick word here about the AIDA64 test suite. At the time of this writing AIDA64 didn’t fully support the new AMD Ryzen architecture. Therefore the test results were somewhat anomalous and inconsistent. As, such we will not be including those test results at this time.

CPU Benchmarks

Up first we’ll take a look at some compression and video rendering benchmarks. This suite of test can really leverage multi-core CPUs and the more cores the better. The first graph will show the CPUs at stock and the second will be all CPUs overclocked to 4.0 GHz taking away any advantage of a higher stock CPU speed. All percentages are based off of the Ryzen 7 1700X being 100% and the remaining CPUs are weighted against it.

Looking at results above and below we can see the 20 threads of the i7 6950x has a definite advantage in most tests and the Intel side walks all over 7Zip both stock and overclocked. One interesting point is the baby brother R7 1700 at stock appears to be right on par with the i7 5960x when both are running their stock configurations. Once overclocked all the AMD Ryzen 7 CPUs give the 6950x a pretty decent run for the money. Speaking of money, the Ryzen 7 1700X and Ryzen 7 1700 come in well below their Intel counterparts at $399.99 and $329.99 at Newegg.com as of today. Compare that to $1069.65 for an i7 5960x and a whopping $1649.99 for the i7 6950x, the AMD Ryzen 7 line appears to be a very good bang for the buck for anyone needing the multi-thread capabilities of these 16 thread CPUs.

I would like to note here that the Ryzen 7 1700X and Ryzen 7 1700 were run on the ASUS ROG Crosshair VI Hero and at this point that gave them a slight advantage with memory tuning. That will explain some of the differences you will see in the overclocked scores when compared to the Ryzen 7 1800X. In general though, memory options are limited at this time. More on this later.

Up next, 2D benchmarks in the form of SuperPi and wPrime.

As you can see in SuperPi stock or overclocked the Intels have the advantage here but not by much. Moving to the wPrime results, the Ryzen CPUs are pretty even with the i7 5960x. Just like our previous set of tests, cores count in this benchmark giving the i7 6950x a definite advantage but the lack of cores is pretty obvious looking at the i7 7700k. The Ryzen7 CPUs still made a good show. SuperPi and wPrime are both heavily affected by ram timings and performance. As I said earlier in the article, there is still room for improvement here and AMD is working hard to remedy this.

IPC Comparison

Just a quick word on IPC (Instructions Per Clock). With AMD’s last generation CPU this was a weak spot for them and they continually lagged behind Intel by a significant margin. I think it’s safe to say they have stepped up their game and closed that gap. As you will see, aside from 7Zip, Intel’s margin is pretty slim. I’ll let the next slide speak for itself.

Gaming Tests

All game tests were run at 1920×1080 and 2560×1440. Please see our testing procedures for details on in-game settings.

• 3DMark Fire Strike Extreme
• Crysis 3
• Dirt: Rally
• Ashes of the Singularity DirectX12
• Rise of the Tomb Raider DirectX12

On to the fun stuff

As you’ll see in the following slides, the gaming performance of the Ryzen 7 CPU is nearly identical across all three models which is to be expected. They also run toe to toe with the i7 6950x at 1080p and 1440p resolutions. I have substituted the Intel i7 7700k results at 4.0 GHz in place of the R7 1800X just for comparison. The 1800X scores are right in line with it’s counterparts and I felt this would be a better representation of gaming performance.

Looking at 3DMark Firestrike Extreme results, the overall scores are very close for all CPUs. The difference is in the physics tests where more cores is definitely an advantage as you can see by the results.

Realistically, I would assume users who adopt this platform gaming isn’t the main use of the system. This shows that when the real work is done, you can kick back and relax a bit. Gaming is still an option and won’t suffer in the least.

Information from AMD Regarding Overclocking

As a general guideline: a CPU voltage of up to 1.35 V is acceptable for driving everyday overclocks of the AMD Ryzen processor. Core voltages up to 1.45 V are also sustainable, but our models suggest that processor longevity may be affected. Regardless of your voltage, make sure you’re using capable cooling to keep temperatures as low as possible.

While there are never guarantees with overclocking, the majority of users should find that an eight core, sixteen thread, AMD Ryzen processor will achieve 4.2 GHz at a core voltage of 1.45 V. Advanced and accomplished overclockers trying to push record frequencies may find more headroom by disabling cores and/or disabling SMT on motherboards that offer these options in the BIOS.

There is no “stock” voltage value for Ryzen, due to how it operates, to base your starting voltage when manually setting vCore. As such, it also does not have a voltage table which it references for Auto settings.

Overclocking

So what did I see in the real world? I can say that neither the Ryzen 7 1700X or the two Ryzen 7 1700s that I have will hit 4.2 GHz at 1.45v. I tried with air cooling and my custom water loop. It wasn’t a temperature limitation that I was hitting. I was a long way from the max temperature of 95 °C which is where these CPU’s will throttle. As you can see below these were max temperatures for a 4.0 GHz overclock on each of the Ryzen 7 CPUs.

No, for me it was a voltage limitation. The Ryzen 7 1700X would run at 4.0 GHz within AMDs recommendation of 1.45v , the Ryzen 7 1700 on the other hand was up to 1.48v for the same overclock. I found that between stock and 3.9 GHz voltage increases were quite moderate. After that mark and quite noticeably with the Ryzen 7 1700, the voltage jump to push to the 4.0 GHz mark was huge. I had a stable overclock at 3.9 GHz with 1.35 v, getting over the next 100 MHz hump required a 0.13 v increase. Needless to say my overclocking adventure was pretty short lived. What I did do was spend a small amount of time working on the memory.

AMD Ryzen 7 and DDR4

The new Ryzen 7 CPU, officially supports memory in the following configurations.

As of this writing most motherboards have very limited DRAM settings in BIOS and a locked or limited base clock adjustment. Hopefully in the near future this will be relaxed a bit. Word on the street is that up until launch AMD was mainly focused on CPU performance and that updates to the AGESA code “should” bring improvements in memory performance and relax limitations in the BIOS. Now that being said, it’s not impossible to improve the RAM performance as it is now. I’ll include a few screen shots below to show the benefits of improving RAM performance. I’m not going into the mechanics of it here but will cover that in my upcoming overclocking guide.

Here’s a shot of SuperPi 32M at the last timings just to give an idea of the improvements.

Final Thoughts

To start, I certainly am relieved that after all the hype leading up to this launch we weren’t handed another Bulldozer. I don’t think I’m alone in thinking that many in the PC world were secretly rooting for AMD to bring some sorely needed competition back to the desktop PC platform and from where I sit today, they certainly have. The Ryzen 7 CPU is a real workhorse and for the money you really can’t beat it. The price to performance ratio here is incredible when compared to its competition, considering its a third of the price.

That being said, I still see some room for improvement. It’s not really “plug and play” at the moment and requires some attention when setting up. Add to that some memory compatibility issues leaving end users a bit frustrated and confused. My recommendation for any one looking to jump into the AM4 platform at present, would be to stick with some low speed DDR4 2666 or even look for one of the new kits intended for this platform being put out by G.Skill – the FlareX and Fortis lines.

Overall, I feel this is a big win for AMD despite a few growing pains that can accompany any new platform. They really came through with their performance and delivered a dramatic gain in IPC when compared to their last generation. The drop in TDP is outstanding coming in at 95 W or less, when comparable platforms are 140 W. Top that off with pricing that makes choosing an 8 core 16 thread CPU pretty straightforward and easy on the wallet. Overclockers Approved!

– Shawn Jennings (Johan45)