As the launch date of AMD's next Zen 2 architecture is fast approaching, the company has raised the curtain and provided insight into the capabilities and improvements of its new model. These new chips include a number of enhancements and benefits to generate both higher cycle instructions (CPIs) and better overall power.
Let's start with some basics. The Ryzen 3000 family is powered by AMD's Zen 2 architecture except for its APU. APUs actually have a generation behind them – the 2000 Series APUs were actually built on the first generation of Ryzen, and the APUs of the 3000 series are based on the second generation Ryzen. The architectural improvements and other features we are going to talk about today do not apply to the Ryzen 3 3200G or the Ryzen 3 3400G.
During his Zen 2 presentation, AMD Corporate Fellow Mike Clark said his 7nm transition was in fact more successful than originally planned.
Some of you may remember rumors that AMD would use Ryzen 3000 processors with clocks much higher than the previous ones. According to AMD engineers, the company did not necessarily expect Zen 2 to reach higher frequencies. at all. This is the intrinsic problem of shrinkage of modern UC nodes. Smaller process requirements mean lower voltages, and lower voltages can negatively impact the absolute operating frequency. In this case, however, the 7 nm TSMC node and AMD's own engineering were able to create components that could reach slightly higher frequencies than the 12/14 nm chips.
It should be kept in mind that AMD did not expect improvements in the clock frequency of 7 nm, which should be taken into account when doing the same. Evaluation of the accuracy of rumors about massive clock jumps in the future.
An important change coming with Zen 2 has nothing to do with the current processor. AMD informed us during the event that new scheduler changes were being incorporated into Windows 10 Scheduler from Windows 10 1903 (May 2019 update). There are two new features: topology recognition and faster clocking of the clock. A faster acceleration of the clock reduces the time required by the processor to change state, thus improving performance and theoretically, inactive power by allowing the processor to move more quickly to different states. lower clock. Knowledge of the topology should help to keep local CCX data relevant and complete a CCX before loading another one.
These gains – a performance of + 15% in 1080p in Rocket League and a 6% improvement in the launch of PCMark 10 applications – are only result from the update of the Windows 10 scheduler and are separate from any additional gain resulting from improvements to the Zen 2 architecture. Taking advantage of these enhancements requires both an updated chipset driver and the upgrade. day Windows 10 1903.
This slide represents the microarchitectural synthesis of AMD. The chip integrates with the new TAGE branch predictor in addition to the BP perceptron it used in the past. The micro-op cache has been increased to 4K instructions, with double the total of L3 on board. (AMD is now referring to its L2 and L3 combinations as "AMD GameCache".) A new Address Generating Unit (AGU) is now associated with the entire side of the core, with full support of the comma floating 256-bit via AVX2.
The slideshow below presents our in-depth analysis of the specific architectural improvements of the third-generation Ryzen processor. Each slide can be clicked to open it in a new window.
According to AMD, these improvements give them an edge over Intel, both in terms of performance per watt and absolute wall power.
Cinebench is not the ideal solution for measuring energy consumption, but it is not a bad test either. The 3700X – which, in all fairness, is probably closer to the ideal position for architecture – is assumed to be 56% more efficient than the Core i7-9700K, while it only consumes 86 % of the power in absolute value.
Energy efficiency gains over the 2700X are even greater. AMD claims that the 3700X is 1.75 times more efficient in performance / watts than the 2700X, while consuming 70% power.
Although we obviously can not decide to launch before having material to test, AMD offers an aggressive and exciting product family. The TDPs have dropped dramatically. The CPI would have risen 1.15 times. Clock speeds have been increased. Planner improvements and doubled floating point capacity should provide their own robust improvements beyond this figure of 1.15. The width of the Infinity Fabric bus has been doubled to allow the use of PCIe Gen 4 bandwidth and a new memory divider at the DDR4-3733 level can reduce IF clocks without compromising the time of day. DRAM scale.
If you're a fan of AMD APUs, 7 nm also has exciting long-term implications for them. Although we do not know when we will see these parts, the company has clearly aggressively targeted a decline in power in all areas. This will clearly pay off when refreshing the APU family at 7 nm. According to one of our 7nm launch theories, AMD would focus on energy efficiency at least on some parts, and we see it perfectly, with a higher-performance 8-core processor in a 65W TDP and a 16-core processor in at 105W TDP.
In January 2018, we discovered a set of vulnerabilities related to how modern processors perform a function known as speculative execution. Spectrum and Meltdown were considered serious, partly because Specter represented a new class of attacks, not a single isolated attack vector. For most of 2018, Intel's "history" was focused on its reaction to these attacks.
Nearly a year and a half later, researchers are still looking for categories of similar problems. Several new vulnerabilities have broken the cover and carry various names such as ZombieLoad, RIDL and Fallout (as named by the researchers). Collectively, Intel summarizes them as MDS – Microarchitect Data Sampling. RIDL, for Rogue In-Flight Data Load, was discovered by researchers at the Vrije Universiteit Amsterdam and the Helmholtz Center for Information Security. Fallout was discovered by a group from the Graz University of Technology, KU Leuven, the University of Michigan, and the Worcester Polytechnic Institute. ZombieLoad was discovered by Graz, Worcester and KU Leven.
