They have to release a Ryzen 7 2800X now, that low power 2800H can't be at the top of the number scheme!

Warrax:

They have to release a Ryzen 7 2800X now, that low power 2800H can't be at the top of the number scheme!

Personally, I don't understand why any of these products go above 2500. They could easily just have the following products: 2500U: 2.2-3.8Ghz, 4c/8t, Vega 11 2400U: 2.0-3.6Ghz, 4c/8t, Vega 8 2300U: 2.0-3.4Ghz, 4c/4t, Vega 6 2100U: 2.0-3.4Ghz, 2c/4t, Vega 3 There's no need for anything more than that, and it removes confusion of numbers that go any higher. We're in a time where CPUs can very reliably scale their performance to meet thermal needs, so when it comes to mobile, all we really need is 1 product per performance tier.

schmidtbag:

Personally, I don't understand why any of these products go above 2500. They could easily just have the following products: 2500U: 2.2-3.8Ghz, 4c/8t, Vega 11 2400U: 2.0-3.6Ghz, 4c/8t, Vega 8 2300U: 2.0-3.4Ghz, 4c/4t, Vega 6 2100U: 2.0-3.4Ghz, 2c/4t, Vega 3 There's no need for anything more than that, and it removes confusion of numbers that go any higher. We're in a time where CPUs can very reliably scale their performance to meet thermal needs, so when it comes to mobile, all we really need is 1 product per performance tier.

Couldn't agree more with you.

schmidtbag:

Personally, I don't understand why any of these products go above 2500. They could easily just have the following products: 2500U: 2.2-3.8Ghz, 4c/8t, Vega 11 2400U: 2.0-3.6Ghz, 4c/8t, Vega 8 2300U: 2.0-3.4Ghz, 4c/4t, Vega 6 2100U: 2.0-3.4Ghz, 2c/4t, Vega 3 There's no need for anything more than that, and it removes confusion of numbers that go any higher. We're in a time where CPUs can very reliably scale their performance to meet thermal needs, so when it comes to mobile, all we really need is 1 product per performance tier.

It is same naming scheme intel used for years. Their Low-end desktop CPU (i3) was marketed in mobile as i3/i5/i7 based on clock. AMD at least differentiates them by GPU.

Fox2232:

It is same naming scheme intel used for years. Their Low-end desktop CPU (i3) was marketed in mobile as i3/i5/i7 based on clock. AMD at least differentiates them by GPU.

Yeah and Intel's naming scheme is notoriously confusing and misleading, so they're not exactly a good reference point. As for AMD differentiating between GPUs, they used to do that with their bulldozer-based APUs, but that doesn't seem to be the case for Ryzen APUs.

schmidtbag:

Personally, I don't understand why any of these products go above 2500. They could easily just have the following products: 2500U: 2.2-3.8Ghz, 4c/8t, Vega 11 2400U: 2.0-3.6Ghz, 4c/8t, Vega 8 2300U: 2.0-3.4Ghz, 4c/4t, Vega 6 2100U: 2.0-3.4Ghz, 2c/4t, Vega 3 There's no need for anything more than that, and it removes confusion of numbers that go any higher. We're in a time where CPUs can very reliably scale their performance to meet thermal needs, so when it comes to mobile, all we really need is 1 product per performance tier.

It's about TDP. CPUs don't just scale to meet thermal needs, they scale to meet TDP needs, which is based on what the cooling system can handle. This is important because laptops have limited cooling capability. A 15W chip essentially means the cooling system must be able to dissipate 15W of heat. If the CPU ignored its TDP limit and only cared about its own thermals, it would eventually turn the inside of the laptop into an oven and cook the other components too (very bad for the battery especially). H parts have 35W TDP, which means it has higher cooling capacity (likely more heatpipes and fans) and means it can sustain higher boost clocks for longer periods of time, perhaps even indefinitely. U parts have 15W which means it can only do boost in short bursts as to avoid overwhelming the cooling system. Also since these are APUs, it means it can sustain higher GPU+CPU capability since they share a common package TDP.

Mundosold:

It's about TDP. CPUs don't just scale to meet thermal needs, they scale to meet TDP needs, which is based on what the cooling system can handle. This is important because laptops have limited cooling capability. A 15W chip essentially means the cooling system must be able to dissipate 15W of heat. If the CPU ignored its TDP limit and only cared about its own thermals, it would eventually turn the inside of the laptop into an oven and cook the other components too (very bad for the battery especially). H parts have 35W TDP, which means it has higher cooling capacity (likely more heatpipes and fans) and means it can sustain higher boost clocks for longer periods of time, perhaps even indefinitely. U parts have 15W which means it can only do boost in short bursts as to avoid overwhelming the cooling system. Also since these are APUs, it means it can sustain higher GPU+CPU capability since they share a common package TDP.

Actually, CPUs do in fact scale to TDP speeds. The TDP rating is for base clocks only. Notice the 2600H - the only difference between that and 2500U is the base clock (which in turn affects the TDP). They have the same boost clocks, caches, GPU, core/thread count, and so on. Both CPUs will reach their rated boost clocks (and the same performance level) if the thermals and power consumption keep up. The point is, both APUs are equally capable chips. The only thing that separates them is the cooling method. So at this point, the only thing that matters to the consumer is which one can retain boost clocks the longest.

schmidtbag:

Actually, CPUs do in fact scale to TDP speeds. The TDP rating is for base clocks only. Notice the 2600H - the only difference between that and 2500U is the base clock (which in turn affects the TDP). They have the same boost clocks, caches, GPU, core/thread count, and so on. Both CPUs will reach their rated boost clocks (and the same performance level) if the thermals and power consumption keep up. The point is, both APUs are equally capable chips. The only thing that separates them is the cooling method. So at this point, the only thing that matters to the consumer is which one can retain boost clocks the longest.

TDP determines the speed the chip runs at, but its for the chip itself, no matter if its base clock or burst. Not all workloads are created equal. Being at "100%" because of a crazy out of control process is not the same as "100%" because of Prime95. The total work the CPU doing is what determines the heat output. GPUs are the same way, when the mining craze was big there were certain algos that were lighter than playing a modern game, and other algos more intense than Furmark. Despite both using your GPU "100%", you would notice the GPU settle on different voltages, temperatures and clock rates depending on the mining algo. This is also why overclock mobos for Intel have a different AVX offset. AVX instructions are particularly intensive and will cause an otherwise stable CPU to fail. The idea is to lower the overclock when AVX is being used to increase stability. With some workloads, a CPU can not even maintain BASE CLOCK speed without going over TDP. The CPU thermals themselves can be fine, but because it is taking so much juice to do the work, the CPU downclocks itself. With 15W vs 35W, the 35W can use a lot more juice before TDP downclocking. To the point it may never occur. With 15W, it would likely frequently occur, since those chips are designed for bursty workloads often seen with laptops - rendering a webpage using turbo speeds for 3-4 seconds, then mostly resting while the user browses the page, for example. They want apps and websites to display quickly, and the CPU model is "hurry up to wait". But in a gaming laptop, the CPU will be consistently used for long periods of time, rather than just short bursts, and you want that higher TDP limit to maintain the burst speeds.