AMD is back again with the launch of their second generation of the ZEN core, the Ryzen 7 2700X and the Ryzen 5 2600X, which AMD refers to as ZEN+. The ZEN+ microarchitecture is a scaled down and a slightly tweaked version of the original Summit Ridge microarchitecture launched last year and is code-named “Pinnacle Ridge”. AMD had a few goals for the new and improved Ryzen CPU such as raising the clock speeds, lowering latency, and improving memory speeds/compatibility.
According to AMD, the ZEN+ core offers some targeted improvements to help optimize latency-sensitive tasks. Those changes include:
- A reduction in L1 Cache latency of approximately 13%
- A reduction in L2 Cache latency of approximately 34%
- A reduction in L3 Cache latency of approximately 16%
- Contributing to a single threaded IPC improvement of approximately 3%
- Official support for JEDEC DDR4-2933 (up from 2667)
In addition to the ZEN+ architecture, the second generation of AMD’s Ryzen CPUs also takes advantage of Global Foundries’ 12LP (Leading Performance) process. Global Foundries 12LP offers 10-15% improvement in transistor performance which extends the clock range and reduces current draw versus the 14nm process used in the last generation of Ryzen CPUs. This process offers the AMD Ryzen CPUs the following:
- +300 MHz clock speed, now reaching up to 4.3 GHz in everyday operation
- A 50 mV reduction in core voltage across the operating range versus 14 nm
- All-core overclocks now in the range of 4.2 GHz
Together, the ZEN+ architecture using the Global Foundries 12LP process has produced a new second generation of the Ryzen CPU that’s measurably faster than its predecessor in all the ways we wanted: clock speed, latencies, DRAM, and overclocking. Further on in this review I’ll be putting this to the test and see how paper measures up to reality.
Specifications and Features
Looking at the specifications table below, we see the new Ryzen+ CPU is produced using the 12 nm FinFET process and has 4.8 billion transistors on a 213 mm² die. AMD is still using solder between the die and IHS on the Ryzen CPUs for improved thermal transfer. Both CPUs are equipped with two CPU Complexes (CCX) with 16 MB shared L3 Cache and 512 MB L2 Cache per core.
The Ryzen 7 2700X has a base frequency of 3.7 GHz and a maximum boost of 4.35 GHz with a 105 W TDP. This is up slightly from its predecessors when comparing to the Ryzen 7 1700X’s base frequency of 3.4 GHz and a 95 W TDP. During stress testing, I noticed an all core, heavy load, boost which hovered between 3.95 GHz and 4.0 GHz. The Ryzen 5 2600X has a base frequency of 3.6 GHz and a maximum boost of 4.25 GHz with a 95 W TDP. Again I observed fairly high frequencies during stress testing with all cores hovering between 3.95 GHz and 4.05 GHz this time, slightly higher than the 2700X. AMD’s new Precision boost 2.0 definitely leverages more of the CPU when comparing these to the original Ryzen CPUs.
The memory situation last year was a bit of a rough spot for the Ryzen launch, which has improved steadily as the BIOS matured. AMD has also improved the memory situation as the new Ryzens officially support memory speeds of DDR4 2933 MHz – up from the 2667 MHz Summit Ridge CPUs officially supported. I can tell you this much, I have tried the G.Skill FlareX 3200 MHZ and the G.Skill SniperX 3400 MHz in both motherboards that AMD supplied. Both sets of RAM would boot at rated XMP speeds/timings and they ran stable for over 30 minutes using AIDA64’s stability test. This was in the MSI X470 Gaming M7 AC as well as the Gigabyte X470 Aorus Gaming 7 WiFi.
Windows 10 is the officially supported platform for the Ryzen CPUs and it does appear as if Windows 7 installations will again be possible with the right drivers.
Specifications below supplied by AMD.
|CPU||AMD Ryzen 7 2700X||AMD Ryzen 5 2600X|
|# of Cores||8 (2 CCX: 4+4)||6 (2 CCX: 3+3)|
|# of Threads||16||12|
|Base Clock Speed||3.7 GHz||3.6 GHz|
|Boost Clock Speed||4.35 GHz||4.25 GHz|
|Instruction Set Extensions||SSE 4.1/4.2/4a, AVX2, SHA||SSE 4.1/4.2/4a, AVX2, SHA|
|Lithography||12 nm FinFET||12 nm FinFET|
|Transistor Count||4.8 billion||4.8 billion|
|TDP||105 W||95 W|
|Thermal Solution Spec||Solder||Solder|
|L1 Cache||64 KB I-Cache|
32 KB D-Cache per Core
|64 KB I-Cache|
32 KB D-Cache per Core
|L2 Cache||4 MB (512 KB per core)||3 MB (512 KB per core)|
|L3 Cache||16 MB Shared||16 MB Shared|
|Max Memory Size||128 GB||128 GB|
|# of Memory Channels||2||2|
|ECC Memory Support||No||No|
The table below is a list of the second generation Ryzen desktop CPU lineup. As I mentioned in the Sneak Peek article AMD has limited the number of available SKUs this time around. You won’t see a 2800X and the four core CPUs have been replaced by the new APUs with Radeon Vega graphics. All CPUs are overclockable, assuming you buy a motherboard with a chipset capable of doing so.
