The high performance/enthusiast class CPU from the AMD Ryzen range, based on the brand new Zen architecture, hits the ground running.
The high performance/enthusiast class CPU from the AMD Ryzen range, based on the brand new Zen architecture, hits the ground running.
It’s been a long wait, but the next generation of AMD’s CPUs have finally arrived: Ryzen is here. Performance reviews of the new CPUs are being rolled out over the course of the day as chips are shipped to consumers, and it seemed reasonable to give you a quick run-down of the brand new processors launch.
—
Ryzen launches today with the Ryzen 7 product range, designed for enthusiasts, prosumers and performance users, but that’s not the end of the Ryzen range. Later this year – provisionally pencilled in for Q2 – are the Ryzen 5 and Ryzen 3 models, with fewer physical CPUs cores tailored to different system requirements and price brackets. Those waiting for alternatives to Intel’s Core i3 and i5 CPUs will have to wait a little longer I’m afraid.
Ryzen 7 is comprised of three CPU models catering to subtly different markets, setting the up against Intel’s Core i7 Z270 and X99 CPUs. Over the course of the coming months further Ryzen CPUs will be launched in the mainstream and performance categories, but for now lets focus on today’s SKUs.
Heading up the lineup is the Ryzen 7 1800X, an 8-core/16-thread part that’s set to disrupt the high end desktop market. Retailing for only $499, it’s a far cry from Intel’s $1050 Core i7-6900K (itself an 8-core 16-thread CPU) but competes closely in synthetic benchmarks and AMD’s own internal testing. Oh, and it doesn’t require an exclusive HEDT motherboard platform; as with all Ryzen CPUs it’s perfectly happy sitting on the entry-level or premium AM4 motherboard.
Going up against the Core i7-6800K is the Ryzen 7 1700X, another 8-core/16-thread CPU. In point of fact, the only stock difference between the 1700X and 1800X is Base and Boost frequencies, dropping a modest 200MHz on both. Nonetheless it comes in at $399, below the $449 MSRP of the i7-6800K. Clearly, price is another aspect that AMD are competing heavily on.
Meanwhile the Ryzen 7 1700 is the ‘entry-level’ design, but it’s no mean CPU. Yes, another 8-core/16-thread design, the Ryzen 7 1700 drops core frequencies significantly whilst also undercutting Intel’s Core i7-7700K in price. The head-to-head between these two CPUs will be fascinating, contrasting the robust multi-threaded performance of the R7 1700 with the raw single-core brute force of the i7-7700K.
All the above would be for naught however if key metrics – IPC and perf/watt especially – didn’t take a large step forward.
The Key Metric – Instructions Per Clock
An oft-criticised aspect of the Bulldozer architecture was poor IPC performance, requiring the part to operate at very high frequencies in order to remain competitive. Staying ahead of the field by using a more advanced lithographic process would have helped the 8000-series CPUs stay in the game, but remaining at 32nm for too long and only transitioning to 28nm relatively late gave Bulldozer the reputation of a hot, slow architecture. Improving IPC became a core design goal early on for Zen.
AMD’s target was 40% improved IPC vs. Piledriver (2nd Gen Bulldozer), an extremely ambitious target. Thanks to the revolutionary design of Zen, and a transition to the Samsung/Globalfoundaries 14nm process, AMD are claiming an average IPC improvement of over 51%. That’s huge whichever way you want to shake it, and goes to show how much of a misstep Bulldozer turned out to be. Make no mistake – by AMD’s own admission Ryzen does still give up a little to Intel in the IPC stakes and hence single-threaded performance, but this is the first generation of this architecture going up against the 7th Gen Intel Core.
Higher IPC has also allowed reduced operational frequencies compared to Bulldozer and sub-100W TDP envelopes. It will also be interesting to see how well the chip can overclock, another core expectation of the enthusiast market.
The CPU Complex (CCX)
AMD’s Bulldozer architecture was a modular design made up of Compute Units – one large core coupled to a smaller one with a shared cache – and that’s one of the fundamental aspects which has changed with Zen. The new architecture defines a CPU Complex, or CCX, which is four fully functional cores married to 512KB L2 cache per core with a further shared L3 cache amounting to 8MB. In total, a Ryzen 7 CPU features 20MB cache – 4MB L2 (8 x 512KB) and 16MB L3 (2 x 8MB).
Clearly one of the advantages of this structure is its scalability – quad-core Ryzen 3’s will include just the one, 8-core Ryzen 7 CPUs feature two CCX modules, and the upcoming Naples server designs are set to be equipped with many more. However fine control is also possible as AMD include tools to disable sections of the CCX on a per-core basis, allowing not only 6-core configurations but also single-core modes for competitive overclocking.
XFR
Ryzen CPUs are specced out with Base and Boost core frequencies, representing typical operating frequencies under heavy multi-core load and lighter load utilising fewer threads. A further twist is a new feature AMD are calling Extended Frequency Range (XFR), which raises frequencies above the Boost state under single-core operation. Sensor data (esp. temperatures and voltages) is analysed to determine whether additional headroom is available above the Boost clock speeds; the CPU will then clock to a higher state beyond boost clock levels, providing just a little more performance.
