**AMD's Power Management and Graphics Technology**

In the latest developments from AMD, it has become clear that the company is prioritizing power management and graphics technology in their upcoming products. The Ryzen 5000 series, which includes the Ryzen 7 3700X and Ryzen 5 5600X, features a new power management system that dynamically allocates power to the CPU and GPU based on the load of each component. This means that when the CPU is under heavy use, more power is directed towards it, while the GPU is able to throttle back its power consumption. On the other hand, during graphics-intensive tests, the power allocation shifts towards the GPU, allowing it to optimize its performance.

The AMD engineers have explained that this approach allows them to balance the power consumption of both the CPU and GPU within the thermal and power limitations of their Raven Ridge chips. This is particularly important for mobile devices, where battery life is a critical consideration. By dynamically allocating power, AMD can reduce power consumption by up to 95% in certain situations, which has significant implications for laptops and other portable devices.

The engineers have also discussed the use of low-dropout regulators (LDOS) as a means of controlling power delivery to the GPU. These regulators are designed to provide stable power to the GPU even when it is under heavy load, while still allowing for power reduction during periods of inactivity. Additionally, AMD has implemented advanced power-gating techniques that enable the disablement of certain components, such as the GPU's compute units or streaming processors, to reduce overall power consumption.

The implementation of these power management techniques allows AMD to achieve a 99% residency on a Windows idle screen static, which means that even when the system is not actively being used, it is still consuming very little power. This is made possible by the advanced thermal and power management capabilities of their Raven Ridge chips, which are designed to optimize performance while minimizing energy consumption.

**AMD's Roadmap for Future Products**

The AMD roadmap for future products has been outlined in recent months, and several new developments have been announced. The company plans to release a range of new products, including the Ryzen 3 3000U and Ryzen 5 2200U APUs, which are expected to offer improved performance and power efficiency.

Additionally, AMD is working on a new graphics processing unit (GPU) that will compete directly with Nvidia's offerings in the deep learning and machine learning market. This GPU will be manufactured on a 7nm process and is expected to provide significant improvements in terms of performance and power efficiency.

In terms of desktop processors, AMD has announced plans to release several new products in the coming months, including the Ryzen 5 4000 series and Ryzen 9 5000X series. These processors will offer improved performance and power efficiency compared to their predecessors and are expected to be highly competitive with Intel's offerings in the market.

**AMD's Future Plans for Motherboards**

In addition to their CPU and GPU developments, AMD has also announced plans for several new motherboard designs that will take advantage of their Ryzen 5000 series processors. The company is working closely with motherboard manufacturers such as ASRock, MSI, and Gigabyte to ensure that their motherboards are optimized for performance and power efficiency.

Some of the features that have been announced include improved overclocking capabilities, advanced cooling solutions, and reduced pricing on certain models. AMD has also stated that they will be focusing on enabling more advanced overclocking capabilities in future products, which could provide users with even greater control over their system's performance.

Overall, it is clear that AMD is committed to delivering high-performance and power-efficient products across a range of markets, from desktops to laptops to mobile devices. With their continued focus on innovation and optimization, it will be interesting to see how these products perform in real-world testing.