|Ryzen 7 2700X||Ryzen 7 2700||Ryzen 5 2600X||Ryzen 5 2600|
|MSRP||USD $329||USD $299||USD $229||USD $199|
|Silicon||12 nm “Pinnacle Ridge”|
|Clock Speed||3.70 GHz||3.20 GHz||3.60 GHz||3.40 GHz|
|Boost Speed||4.35 GHz||4.10 GHz||4.25 GHz||3.90 GHz|
|Cooler||Wraith Prism||Wraith Spire||Wraith Spire||Wraith Stealth|
|L2 Cache||512 KB per core|
|L3 Cache||16 MB shared|
|New Features||XFR 2.0 and Precision Boost 2.0|
|TDP||105 W||65 W||95 W||65 W|
|Memory||Dual-Channel DDR4-2933 JEDEC up to 64 GB|
|PCIe||PCIe Gen 3.0 x16 PEG (x16 or x8 + x8) + x4 M.2 + x4 Chipset|
|SoC Connectivity||2xSATA 6 Gbps, 4x USB 3.1, 1xM.2- PCIe 32 Gbps|
|Chipset||AMD 300 and 400 Series|
The following information was provided by AMD:
Practical Characterizations of “Zen+” and 12LP
With the above changes in mind, we can now begin to explore the practical implications in production 2nd Gen AMD Ryzen™ processors. First, we’ll look at analogous “Summit Ridge” and “Pinnacle Ridge” parts operating at the same clock speed: here we see the 12LP’s improved transistor performance manifesting as ~11% better power per clock—right in the 10-15% expected range.
Next, we can evaluate the YoY performance improvement for consumers by comparing two analogous processors with the same TDP and power envelopes at stock settings. In such a comparison, we see a fantastic performance improvement of up to 16%.
Finally, the combination of “Zen+” and 12LP makes it possible for AMD to explore product definitions that simply weren’t possible with “Zen” and the 14nm process. That headroom allows AMD to meet consumer appetite for faster flagship processors in the AM4 Socket.
This year, that “maximum performance” product is the AMD Ryzen 2700X with a fantastic new performance threshold and 105W TDP. Thanks to performance-maximizing decisions like Precision Boost 2 and a new, higher 4.3GHz clock, the 2700X is approximately +12% faster than the 1800X in multi-thread scenarios. When fewer threads are in flight, Precision Boost 2 also shines, driving as much as 500MHz extra clock speed that the 1800X couldn’t capture. Such decisions naturally have electrical and thermal implications, but the Socket AM4 infrastructure was designed for this moment.
Overall, we know consumers will be pleased by the range of offerings enabled by “Zen+” and 12LP. Whether your purchasing goals revolve around maximizing performance or maximizing power efficiency, the 2nd Gen AMD Ryzen Processor family truly offers something for everyone.
The “Zen” X86 Microarchitecture
On the performance side, the Zen micro-architecture represents a quantum leap in core execution capability versus AMD’s previous desktop designs. Notably, the Zen architecture features a 1.75x larger instruction scheduler window and 1.5x greater issue width and resources; this change allows Zen to schedule and send more work int the execution units. Further, a micro-op cache allows Zen to bypass L2 and L3 cache when utilizing frequently-accessed micro-operations. Zen also gains a neural network-based branch prediction unit which allows the Zen architecture to be more intelligent about preparing optimal instructions and pathways for future work. Finally, products based on the Zen architecture may optionally utilize SMT to increase utilization of the compute pipeline by filling app-created pipeline bubbles with meaningful work.
A high-performance engine requires fuel, and the Zen architecture’s throughput characteristics deliver in this regard. Chief amongst the changes are major revisions to cache hierarchy with dedicated 64 KB L1 instruction and data caches, 512KB dedicated L2 cache per core, and 8 MB of L3 cache shared across four cores. This cache is augmented with a sophisticated learning prefetcher that speculatively harvests application data into the caches so they are available for immediate execution. Altogether, these changes establish lower level cache nearer to the core netting up to 5x greater cache bandwidth into a core.