As a reminder: all these flaws, including Specter and Meltdown, are related to the way processors in general or Intel processors, in particular, work speculative execution. In the case of RIDL, ZombieLoad, Fallout and MDS more generally, underlined faults seem to be specific to Intel processors. These problems are due to differences between the architecture of a CPU (its functioning on paper is documented) and its microarchitecture (the way the CPU performs operations "upside down".) The speculative execution is exactly what It looks like: The CPU speculates on the operations to be executed next, and then executes them so that the results are ready if they are needed, rather than performing these operations afterwards. know they are necessary.
As far as the architecture is concerned, all the operations are done in sequence and the only data is kept by the CPU. But it is possible to observe the microarchitecture to look for subtle clues about the storage of data on a chip, depending on the time needed to access the information. Measuring these differences can allow attackers to infer data values stored in the cache or in built-in buffers. Previous flaws in the Specter class typically focused on cached data leakage, but new MDS flaws leaked buffered data – tiny data stores that the chip uses to move data internally.
There has been a bit of controversy about the severity of these new attacks, and I will honestly say that I am a little dissatisfied with how some of this news has been published. Some of you may remember last year, when a supposed security research firm, CTS-Labs, appeared to be collaborating with a short-sellers firm blatant attempt to attack the price of AMD shares by making public a set of security flaws allegedly critical that disclosure literally implicit could put lives in danger. Absolutely nothing came of these defects, which, as literally predicted by the short seller, Viceroy Research, could reduce the value of AMD's shares to zero. As we discussed at the time, contaminating security disclosures with hyperbolic marketing claims to portray problems as being far worse than they were actually a bad idea, no matter who is targeted or for what reason.
The situation with Intel is not closeand that is bad, but it shows some of the same disturbing trends that I talked about last year. The researchers chose to make their efforts known on a website named "CPU.fail"With scary graphics and an FAQ that seems more designed to scare than to inform, when asked if problems are highlighted in nature, for example, they simply state," We do not know. "But the question of the gravity of these problems is it? in practice is authentic.
Until now, no attack using Spectrum and Meltdown actually has been spotted in nature, apart from the concept validation work submitted by the researchers. Similarly, taking advantage of MDS is more difficult than what this website entails. Attackers can not directly control what is in the targeted buffers, for example, which means that the exploit may leak old, obsolete data without interest. Microcode updates for systems operating between Sandy Lake and Kaby Lake have already been delivered to customers. The first generation Coffee Lake and Whiskey Lake processors are already immune to this attack. The impact of the patch on performance is estimated at about 3%.
D & # 39; Intel official statement said:
Microarchitectural Data Sampling (MDS) is already supported at hardware level in many of our recent 8th and 9th generation Intel Core processors, as well as in the 2nd generation Intel Xeon scalable processor family. For the other affected products, mitigation is available via microcode updates, associated with the corresponding operating system updates and hypervisor software available from today. # 39; hui. We have provided more information on our website and continue to encourage everyone to keep their systems up to date because it is one of the best ways to stay protected. We would like to thank the workers who have worked with us and our industry partners for their contribution to the coordinated disclosure of these issues.
So, the tone of the cover on this question has varied a lot. Wired takes a your alarmist, arguing that these flaws "allow attackers to listen to virtually every bit of raw data that the victim's processor touches" and claiming that the researchers are right in saying that these flaws are quite serious. Intel argues that they are of medium to low severity, given the difficulty of removing them, the lack of practical attacks in the wild and the fact that microcode updates and fixed hardware processors are already on the market. Like PCMag marks:
(T) The micro-architectural data sampling vulnerabilities revealed today seem to be more academic at this point. For the moment, no real attack involving flaws has been encountered and made public. One of the main reasons is that this is possible because hackers can simply use traditional malware to steal data on your PC rather than having to falsify the Intel processor.
The fact that they have not been tested with these PCs is part of the difficulty in solving these problems, but these are the problems we face in terms of PC security since we have PCs. A serious flaw could be which experts to listen to. Last year, for example, Theo de Raadt decided to change the default behavior of OpenBSD in disable hyper-threadingconsidering it as a fundamental security risk. Other operating system vendors have not gone so far as to disable this feature preemptively. Is hyper-threading a potential safety risk? Yes Is there a potential security risk large enough for existing users to disable this feature? The experts are literally at odds. The honest answer is: "It depends", because everyone wants to be wanted, but because of the appropriate security practices in a given situation, depending on the threat and the cost of activating the patches in question.
How badly do you take these threats? Serious enough to patch up, certainly. But past that, the practical implications of the real world are still unknown. To date, we have not seen any Spectrum or Meltdown attacks in the wild that pose a threat to Intel processors (or anyone's processors) of any generation. This does not mean that we can not arrive, and that does not relieve Intel of the responsibility of securing its products. But that does not mean that invisible hackers are sneaking between your pockets now via hardware attacks that you are not aware of. All security vulnerabilities do not become a practically exploited avenue of attack. Until now, these attacks do not have.