"WEBVTTKind: captionsLanguage: enour first major AMD video for the week showcased an X 470 motherboard from gigabyte with the m4 socket of course just with an updated chipset and some changes post rise and launch so as the motherboard vendors have been ironing out their side of Rison and the two has been working on their next iteration of verizon the Rison 2 or as it's been called sometimes Rison plus family of CPUs which will exist on the zen architecture not Zen - but a Zen architecture and also there's news on Raven Ridge the APU integration of the Zen or Rison core complexes and the Vega GPU so I'll be talking about that briefly today before that this coverage is brought to you by Thermaltake and the flow liquid cooler the flow is a 360 millimeter radiator with three ring RGB LED fans you can program the fans our custom lighting through software and then of course benefit from the larger radiator size and cooling performance learn more at the link in the description below this one's gonna be a quick one for us we didn't go to the AMD tech day but we collected the information separately most of our early efforts were focused on the x4 70 motherboard which is one of the first one that's been shown officially you can find that out on our channel by gigabyte so the main information here especially as it pertains to Raven Ridge and the AP is is that the architecture as expected is focused on power savings this is something Andy's been big on it with Rison and the Zen architecture in general and they've had some trouble with it with Vega but combining the two is a bit different than just Vega stand alone on the big d GPU PCB so combining them Vega and risin as expected will communicate via and these high speed interconnect which they call infinity fabric and that also bridges to them the system IO via the chipset and the integrated memory controller the IMC so all these things are bridged together via infinity fabric and then the other kind of major changes there are pertaining to precision boost - which is what Andy is calling its updated version of its boost technology for Zen so the boost technology primarily has hinged on the things you would expect how much current is going to the CPU on the GPU side you're looking at current voltage power consumption and thermals and a lot of the same stuff is true for the CPUs so we look at these elements for precision boost - with the addition of implementing a new volt frequency curve where unlike with Zen one where you had very hard to set frequencies for core utilization you use one core versus two versus all of them there's a very defined almost binary toggle between the frequencies unlike that now it's a more gradual gradient for the frequencies versus the core utilization and the voltage so that's a big change they have charts for that where you can see it's it's kind of a as you increase core count it kind of asymptotes out at the bottom so that's different and it's somewhat mirrors what Intel has been doing not to say that Andy and Intel are copying one another but that this is clearly one of the more optimal ways to configure your core frequencies versus your power consumption another item here is that AMD is looking at which thread is under the most demand so the thread under the most active load is the one that will get the most power sent to it or the most current sent to it so they are modulating based on that not just based on a hard lookup table or something like that further the power between the CPU and the GPU package are allocated based upon the current requirements of the applications running so AMD is using firestrike as an example here on one of their presentation slides which makes a great example because fire strike for 3dmark fire strike it has a few tests it's got graphics tests and physics tests physics uses almost zero graphics compute and graphics uses very little CPU so between the two you can see the power allocation go up for the Vega components the graphics component during the graphics test you see the allocation go up for the CPU component during the physics test which means that they're able to min/max their power and current budget because you can't just push all of it through to both components they don't have the budget for it so you pick and choose to fit the smaller thermal and power consumption limitations of these Raven Ridge chips and that's what they're doing here and they're able to dynamically regulate where the power and the current is going based on the load of the CPU at that time or the GPU as it were and the unfortunately some of their slides aren't all that helpful they are employing the bigger bar better mentality for them where you've got a y-axis with no label well that have a label but no numbers an x-axis with no numbers so it's just sort of nebulous infinitely going in each direction with no actual understanding of what it means which is very unfortunate because it's an interesting slide otherwise but this is something we'll look at more in testing later on just like we talked about when Rison launched AMD is still using LDOS or low dropout regulators for their power management and the LDOS here double as power gates so they can turn things on and off based on the current needs of the system this goes to the extreme of being able to disable 95% of the GPU if it's not needed which of course is really important for any kind of mobile chip because you care a lot about battery life and if you have a GPU sitting there burning power waiting to be used you're just burning battery life so they can paragate down to 95% of it off and they can further power gate with GPU uncor so all of the uncor components things like cash would be mitigated in their consumption as well and save more power power