Beyond adopting the more power efficient 12 nm FinFET process, the overall Zen architecture also incorporates AMD’s latest low power design methodologies, such as: the previously mentioned micro-op cache to reduce power-intensive faraway fetches, aggressive clock gating to zero out dynamic power consumption in minimally utilized regions of the core, and a stack engine for low-power address generation into the dispatcher.
It is in this realm, especially, that the power management wisdom of AMD’s APU teams shines through to impart in Zen the ability to scale from low-wattage mobile to HEDT configurations.
Scalability in the Zen architecture starts with the CPU Complex (CCX), a natively four core eight thread module. Each CCX has 64 KB L1 I-Cache, 64 KB L1 D-Cache, 512 KB dedicated L2 cache per core, and 8 MB L3 cache shared across cores. Each core within the CCX may optionally feature SMT for additional multi-threaded capabilities.More than one CCX can be present in a Zen-based product, wherein the AMD Ryzen processor features two CCX’s consisting of eight cores and 16 threads (total). Individual cores within the CCX may be disabled by AMD, and the CCX’s communicate across the high-speed Infinity Fabric. This modular design allows AMD to scale core, thread, and cache quantities as necessary to target the full spectrum of the client, server, and HPC markets.
The Infinity Fabric, meanwhile, is a flexible and coherent interface/bus that allows AMD to quickly and efficiently integrate a sophisticated IP portfolio into a cohesive die. These assembled pieces can utilize the Infinity Fabric to exchange data between CCX’s, system memory, and other controllers (e.g. memory, I/O, PCIe) present on the AMD Ryzen SoC design. The Infinity Fabric also gives Zen architecture powerful command and control capabilities, establishing a sensitive feedback loop that allows for real-time estimations and adjustments to core voltage, temperature, socket power draw, clock speed, and more. This command and control functionality is instrumental to AMD SenseMI technology.
AMD SenseMI Technology
First and foremost, it is important to understand that each AMD Ryzen processor has a distributed “smart grid” of interconnected sensors that are accurate to 1 mA, 1 mV, 1 mW, and 1 °C with a polling rate of 1000/sec. These sensors generate vital telemetry data that feed into the Infinity Fabric control loop, and the control loop is empowered to make real-time adjustments to AMD Ryzen processor’s behavior based on current and expected future operating conditions.
AMD SenseMI is a package of five related “senses” that rely on sophisticated learning algorithms and/or the command-and-control functionality of the Infinity Fabric to empower AMD Ryzen processors with Machine Intelligence (MI). This intelligence is used to fine-tune the performance and power characteristics of the cores, manage speculative cache fetches, and perform AI-based branch prediction.
The distributed network of smart sensors that drive Precision Boost can do double duty to streamline processor power consumption with any given workload. And for next-level brilliance: telemetry data from the Pure Power optimization loop allows each AMD Ryzen processor to inspect the unique characteristics of its own silicon to extract individualized power management.
NEW:Precision Boost 2
Starting with the AMD Ryzen desktop CPU, AMD unveiled “Precision Boost”: a telemetry-aware DVFS technology that could adjust CPU frequencies with an industry-leading granularity of 25 MHz. The Ryzen 1000 Series desktop CPU employed this boost algorithm in two discrete states
- Three or greater CPU threads (all core boost)
- Two or fewrwer CPU threads (Precision Boost)
After the unveiling of Precision Boost and the AMD Ryzen desktop processor, AMD has observed scenarios where 3+ cores are in use, yet the overall size of the workload is relatively small. This creates a scenario where the “all core boost” state is triggered, even though there is no imminent electrical, thermal, or utilization boundary that would practically halt further clock speed increases.
This scenario represents additional opportunity to drive higher performance. The thermal, electrical, and utilization headroom of the product can be converted into higher clock speeds to capitalize on the opportunity.
Precision Boost 2 carries forward the 25 MHz granularity of its predecessor, but importantly transitions to an algorithm that will intelligently pursue the highest possible frequency until an aforementioned limit is encountered, or the rated frequency of the part is met (whichever comes first). This applies to any number of threads in flight, without arbitrary limits. Precision Boost 2 could be described as opportunistic, linear, or graceful, and a conceptual comparison of Precision Boost 1 VS 2 has been plotted for clarity below.
