The launch of the new AMDU APUs reaches the top of the 2017 design that tends to black balance. After returning to high performance on x86 with Ryzen's CPU product line, and 'we can not produce enough' graphics Vega, AMD has included several product lines that combine both. Today is the launch of the APU desktop socket edition, with four core Zen and up to 11 Vega computing units. AMD is historically aggressive in low-end desktop space, which effectively kills sub-$ 100 discrete graphics market. The new APU now regulates the bar even higher. In this review we focus on Ryzen 5 2400G, but also test Ryzen 3 2200G.
Ryzen 5 2400G and Ryzen 3 2200G: Series Ryzen 2000
Two APUs launched today AMD are Ryzen 5 2400G, Zen quad-core processor for $ 169 14nm with multithreading graphics and simultaneous 'Vega 11', and Ryzen 3 2200G, Zen 14nm quad-core processor without simultaneous and simultaneous multithreading processors. with the 'Vega 8' graphics. Both parts can be distinguished from Ryzen's graphical processor that is not integrated with 'G', which is similar to how Intel markets its own Vega-enabled processor.
Although Ryzen 5 2400G is classified as 'Ryzen 5', chip specifications are quite a few of the peak specs expected to be silicon to offer. AMD has stated that there is currently no planned Ryzen 7 special plan at this time. Ryzen 5 2400G has four complete cores with simultaneous multi-threading, and a complete set of 11 unit computations on integrated graphics. This is a much larger unit of calculation than the Mobile 73000U Ryzen 7 processor, which only has 10 counting units but is limited to TDP 15W. 11 unit counts for 2400G is translated as streaming processor 704, compared with 640 SPs at Ryzen 7 2700U or 512 SPs on previous generation desktop APU: 25% effective automatic upgrade from generation to generation desktop AP without considering Vega architecture or frequency fix.
The integrated graphics frequency will be the default 1250 MHz and the total TDP chip is 65W. The maximum supported memory frequency will vary depending on how much memory is used and what type, but the AMD DDR4-2933 list as support for one side module per channel. In addition to full hardware, the CPU frequency of 2400G is very high, similar to the standard Ryzen 7 desktop processor: 3.6 GHz basic frequency and 3.9 GHz turbo will leave less room for overclocking. (Yes, that means the chip is overclockable.)
Ryzen 5 2400G somewhat replaces Ryzen 5 1400 at $ 169. Both of these chips will continue to be sold, but at this price point AMD will promote 2400G above 1400. 2400G has a higher frequency (base frequency 3.6G vs. 3.2G, turbo frequency 3.9 volts vs 3, G), higher Memory support (DDR4-2933 vs DDR4-2666), no cross-CCX latency between core sets but L3 less cache per core (1 MB vs. 2 MB). In almost all scenarios, even if the user does not use Ryzen 5 2400G integrated graphics, Ryzen 5 2400G seems to be a better choice on paper.
The cheaper $ 99 processor is Ryzen 3 2200G. Specifications follow other existing Ryzen 3 processors on the market: four cores, and no multi threading simultaneously. The rated frequency, 3.5 GHz for base and 3.7 GHz for turbo, slightly below Ryzen 5 2400G but still high enough - although these chips are rated at 65W, similar to Ryzen 5 2400G, users may expect this processor. to a longer turbo inside the window of its power during thermal limits (we see this in some benchmarks in the review). Recommended retail price of $ 99 means that this is the cheapest Ryzen desktop processor available on the market, and this is a fantastic channel for consumers: four high performance x86 cores under the $ 100 mark. Integrated graphics provide 512 streaming processors, identical to processors $ 169 from previous generations, but this time is enhanced by the Vega architecture.
In a presentation on Tech Day, AMD usually provides a lot of performance data from their own lab. Of course, we prefer to present our own data obtained in our lab, but combing the AMD numbers gives a point of pointing to AMD's confidence in its low-end units: using 3DMark 11 Performance Ryzen 3 2200G (according to AMD) prints 3366 points, while on the same benchmark as Intel's integrated graphics, Core i7-5775C with embedded DRAM, prints only 3094. If we take this data point as everything and the end of it all, it will find that AMD has broken its integrated graphics strategy Intel. We have some other interesting numbers in today's reviews.
One of the other important elements for the launch of APU Ryzen is that both processors, including Ryzen 3 2200G at $ 99, will be bundled with reinforced 65W AMD Wraith Stealth (non-RGB) cooling. This is not a high-end AMD cooler, but as far as stock cooler comes in, it easily introduces $ 30 savings to any PC, reducing the need to buy a good standard cooling.
Combining Performance with Performance: Featured Strategy (on Paper)
Over the past 10 years, joining the CPU and GPU together, either as two silicon bits in one package or both at the same bit of silicon, match the holes that drive the low-end market. This really reduces the need for discrete graphics cards if all users are required is a basic desktop experience. It also has a knock-on effect for mobile devices, reducing total power needs even in lightweight workload scenarios. From then on, however, integrated graphics have been constantly asked to do more. In addition to 2D layers, we now ask them to deal with interactive web pages, new graphics APIs, and new video decoding formats. Lined up to show higher resolution means a new complex video encoding information has been developed to minimize file size but maintain quality, which can stretch integrated basic graphics solutions, so hardware decoding is dedicated to being added to future versions of the hardware.
