AMD Ryzen 5 2400G And Ryzen 3 2200G Review: Raven Ridge Desktop Debuts
AMD Raven Ridge Ryzen 3 2200G And Ryzen 5 2400G APUs
AMD continues its assault on the desktop PC market with a pair of new, affordable APUs today, the Raven Ridge-based Ryzen 3 2200G and Ryzen 5 2400G. At this point, if you follow PC technology with any sort of regularity, you’re probably familiar with Raven Ridge. AMD has been talking about its plans to mate the Zen CPU and Vega GPU microarchitectures for quite some time and has revealed a number of details regarding these new hybrid chips over the last few months. In fact, some feature additions and initial specifications were just revealed as the CES 2018 show kicked off a couple of weeks back.
Today though, we can give you the full scoop. We have had the Ryzen 3 2200G and Ryzen 5 2400G APUs in house for a few days now and have put them through their paces with an array of benchmarks. We’ve got CPU and GPU tests on tap, along with some power and temperature data, and overclocking thrown in for good measure. Check out the Ryzen 3 2200G and Ryzen 5 2400G’s main features and specifications below and then we’ll dig in a little deeper and see how these puppies perform with a variety of workloads...
We can’t discuss Raven Ridge without mentioning Zen, Vega, and Infinity Fabric. We have covered AMD’s Zen CPU microarchitecture and the Vega GPU at length in the past, so we won't re-hash the meaty technical details again here. We will, however point you to a few articles should you need a refresher. In this piece, we cover all of the features and capabilities of the Zen CPU microarchitecture, including SenseMI, Precision Boost, XFR, and more. In this piece, we cover all of Vega’s technical attributes, we have benchmarks here, and all three pieces are sprinkled with details of AMD’s Infinity Fabric, which ties all of the company’s IP together over high speed communication links.
There are six main clients linked together on Raven Ridge with AMD’s Infinity Fabric. The APUs consist of a single, monolithic piece of silicon (it is not an MCM) with a die size of 209.78mm2, comprised of roughly 4.9B transistors. A Zen processor core complex (CCX) with quad-CPU cores and a Vega graphics engine (with 11 or 8 CUs), are linked to multi-media and display engines, along with the DDR4 memory controller and IO.
The model numbers of these new APUs may give the impression that AMD just bolted Vega onto the Ryzen 5 1400 and Ryzen 3 1200, tweaked the clocks, and called it a day, but that is not the case. These first Raven Ridge APUs have only a single CCX, whereas first-gen quad-core Ryzen processors had two, but with fewer active cores in each. A Ryzen 5 1400, for example, had a 2+2 CCX configuration, with two active CPU cores in each CCX. The Ryzen 5 2400G has only a single, quad-core CCX. This configuration has the benefit of lowering latency should workloads have to bounce between threads, but this comes at the expense of L3 cache. Raven Ridge has 4MB of available L3 cache, while a first-gen Ryzen 5 1400 (and other Ryzen processors with a 2+2 CCX configuration) has 8MB. The number of PCIe lanes available with Raven Ridge has been reduced as well; there are only 8 GPU PCIe lanes available versus 16 in standard Ryzen desktop processors (sans GPU).
AMD aims to offset the potential performance penalty of a reduced L3 cache size in a couple of ways. Raven Ridge APUs have higher clocks than their first-gen counterparts. The Ryzen 5 2400G’s clocks range from 3.6GHz – 3.9GHz (base / boost) and the Ryzen 3 2200G clocks-in at 3.5GHz - 3.7GHz (both chips have a 65W TDP). These new APUs also employ Precision Boost 2. The new Precision Boost 2 algorithm is designed to improve performance, responsiveness, and power characteristics. It employs the same 25MHz granularity of Precision Boost 1, but intelligently boosts to the highest possible frequency until a thermal or power limit is reached and continually adjusts each of the CPU core frequencies. The original Precision Boost essentially had only two states – “all core” and “two core” boost, and that was it.