If a power or temperature limit is encountered, Precision Boost 2 is designed to employ its granular clock selection to dither at a frequency within the power or temperature limits. This process is a continuous adjustment loop managed by the AMD Infinity Fabric, and it cycles up to 1000 times per second. A real-world example of this is shown below with OCCT, where the boost gracefully transitions across one to eight threads and then maintains a max-thread clock speed well above the base. Taken as a whole, the Precision Boost 2 invests the AMD Ryzen Processor with Radeon Vega Graphics with greater performance in real-world multi-threaded applications by freeing the CPU to make the most performant clock selection for its defined electrical/thermal/load/frequency capacity — regardless of the number of threads in flight. On the AMD Ryzen 7 2700X, for example, users could see clock speeds in modestly-threaded tasks (e.g. games) that are ~500 MHz higher than what was possible on the Ryzen 7 1800X.
NEW: Extended Frequency Range 2 (XFR2) By now you understand that Precision Boost 2 is an opportunistic algorithm governed by built-in limits concerning temperature, current, and maximum clock speed. Those AMD-specified limits are what inform the official processor specifications that are published and ensure consistent performance. By their very nature, these specifications must be somewhat conservative to accommodate sub-optimal but inevitable conditions like warm and humid climates. But the thermal environment is an easy variable for the user to influence: large heatsinks, high-airflow cases, air-conditioned rooms, AMD believes that the user should be rewarded for exceeding recommended thermal specification with more performance, and that is achieved through the XFR2 capability. A superior-than-recommended thermal environment effectively extends the time the processor can boost dock speeds before encountering a thermal boundary, and XFR2 will let the AMD Ryzen processor run at a higher average frequency as a result. Whereas this capability was restricted to a small number of cores with the 1st Gen AMD Ryzen, XFR2 now operates across any number of cores and threads – just like Precision Boost 2. As a result, XFR2 can enable up to 7% additional processor performance even when every core and thread is loaded.
- Neural Net Prediction
A true AI inside every AMD Ryzen processor harnesses a neural network to do real-time learning of an application’s behavior and speculate on its next moves. The predictive AI readies vital CPU instructions so the processor is always primed to tackle a new workload.
Sophisticated learning algorithms understand the internal patterns and behaviors of applications and anticipate what data will be needed for fast execution in the future. Smart Prefetch predicatively pre-loads that data into large caches on the AMD Ryzen processor to enable fast and responsive computing.
SMT (Simultaneous Multi-Threading)
This is AMD’s new equivalent to Intel’s HyperThreading (HT) technology. It allows each core to function as two threads, adding performance in multi-threaded applications.
Every Processor is Unlocked
AMD is allowing overclocking on all CPU models, much as they have in the past. The only caveat this time around is you must have a motherboard with a chipset supporting overclocking.
Here are a couple more slides from AMD, first up is a shot of the ZEN+ die.
Next, we have AMD’s predicted rollout for the remainder of 2018 and their second generation, Zen+ based CPUs
Precision Boost 2
Just a few words on my observations of AMD’s improved boost function. First off, it does behave differently than their first iteration of Precision Boost. The 2700X has a base clock of 3.7 GHz and boost to 4.35 GHz with XFR2 and from what I observed, the CPU would boost to 3.95-4.0 GHz under heavy loads such as AIDA64 stability, Cinebench R15 or HWBot x265. During single threaded operations, it would boost one or more cores up to 4.35 GHz but the load appeared to move between different cores, it almost seemed erratic. I retested and set affinity to a single core so it was the only one that boosted, that core stayed at 4.35 GHz during the full test but the benchmark scored the same so the stock behavior didn’t affect the outcome. The Ryzen 5 2600X displayed very similar behavior but a slightly higher all core boost up to 4.05 GHz under heavy loads and only 4.25 GHz with lightly threaded loads.
Precision Boost Overdrive
This is a nice addition that AMD decided to incorporate in the second generation of Ryzen CPUs. Overdrive allows the user, through the BIOS, to raise the all core boost function. The maximum I could achieve with the Ryzen 7 2700X was 4.15 GHz all core boost, I am assuming there were other factors such as temperatures limiting me since this was done on the AMD supplied Wraith Prism. Despite the fact that the stock cooler was being used, the Ryzen 7 2700X held stable under the AIDA64 stability test for 30 minutes. I should say I hadn’t altered any voltages in BIOS for this either, this was all done automatically by only raising the Precision Boost Overdrive setting.
I covered the review kit from AMD in the “Unboxing” article, you can follow this link for more details but I’ll also include a quick rundown of the components here.