gating for SOC functions is split into two containers so they begin to slide for this one where the containers are the theirs they label them a and B one of them contains the CPU interface the GPU interface and the i/o hub that would be all in one and the other one contains the multimedia hub and the memory controller in a second container and then these can be gated based on which region or container they are in they Assembly load of the system to reduce power consumption further and all of this allows a 99% residency on a Windows idle screen static so they're able to use very little power and just sit there idle and have a lawn standby life basically on the laptops now we haven't tested this these are all claims and their marketing slides but some of this stuff's already out there some of the rise in seven twenty seven hundred you and the twenty five hundred you compose have been out for a little while now there are more coming forward there's the well there's a couple things interesting the r3 is twenty three hundred you the twenty-two hundred you are both coming out those are three point four gigahertz boost and 2.0 gigahertz for twenty three hundred you and two point five for the twenty two hundred you they are twenty three hundred years for core for thread with six compute units and that would be 64 x six for the streaming processors three compute units for the twenty two hundred you to court four threads on that one other than these the roadmap has been laid out as well for for the future for AMD including a card that will compete with Nvidia in the deep learning and machine learning space so that will be a new Vega GPU on seven nanometer process the immediate road map however lists a rise in three mobile APU for January ninth aka now the horizon desktop AP use for February 12th and second gen rise and desktop processors in April which we already knew from our earlier coverage rise in Pro mobile quarter 2 and second gen thread Ripper a lot of you were interested in that second half 2018 second gen rise and pro so I can have 2018 other smaller items here included the launch of the 400 series we already covered that there have been some changes to overclocking some of the motherboard vendors have told us directly that they are focusing on better enabling piece date overclocking not necessarily that useful but soozee a thing that helps and then they have some see be cooler updates and a reduction permanent reduction in pricing for rise in desktop CPUs going forward for the ones that are out now so that's what we have for you just a kind of quick recap of the AMD news as we've seen it you can find some more unique coverage in our gigabyte X 470 motherboard analysis I suppose show floor analysis news video if you're interested subscribe for more CES coverage as always go to patreon.com/scishow gamers Nexus helps out directly and I'll see you all next timeour first major AMD video for the week showcased an X 470 motherboard from gigabyte with the m4 socket of course just with an updated chipset and some changes post rise and launch so as the motherboard vendors have been ironing out their side of Rison and the two has been working on their next iteration of verizon the Rison 2 or as it's been called sometimes Rison plus family of CPUs which will exist on the zen architecture not Zen - but a Zen architecture and also there's news on Raven Ridge the APU integration of the Zen or Rison core complexes and the Vega GPU so I'll be talking about that briefly today before that this coverage is brought to you by Thermaltake and the flow liquid cooler the flow is a 360 millimeter radiator with three ring RGB LED fans you can program the fans our custom lighting through software and then of course benefit from the larger radiator size and cooling performance learn more at the link in the description below this one's gonna be a quick one for us we didn't go to the AMD tech day but we collected the information separately most of our early efforts were focused on the x4 70 motherboard which is one of the first one that's been shown officially you can find that out on our channel by gigabyte so the main information here especially as it pertains to Raven Ridge and the AP is is that the architecture as expected is focused on power savings this is something Andy's been big on it with Rison and the Zen architecture in general and they've had some trouble with it with Vega but combining the two is a bit different than just Vega stand alone on the big d GPU PCB so combining them Vega and risin as expected will communicate via and these high speed interconnect which they call infinity fabric and that also bridges to them the system IO via the chipset and the integrated memory controller the IMC so all these things are bridged together via infinity fabric and then the other kind of major changes there are pertaining to precision boost - which is what Andy is calling its updated version of its boost technology for Zen so the boost technology primarily has hinged on the things you would expect how much current is going to the CPU on the GPU side you're looking at current voltage power consumption and thermals and a lot of the same stuff is true for the CPUs so we look at these elements for precision boost - with the addition of implementing a new volt frequency curve where unlike with Zen one where you had very hard to set frequencies for core utilization you use one core versus two versus all of them there's a very defined almost binary toggle between the frequencies unlike that now it's a more gradual gradient for the frequencies versus the core utilization and the voltage so that's a big change they have charts for that where you can see it's it's kind of a as you increase core count it kind of