The Ryzen 5 2400G is a quad-core / eight-thread machine with an on-die, 11 CU (compute unit) Vega graphics core with 704 stream processors, priced at $169. And the Ryzen 3 2200G is a quad-core / quad-thread chip with an 8 CU Vega-based graphics engine with 512 stream processors for only $99. Externally, the APUs look just like any other Ryzen processor designed for the AM4 platform, but there are actually some differences with the packaging here as well. AMD is switching to a non-metallic TIM for the 2400G and 2200G, and the new CPU package allowed the company to officially support JEDEC DDR4-2933 memory speeds as well.
Today though, we can give you the full scoop. We have had the Ryzen 3 2200G and Ryzen 5 2400G APUs in house for a few days now and have put them through their paces with an array of benchmarks. We’ve got CPU and GPU tests on tap, along with some power and temperature data, and overclocking thrown in for good measure. Check out the Ryzen 3 2200G and Ryzen 5 2400G’s main features and specifications below and then we’ll dig in a little deeper and see how these puppies perform with a variety of workloads...
 |
|
We can’t discuss Raven Ridge without mentioning Zen, Vega, and Infinity Fabric. We have covered AMD’s Zen CPU microarchitecture and the Vega GPU at length in the past, so we won't re-hash the meaty technical details again here. We will, however point you to a few articles should you need a refresher. In this piece, we cover all of the features and capabilities of the Zen CPU microarchitecture, including SenseMI, Precision Boost, XFR, and more. In this piece, we cover all of Vega’s technical attributes, we have benchmarks here, and all three pieces are sprinkled with details of AMD’s Infinity Fabric, which ties all of the company’s IP together over high speed communication links.
The model numbers of these new APUs may give the impression that AMD just bolted Vega onto the Ryzen 5 1400 and Ryzen 3 1200, tweaked the clocks, and called it a day, but that is not the case. These first Raven Ridge APUs have only a single CCX, whereas first-gen quad-core Ryzen processors had two, but with fewer active cores in each. A Ryzen 5 1400, for example, had a 2+2 CCX configuration, with two active CPU cores in each CCX. The Ryzen 5 2400G has only a single, quad-core CCX. This configuration has the benefit of lowering latency should workloads have to bounce between threads, but this comes at the expense of L3 cache. Raven Ridge has 4MB of available L3 cache, while a first-gen Ryzen 5 1400 (and other Ryzen processors with a 2+2 CCX configuration) has 8MB. The number of PCIe lanes available with Raven Ridge has been reduced as well; there are only 8 GPU PCIe lanes available versus 16 in standard Ryzen desktop processors (sans GPU).
AMD aims to offset the potential performance penalty of a reduced L3 cache size in a couple of ways. Raven Ridge APUs have higher clocks than their first-gen counterparts. The Ryzen 5 2400G’s clocks range from 3.6GHz – 3.9GHz (base / boost) and the Ryzen 3 2200G clocks-in at 3.5GHz - 3.7GHz (both chips have a 65W TDP). These new APUs also employ Precision Boost 2. The new Precision Boost 2 algorithm is designed to improve performance, responsiveness, and power characteristics. It employs the same 25MHz granularity of Precision Boost 1, but intelligently boosts to the highest possible frequency until a thermal or power limit is reached and continually adjusts each of the CPU core frequencies. The original Precision Boost essentially had only two states – “all core” and “two core” boost, and that was it.
The Ryzen 5 2400G is a quad-core / eight-thread machine with an on-die, 11 CU (compute unit) Vega graphics core with 704 stream processors, priced at $169. And the Ryzen 3 2200G is a quad-core / quad-thread chip with an 8 CU Vega-based graphics engine with 512 stream processors for only $99. Externally, the APUs look just like any other Ryzen processor designed for the AM4 platform, but there are actually some differences with the packaging here as well. AMD is switching to a non-metallic TIM for the 2400G and 2200G, and the new CPU package allowed the company to officially support JEDEC DDR4-2933 memory speeds as well.