The reviewer’s kit included the following
- AMD Ryzen 7 2700X CPU with Wraith Prism CPU cooler
- AMD Ryzen 5 2600X CPU with Wraith Spire CPU cooler
- MSI X470 Gaming M7 AC Motherboard
- Gigabyte X470 Aorus Gaming7 WiFi Motherboard
- G.Skill SniperX 2 x 8 GB DDR4 3400 CL16
I have tested all the components and they all work as they should. The SniperX RAM would boot at XMP and run stable on both motherboards. The majority of my testing was done on the Gigabyte Aorus, the MSI Gaming had a minor BIOS issue when I received it which has since been fixed. I was well into testing by then and wanted to keep things consistent so I stuck with the Aorus. As usual, I would advise an early adopter of the X470 platform to update their BIOS first thing to avoid such issues.
All benchmarks were run with the motherboard being set to optimized defaults (outside of some memory settings which had to be configured manually). When “stock” is mentioned along with the clock speed, this does not include any boost or turbo options. I disabled these in the BIOS to keep core clocks consistent and remove any variance that the boost options may introduce. I would also like to add that both AMD CPUs were tested with their respective, included Wraith coolers. This is also true for all tests performed at 4.0 GHz later in the head to head benchmarks and gaming benchmarks. One final note here, I also updated all the testing for the Ryzen 7 1700X to reflect the BIOS and driver improvements AMD has made since the launch in 2017.
|AMD Ryzen7 2700X||AMD Ryzen 5 2600X||AMD Ryzen 7 1700X||Intel i7- 8700K|
|Motherboard||Gigabyte X470 Aorus Gaming7 WiFi||Gigabyte X470 Aorus Gaming7 WiFi||ASUS X370 ROG Crosshair VI Hero||ASUS ROG Strix X370-E Gaming|
|Memory||G.Skill FlareX 2×8 GB DDR4-3200 MHz 14-14-14-34||G.Skill FlareX 2×8 GB DDR4-3200 MHz 14-14-14-34||G.Skill FlareX 2×8 GB DDR4-3200 MHz 14-14-14-34||G.Skill FlareX 2×8 GB DDR4-3200 MHz 14-14-14-34|
|Graphics Card||ASUS ROG GTX 1080 Ti||ASUS ROG GTX 1080 Ti Strix||ASUS ROG GTX 1080 Ti Strix||ASUS ROG GTX 1080 Ti Strix|
|HDD||Samsung 512 GB 950 Pro||Samsung 512 GB 950 Pro||Samsung 120 GB 840 EVO||Samsung 512 GB 950 Pro|
|Game Storage||Samsung T5 1 TB Portable SSD||Samsung T5 1 TB Portable SSD||Samsung T5 1 TB Portable SSD||Samsung T5 1 TB Portable SSD|
|Power Supply||Super Flower 1000W Platinum||Super Flower 1000W Platinum||Super Flower 1000W Platinum||Super Flower 1000W Platinum|
|Cooling||AMD Wraith Prism||AMD Wraith Spire||EKWB 360 Predator XLC||EKWB 360 Predator XLC|
|OS||Windows 10 x64||Windows 10 x64||Windows 10 x64||Windows 10 x64|
- AIDA64 Engineer CPU, FPU, and Memory Tests
- Cinebench R11.5 and R15
- HWBot x265 1080p Benchmark
- SuperPi 1M/32M
- WPrime 32M/1024M
All CPU tests were run at their default settings unless otherwise noted.
All game tests were run at 1920×1080 and 2560×1440 with all CPUs at 4.0 GHz. Please see our testing procedures for details on in-game settings.
- 3DMark Fire Strike Extreme
- Middle Earth: Shadow of Mordor
- Metro Last Light
- Ashes of the Singularity
- Rise of the Tomb Raider
I have to add here that all testing was done with the HPET (High Precision Event Timer) active. It has recently come to light that having HPET active on the Intel platform can have an adverse effect on the gaming results in particular. Since the average user isn’t likely to activate this feature in Windows 10, I will be retesting the i7-8700k without having the HPET active to compare the results. I will include these results in the Ryzen 7 2700 and Ryzen 5 2600 review which is coming in the next few weeks.
Just a note here, I used the latest AIDA64 Engineer Beta for testing but it also needed an update to work with the new X470 motherboards. Thank you to the crew at Finalwire.
|AIDA64 Cache and Memory Benchmark|
|Ryzen 7 2700X @ 3.7 GHz||49287||47651||45392||68.9|
|Ryzen 5 2600X @ 3.6 GHz||49128||47767||45356||69.8|
|Ryzen 7 1700X @ 3.4 GHz||48648||48499||44929||74|
|Intel i7-8700K @ 3.7 GHz||42675||45393||42191||46.8|
As you can see the Ryzen is working much better with ram than it was a year ago but that latency is still quite high when compared to Intel. Up next the AIDA64 CPU benchmarks.