asymptotes out at the bottom so that's different and it's somewhat mirrors what Intel has been doing not to say that Andy and Intel are copying one another but that this is clearly one of the more optimal ways to configure your core frequencies versus your power consumption another item here is that AMD is looking at which thread is under the most demand so the thread under the most active load is the one that will get the most power sent to it or the most current sent to it so they are modulating based on that not just based on a hard lookup table or something like that further the power between the CPU and the GPU package are allocated based upon the current requirements of the applications running so AMD is using firestrike as an example here on one of their presentation slides which makes a great example because fire strike for 3dmark fire strike it has a few tests it's got graphics tests and physics tests physics uses almost zero graphics compute and graphics uses very little CPU so between the two you can see the power allocation go up for the Vega components the graphics component during the graphics test you see the allocation go up for the CPU component during the physics test which means that they're able to min/max their power and current budget because you can't just push all of it through to both components they don't have the budget for it so you pick and choose to fit the smaller thermal and power consumption limitations of these Raven Ridge chips and that's what they're doing here and they're able to dynamically regulate where the power and the current is going based on the load of the CPU at that time or the GPU as it were and the unfortunately some of their slides aren't all that helpful they are employing the bigger bar better mentality for them where you've got a y-axis with no label well that have a label but no numbers an x-axis with no numbers so it's just sort of nebulous infinitely going in each direction with no actual understanding of what it means which is very unfortunate because it's an interesting slide otherwise but this is something we'll look at more in testing later on just like we talked about when Rison launched AMD is still using LDOS or low dropout regulators for their power management and the LDOS here double as power gates so they can turn things on and off based on the current needs of the system this goes to the extreme of being able to disable 95% of the GPU if it's not needed which of course is really important for any kind of mobile chip because you care a lot about battery life and if you have a GPU sitting there burning power waiting to be used you're just burning battery life so they can paragate down to 95% of it off and they can further power gate with GPU uncor so all of the uncor components things like cash would be mitigated in their consumption as well and save more power power gating for SOC functions is split into two containers so they begin to slide for this one where the containers are the theirs they label them a and B one of them contains the CPU interface the GPU interface and the i/o hub that would be all in one and the other one contains the multimedia hub and the memory controller in a second container and then these can be gated based on which region or container they are in they Assembly load of the system to reduce power consumption further and all of this allows a 99% residency on a Windows idle screen static so they're able to use very little power and just sit there idle and have a lawn standby life basically on the laptops now we haven't tested this these are all claims and their marketing slides but some of this stuff's already out there some of the rise in seven twenty seven hundred you and the twenty five hundred you compose have been out for a little while now there are more coming forward there's the well there's a couple things interesting the r3 is twenty three hundred you the twenty-two hundred you are both coming out those are three point four gigahertz boost and 2.0 gigahertz for twenty three hundred you and two point five for the twenty two hundred you they are twenty three hundred years for core for thread with six compute units and that would be 64 x six for the streaming processors three compute units for the twenty two hundred you to court four threads on that one other than these the roadmap has been laid out as well for for the future for AMD including a card that will compete with Nvidia in the deep learning and machine learning space so that will be a new Vega GPU on seven nanometer process the immediate road map however lists a rise in three mobile APU for January ninth aka now the horizon desktop AP use for February 12th and second gen rise and desktop processors in April which we already knew from our earlier coverage rise in Pro mobile quarter 2 and second gen thread Ripper a lot of you were interested in that second half 2018 second gen rise and pro so I can have 2018 other smaller items here included the launch of the 400 series we already covered that there have been some changes to overclocking some of the motherboard vendors have told us directly that they are focusing on better enabling piece date overclocking not necessarily that useful but soozee a thing that helps and then they have some see be cooler updates and a reduction permanent reduction in pricing for rise in desktop CPUs going forward for the ones that are out now so that's what we have for you just a kind of quick recap of the AMD news as we've seen it you can find some more unique coverage in our gigabyte X 470 motherboard analysis I suppose show floor analysis news video if you're interested subscribe for more CES coverage as always go to patreon.com/scishow gamers Nexus helps out directly and I'll see you all next time\n"