Sisyphean's task, the Holy Grail for graphics, is always a game. Higher equality, higher resolution, and more ambiguous environments such as virtual reality, far beyond the integrated graphics scope. For the most part, complex tasks are still there now - do not let me fool you in this regard. But AMD did manage to change the status quo when introducing the VLIW design later, followed by its GCN graphics architecture, several generations ago. The argument at that time was that most budget users were limited, and by saving money for decent integrated graphics solutions, low-end gamers could have a much better experience. It seems strange at the time, given the success of AMD in the low-end discrete graphics market - they advertise the sale of one product to another with more complex designs and lower margins. This was clearly seen during our review of the time.
For several years of core Bulldozer processing and integrated graphic design, AMD competed in two main places: performance per dollar, and peak performance. In this market, its competition is Intel, with graphic design 'Gen'. Both companies made great strides in graphics, but bifurkation soon began developing Intel Gen graphics that were easy enough for office work on mobile devices, using higher performance processors, and more efficient CPU power with good margins. AMD compete more for desktop market share, where power limits are less concerned, and provide similar or better peak performance graphics at much lower costs. For the low-end graphics market, this is perfect for them alright, though AMDs are still lagging in general CPU performance which indeed makes certain segment users inactive.
What AMD noticed is one of the limitations for this integrated design is memory bandwidth. For several years, they continue to release products that have higher base memory support than Intel: when Intel still has DDR3-1600 registered as a supported frequency, AMD switched to DDR3-2133, which improves graphics performance with a fair margin. You can look in our memory scaling articles with Intel Haswell products that DDR3-1600 is effectively a black hole to unlock integrated graphics performance, especially when reaching the minimum frame rate.
At this stage in history, the memory bandwidth to the CPU is about 20 GB / s, compared to discrete graphics that drive 250GB / s. Memory bandwidth problems were not addressed by Intel, and so did Broadwell, they introduced the Broadwell 'Crystalwell' processor lineup: it featured the largest implementation of Intel's latest graphic design, which was paired with DRAM silicon embedded in its packaging. This eDRAM, up to 128MB, is a victim cache, which allows processors for data (such as textures) that have been retrieved from memory and already used at 50 GB / s re-use (bi-directional) levels. The ability to store relevant data with graphics rendering is closer to the processor, at faster bandwidth than main memory, paired with Intel's best integrated graphics design, touting the new halo product in this category. This eDRAM processor circuit also gives speed to the additional bandwidth bandwidth assignments that are reused, as stated during our review. The big drawback of this is the price: adding a bit of silicon to packets, some accounts, is quite cheap: but Intel sells it at high premium prices, aimed at a particular customer with a fruit logo. Some parts are also available to end-users, very short before being excluded from sales, and are quoted by other media that OEMs do not like the price.
During 2017, AMD launched its Ryzen desktop processor, using the new xen zen microarchitecture. This is a return to high performance, with AMD citing 52% gains from previous generations at the same frequency, by redesigning fundamentally how the core should be created. We reviewed the ranks of Ryzen 7 processors, as well as Ryzen 5, Ryzen 3, Ryzen Threadripper, and the company's EPYC processors, all built with the same core layout, concluding that AMD now has a high-performance design in competition cries in a market that respects single performance -threaded. AMD is also competing on per dollar performance, reducing competition and making the Ryzen family's headline into a number of popular Buyers Guide, including our own. AMD also launched a new graphic design, called Vega. AMD is positioning Vega products to compete against the NVIDIA dollar for the dollar, and although power consumption for high-end models (up to 64 computing units) is in doubt, AMD can not currently make enough Vega chips to meet demand, as some workloads are done. best in Vega In a recent financial call, CEO Dr. Lisa Su stated that they are constantly increasing (increasing) the production of discrete Vega graphics cards because of that demand. Despite the power consumption for graphical workloads on high-end discrete graphics, it is always accepted that the peak efficiency point for the Vega design is the smaller and lower frequency. It seems that Intel partially agrees with this statement, as it recently introduced an Intel Core graphics processor with Radeon RX Vega, combining its high performance core with medium-sized Vega chips, powered by high bandwidth memory. The reason for choosing an AMD graphics chip instead of scrolling it yourself, according to Intel, is the right section for that product segment.
So, a similar reason for today's launch: combine high performing cores with high-performance graphics. For the newly launched Ryush Desktop APUs today, AMD has incorporated four high performance Zen x86 cores and smaller versions of the Vega graphics into the same silicon part. Like all silicon manufacturers, APU design should reach the right point of performance, power, dead area, and cost, and with this product, AMD focuses precisely on entry-level gaming performance metrics, for users who spend $ 400- $ 600 on entire PC, including motherboard, memory, case, storage, power supply, and operating system. The idea is that high-performance processor cores, combined with high-performance graphics, can create unequal products for the market.