|AIDA64 CPU Tests|
|Ryzen 7 2700X @ 3.7 GHz||85853||23874||699,8||66210||22338|
|Ryzen 5 2600X @ 3.6 GHz||65315||22464||505.9||48288||16297|
|Ryzen 7 1700X @ 3.4 GHz||78723||23595||618.6||60045||20502|
|Intel i7-8700K @ 3.7 GHz||62220||25397||493.2||25290||6362|
The CPU tests show the Ryzen 5 2600X is holding up well compared to the i7-8700k even with a 100 MHz speed disadvantage.
|AIDA64 FPU Tests|
|Ryzen 7 2700X @ 3.7 GHz||7475||38553||20210||12676|
|Ryzen 5 2600X @ 3.6 GHz||7439||28132||14741||9248|
|Ryzen 7 1700X @ 3.4 GHz||7410||35360||18531||11633|
|Intel i7-8700K @ 3.7 GHz||7659||47205||25401||5402|
The floating point tests seem to be a bit of a weak spot for the Ryzen-based CPUs even with the extra threads they were left behind in all but the SinJulia test.
Real World Tests
Next, we will move on to something a bit more tangible/productivity-based with compression, rendering, and encoding benchmarks.
|Cinebench R11.5/R15, POVRay, x265 (HWBot), 7Zip – Raw Data|
|Ryzen 7 2700X @ 3.7 GHz||17.8||1691||3368.94||43.31||43912|
|Ryzen 5 2600X @ 3.6 GHz||13.31||1254||2566.16||31.85||34731|
|Ryzen 7 1700X @ 3.4 GHz||16.75||1521||3118.42||38.24||39261|
|Intel i7-8700K @ 3.7 GHz||13,42||1239||2593,15||42.57||36116|
Here again, the extra threads gave the Ryzen 7 CPUs a bit of an advantage over the other CPUs in all but HWBot’s X265 benchmark. The last generation of Intel CPUs got a real boost in this benchmark when compared to their predecessors. The Ryzen 5 is holding its own against the 8700K which is nice to see.
Moving on from all the multi-threaded goodness above, we get to some Pi and Prime number based tests. SuperPi and WPrime, specifically.
|SuperPi and wPrime Benchmarks – Raw Data|
|CPU||SuperPi 1M||SuperPi 32M||wPrime 32M||wPrime 1024M|
|Ryzen 7 2700X @ 3.7 GHz||11.454||612.617||3.466||92.37|
|Ryzen 5 2600X @ 3.6 GHz||11.517||631.833||4.358||126.839|
|Ryzen 7 1700X @ 3.4 GHz||12.237||678.93||3.857||99.373|
|Intel i7-8700K @ 3.7 GHz||9.907||533.12||4.205||116.249|
Clock speeds and thread counts play a big role in these last tests. The results will be better interpreted when they are all running the same speed.
As far as the games go, tests were done at 1920 x 1080p and 2560 x 1440p according to our Graphics Testing Procedure which was linked earlier. The CPUs are all at 4.0 GHz with the 3200 MHz FlareX to keep the field as even as possible.
As you can see above in the gaming results, AMD has made some nice improvements here over the last generation of Ryzen CPUs. In the 1080p tests, the gains were over 15% in Ashes of the singularity and on average in the 10% range for the other three tests. The 2700X and 2600X even managed to beat the i7-8700K in all but the Ashes test as well. The 1440p tests closed the gap a bit but there were still impressive gains across all the tests.
On to the synthetic benchmark, 3DMark Fire Strike, you can see the results are very close across the board except for the Physics test. The added cores of the 1700X and 2700X gave them a big advantage here but even the 2600X edged past the i7-8700K on an even playing field.
Head to Head
In this round of testing all CPUs were set to a static frequency of 4.0 GHz and all were using the same FLAREX, 3200 MHz RAM from G.Skill. This was done to even the playing field as much as possible and give a realistic impression of performance at matched speeds regardless of what a CPUs maximum frequency potential may be. Keep in mind the AMD Ryzen 2600X and Intel i-7 8700K are both six-core, twelve thread CPUs making them the most closely matched out of the four.
So let’s see how they stacked up.
|AIDA64 CPU Tests|
|Ryzen 7 2700X @ 4.0 GHz||92700||24624||753.1||71054||24147|
|Ryzen 5 2600X @ 4.0 GHz||72554||23314||561.4||53676||18109|
|Ryzen 7 1700X @ 4.0 GHz||90805||24367||717.6||69652||23782|
|Intel i7-8700K @ 4.0 GHz||66886||27302||530||27186||6839|
The Ryzen CPUs have always been fairly strong in the AIDA64 CPU tests, except for PhotoWorxx, that’s where even the lower core count couldn’t hold the 8700K back. Aside from that, the Ryzen 2600K gave it a good run for the money taking 4/5 of the CPU tests from its Intel competitor. You will notice as well that the second generation eight-core Ryzen is slightly ahead of its predecessor throughout all the testing here and on into the FPU tests.