The XFR boost levels are indicated by the CPU SKU. Those with feature an ‘X’ suffix are equipped with a 100MHz XFR, whereas those without have an XFR of half that. It’s worth noting that when user overclocking tools are enabled Boost and XFR frequencies are ignored, so they’re only applicable to stock clocked configurations. The overall impact of XFR is likely to be quite small with Ryzen 7, but there’s substantial scope for AMD to build on the technology or leverage it more aggressively in lower SKUs.
Simultaneous Multithreading (SMT)
Intel’s proprietary Hyperthreading technology has been core to their CPU design for fifteen years now. It has become a key selling point for Intel’s premium SKUs, but arguably also resulted in a stagnation of core counts such that entry-level chips are often still limited to two physical cores. In fairness, only relatively recently have applications begun to exploit these capabilities; notably game engines tended to rely on strong single-core performance rather than evenly distributed multi-core processing.
Although proprietary, Hyperthreading is an implementation of a more general technique known as Simultaneous Multithreading, where two threads with broadly shared resources can be processed at the same time. Bulldozer featured partial SMT – the integer cores were single-threaded whilst other sections were multi-threaded – but Zen implement 2-way SMT (or two threads per CPU core). Performance enhancements due to SMT are generally reliant on threads having similar workloads such that cached instructions and data can be shared, whilst the more optimal use of processor resources can significantly improve power efficiency, which is a key factor for enterprise markets in particular.
Although details on Ryzen 3 and 5 CPUs are still sparse it does appear that both lower-end ranges will feature SMT as standard. Thanks to the implementation of SMT developers can be reassured that hardware resources are available to take advantage of multi-threaded design even at an entry level, perhaps up to eight logical cores. It will be interesting to see how, if at all, Intel respond to this move by AMD.
Memory Support
AMD Ryzen on the AM4 platform supports dual-channel DDR4 up to a maximum frequency of 2667MHz according to JEDEC standards. AMD representatives indicated that the maximum memory capacity possible is 128GB, but the cost of 32GB DIMMs effectively limits support to 64GB or 32GB in two-DIMM configuration. Ryzen does not support XMP, Intel’s proprietary extension to the JEDEC SPD information resident on RAM DIMMs. Consequently additional memory timings will need to be input manually, but this will also allow support for frequencies beyond the stock 2667MHz defined by JEDEC.
Where Ryzen Shines – Multithreaded Workloads and Megatasking
Without a doubt Ryzen’s speciality is multithreaded performance. AMD’s own benchmarks – benchmarks which we should stress are the same tools that Intel utilised in the past to support their own performance claims – indicates that they’re competitive with the Intel chips they’re lined up against. The major questionmark hovering over Ryzen – much like Bulldozer – is single-threaded performance especially in gaming.
It’s in this respect that AMD want to emphasise new opportunities which are brought about by 8-core, 16-thread parts, namely ‘Megatasking’. In principle, Megatask workloads are those which feature multiple intensive tasks simultaneously; the classic example given is playing a computergame and simultaneously encoding the video for streaming using CPU resources. Here both tasks should be completed in a timely fashion to ensure both smooth gameplay and a smooth stream, far more difficult than you might think.
Megatasking benchmarks are inherently ad hoc. To compare the Ryzen 7 1700 to Intel’s Core i7-7700K AMD measured the number of dropped frames of an encoded stream of DOTA 2 gameplay, finding that the 7700K dropped ~15% whilst its Ryzen counterpart dropped a total of zero. This would equate to a far better viewing experience for those tuning in over Twitch, for example.
The value of megatasking is hard to pinpoint, and further undermined due to the range of resources available on other components in a modern PC (such as hardware encoders on discrete GPUs). However as the demands of high quality streaming increase, and Virtual Reality enters the fray, more cores rather than higher core frequencies, could become ever more important.
Motherboards – Simple and Cost-Effective
With Ryzen AMD are beginning the process of adopting a new motherboard platform across their range. This platform will be defined by five chipsets – X370, B350, A320, X300 and A300 – which cater to the whole of the desktop and small form factor market from enthusiast down to the small form factor entry-level, with the X370 sitting at the top of the tree.
The differences in functionality are outlined above, and as you can see there’s not a huge amount of daylight between each. Multi-GPU configurations will require X370, if you only need robust overclocking support a B350 model will do fine. NVMe storage support via M.2 is possible throughout the range, as is USB 3.1 Gen 2 I/O and more common storage options. Furthermore even the X370 motherboards are more affordable than their X99 or X270 equivalents, and once Ryzen 5 and 3 CPUs enter the market excellent value systems which perform above mainstream expectations could well be the order of the day.
Fundamentally, the new Ryzen platform as a whole is extremely flexible, where each CPU is fully unlocked and only the motherboard determines functional platform limitations. As each CPU is compatible with each motherboard upgrades should be simple, even to the extent of replacing a processor with an APU.
Summary – Just the Beginning
So, that’s just a taste of AMD Ryzen. Benchmarks are now available across the web for you to assess, and the first round of CPUs and motherboards are available now. There’s still a lot of ground for AMD to cover that will show the world if they can once again compete in the mainstream space against Intel, but for now they appear to have made a very solid start that dispels the spectre of Bulldozer once and for all.
For more information on AMD Ryzen visit http://www.amd.com/en/ryzen. Also check out our review of the Ryzen 7 1800X and GIGABYTE GA-AX370 AORUS GAMING 5 motherboard.