The FPU testing, on the other hand, is where the 8700K really digs in and shows its strength. Despite the fact that it has a thread disadvantage, the 8700K topped the eight-core Ryzens by over 20% in the Julia and Mandel tests exposing one of the weaknesses in the Ryzen architecture. For everyday tasks, this really doesn’t seem to hold the AMD Ryzen back though. You can also see some weak spots across both sets of AIDA64’s test suites where the Intel CPU is scoring much lower than its Ryzen competitors.
|AIDA64 FPU Tests|
|Ryzen 7 2700X @ 4.0 GHz||8595.6||41017||21287||13494|
|Ryzen 5 2600X @ 4.0 GHz||8071||31244||16375||10278|
|Ryzen 7 1700X @ 4.0 GHz||7866||40798||20649||13197|
|Intel i7-8700K @ 4.0 GHz||8233||50745||27306||5807|
In our compression, rendering, and encoding tests you can see the 2600X putting up a good fight against the i7-8700K once again. All the tests look as they should with Intel taking the crown for the HWBot X265 benchmark.
|Cinebench R11.5/R15, POVRay, x265 (HWBot), 7Zip – Raw Data|
|Ryzen 7 2700X @ 4.0 GHz||19.51||1828||3711.25||45.01||46403|
|Ryzen 5 2600X @ 4.0 GHz||14.78||1385||2841.3||34.91||37265|
|Ryzen 7 1700X @ 4.0 GHz||19.63||1769||3643.98||44.33||44333|
|Intel i7-8700K @ 4.0 GHz||14.42||1303||2792.3||45.42||38505|
The SuperPi benchmarks have never been a strong point for AMD and again this shows here, on the other hand, in WPrime the six-core Ryzen is only slightly trailing the 8700K. The extra cores of the 1700X and 2700X give them an obvious advantage in this last benchmark.
|SuperPi and wPrime Benchmarks – Raw Data|
|CPU||SuperPi 1M||SuperPi 32M||wPrime 32M||wPrime 1024M|
|Ryzen 7 2700X @ 4.0 GHz||10.52||581.588||3.109||86.122|
|Ryzen 5 2600X @ 4.0 GHz||10.532||566.701||3.92||114.314|
|Ryzen 7 1700X @ 4.0 GHz||10.42||586.477||3.766||85.06|
|Intel i7-8700K @ 4.0 GHz||9.204||477.977||3.89||108.124|
AMD Ryzen 7 2700X @ 4.0 GHz Screenshots
AMD Ryzen 5 2600X @ 4.0 GHz Screenshots
IPC and SMT
AMD held true and did improve the IPC (Instructions per Clock) of the second generation Ryzen CPUs. Comparing the 2700x to the 1700x in this limited amount of testing averages out to 3.15% which is right where AMD placed it. Comparing Ryzen+ to the competition puts them slightly behind the 8700K but not by much and these numbers do not include SMT, which, when compared to Intel’s HT is more efficient in most scenarios as you’ll see below.
The chart below compares AMD’s SMT efficiency to Intel’s HT as well as the previous generation of Ryzen. The results speak for themselves really, AMD is way ahead in all but the 7Zip test. One thing you will notice is the SMT efficiency has dropped a bit when compared to the original ZEN core. I’m assuming in favor of the Cache tweaks to improve latency and IPC.
Power Consumption and Temperatures
In the graph below we tested power use of the system across multiple situations from idle, to Prime 95 Small FFT (with FMA3/AVX). The 2700X and 2600X systems, both pulled the most power during the Prime 95 small FFT test with 225 W and 185 W respectively. Keep in mind this is full system power usage.
Temperatures were surprisingly well-controlled with the included Wraith coolers, I saw no throttling at any point. The highest temperature when at 4.0 GHz was 85 °C with the Ryzen 5 2600X paired with the Wraith Spire cooler, during Prime95 Small FFT. This shows that the stock cooler is adequate for the 2600X at 4.0 GHz settings. For the Ryzen 7 2700X, the Wraith Prism did a great job at full load and 4.0 GHz slightly passing 78 °C during P95 small FFT testing. I was also able to achieve a stable (30 minutes AIDA64) 4.2 GHz with the 2700X/Wraith Prism combination. I tried the same with the 2600X/Wraith Spire and it was falling short with improved cooling I was able to run at 4.2 GHz stable.
Pushing the Limits
Now the moment you’ve all been waiting for, “so how fast will they go?” I have a few pictures to post here, I spent most of my limited time working with the Ryzen 7 2700X, but for this adventure, I switched over to my EKWB Predator 360 XLC closed loop cooler for improved cooling which took me up to 4.3 GHz and fairly stable at that. I was able to run the AIDA64 stability test for over an hour, because of time restraints that was all the stability testing I did, but it gives you an idea of what these CPUs are capable of. At 4.3 GHz, I was also at my voltage limit of 1.45 V but I’m certain with a bit more time that frequency would be completely stable for 24/7 usage. I’m also including a screenshot at 4.2 GHz that was done with the Wraith Prism cooler and the G.Skill SniperX 3400 MHz running at 3666 MHz. The last screenshot is all out tuning at 4.3 GHz on the MSI Motherboard, I have the FlareX 3200 MHz running at 3733 MHz with tight timings.
Just one more screenshot, this one is really all out. If there are any that haven’t figured this out yet, I’m an extreme bencher at heart. I have a custom loop consisting of a Phobya 1260 SuperNova, a 750 GPH pump, and a Koolance 380A block modified to fit the AM4 socket. To top it off the liquid is chilled well below ambient, let’s just say in the snowflake range. This is what I use for testing before I go to LN2 in order to get an idea of how hardware will scale with colder temperatures. The hardware is the same with the same settings as the previous screenshot but I have upped the CPU speed to 4.55 GHz and bumped the voltage to 1.5 V for the CPU core. It needed that extra 0.05V to go from 4.5 GHz to 4.55 GHz, 4.6 wasn’t going to happen at that voltage.
I need to add a disclaimer here. What I have done is well out of the normal operating parameters of this hardware. If you try this at home you most likely will break something, so here it comes “Don’t try this at home” this will definitely void your warranty!
With a bit more time I was able to push the Memory envelope of the second generation Ryzen to new heights. This falls into the same category as the last screenshot as I was using excessive DRAM voltage to get the RAM to such a low CL at a speed of 4000 MHz that’s right 4000 MHz. With some work it may be possible to run this speed 24/7 but it will take a lot of tuning to do so.
I would like to mention that both CPUs behaved nearly identical when it came to voltage requirements. Both CPUs were stable at 4.0 GHz and 1.3V set in BIOS, 4.2 GHz with 1.4V and 4.3 GHz with 1.45V set in BIOS. Anyone familiar with the Ryzen platform will also know about the V-Droop associated with it. On both motherboards that is still present but easily addressed with some LLC (Load Line Calibration) tuning.
Overall, the performance of AMD’s second-generation Ryzen CPU is quite impressive, leveraging higher clock speeds and efficiency from the 12nm process improvements. To me, it appears as if they have accomplished all their goals. Overall speed has increased by about 300 MHz, I know it’s not groundbreaking but it is an improvement considering some CPUs were topping out at 3.8-3.9 GHz on the previous generation. There has definitely been an improvement in latency. This is quite apparent by the boost the new Ryzen has gotten in tests such as 7Zip and even more so in the gaming tests where we saw roughly a 15% improvement at 1080p. The RAM situation also seems to have improved, they still run anything on “Samsung B die” with ease but there are also new alternatives such as the G.Skill SniperX which isn’t based on the same IC although it is Samsung. My testing of Hynix or Micron based kits was limited by the fact that I don’t have any so I was unable to check any compatibility improvements there.
AMD has delivered on their promises for Ryzen+, it does run faster and more efficiently showing some nice improvements on the ZEN microarchitecture. The new Ryzen CPU should be widely available by the time anyone is reading this review. Currently, they are on pre-order at Newegg.com, the Ryzen 7 2700X is listing at $329.99 and the Ryzen 5 2600X is available for $229.99. The new Ryzen non-X CPUs are also available for ~ $30 less than their “X” counterparts. Comparing these prices to the Intel i7-8700K at $349.99 makes eight-cores for less than the price of six an attractive offer which has always been another one of AMD’s strong points.
A bit more overclocking headroom and improved performance, plus an improved Precision Boost leveraging additional performance at stock settings. Pairing these with an improved 400-series chipset on a new line of motherboards makes the Ryzen 2000 series CPUs very competitive. One added bonus is that the growing pains seen with the original Ryzen launch have all been overcome offering the end-user a rewarding, stress-free experience. A big thumbs up and Overclockers Approved!
Shawn Jennings – Johan45