Core-to-Core, Cache Latency, Ramp

For some of our standard tests, we look at how the CPU performs in a series of synthetic workloads to example any microarchitectural changes or differences. This includes our core-to-core latency test, a cache latency sweep across the memory space, and a ramp test to see how quick a system runs from idle to load.

Core-to-Core

Inside the chip are eight cores connected through a bi-directional ring, each direction capable of transmitting 32 bytes per cycle. In this test we test how long it takes to probe an L3 cache line from a different core on the chip and return the result.

For two threads on the same core, we’re seeing a 7 nanosecond difference, whereas for two separate cores we’re seeing a latency from 15.5 nanoseconds up to 21.2 nanoseconds, which is a wide gap. Finding out exactly how much each jump takes is a bit tricky, as the overall time is reliant on the frequency of the core, of the cache, and of the fabric over the time of the test. It also doesn’t tell us if there is anything else on the ring aside from the cores, as there is also going to be some form of external connectivity to other elements of the SoC.

However, compared to the Zen3 numbers we saw on the Ryzen 9 5980HS, they are practically the same.

Cache Latency Ramp

This test showcases the access latency at all the points in the cache hierarchy for a single core. We start at 2 KiB, and probe the latency all the way through to 256 MB, which for most CPUs sits inside the DRAM.

Part of this test helps us understand the range of latencies for accessing a given level of cache, but also the transition between the cache levels gives insight into how different parts of the cache microarchitecture work, such as TLBs. As CPU microarchitects look at interesting and novel ways to design caches upon caches inside caches, this basic test proves to be very valuable.

The data here again mirrors exactly what we saw with the previous generation on Zen3.

Frequency Ramp

Both AMD and Intel over the past few years have introduced features to their processors that speed up the time from when a CPU moves from idle into a high-powered state. The effect of this means that users can get peak performance quicker, but the biggest knock-on effect for this is with battery life in mobile devices, especially if a system can turbo up quick and turbo down quick, ensuring that it stays in the lowest and most efficient power state for as long as possible.

Intel’s technology is called SpeedShift, although SpeedShift was not enabled until Skylake.

One of the issues though with this technology is that sometimes the adjustments in frequency can be so fast, software cannot detect them. If the frequency is changing on the order of microseconds, but your software is only probing frequency in milliseconds (or seconds), then quick changes will be missed. Not only that, as an observer probing the frequency, you could be affecting the actual turbo performance. When the CPU is changing frequency, it essentially has to pause all compute while it aligns the frequency rate of the whole core.

We wrote an extensive review analysis piece on this, called ‘Reaching for Turbo: Aligning Perception with AMD’s Frequency Metrics’, due to an issue where users were not observing the peak turbo speeds for AMD’s processors.

We got around the issue by making the frequency probing the workload causing the turbo. The software is able to detect frequency adjustments on a microsecond scale, so we can see how well a system can get to those boost frequencies. Our Frequency Ramp tool has already been in use in a number of reviews.

A ramp time of within one millisecond is as expected for modern AMD platforms, although we didn’t see the high 4.9 GHz that AMD has listed this processor as being able to obtain. We saw it hit that frequency in a number of tests, but not this one. AMD’s previous generation took a couple of milliseconds to hit around the 4.0 GHz mark, but then another 16 milliseconds to go full speed. We didn’t see it in this test, perhaps due to some of the new measurements AMD is doing on core workload and power. We will have to try this on a different AMD Ryzen 6000 Mobile system to see if we get the same result.

AMD's Ryzen 9 6900HS Rembrandt Benchmarked Power Consumption
Comments Locked

92 Comments

View All Comments

  • ingwe - Tuesday, March 1, 2022 - link

    While I understand not looking at battery life in this, not looking at graphics is a big miss.
  • _abit - Tuesday, March 1, 2022 - link

    Back to pimping intel ever so gently
  • SarahKerrigan - Tuesday, March 1, 2022 - link

    In 35/45W laptops, including this one, a dGPU will generally be present, making it a little moot. I expect more of a focus on graphics in the 15W review.
  • Kangal - Wednesday, March 2, 2022 - link

    Not at all.
    Most laptops (+80%) are below 45W TDP and they don't have a dGPU, and rely on the iGPU. It's a shame AMD has dragged their heels in this segment. For that reason, RDNA-2 is a substantial upgrade for most laptop users, but if you want a thick, heavy, Gaming Laptop then you are better off with the (12th-gen) Intel CPU and Nvidia dGPU (GeForce 30).

    Besides, I see AMD's tech as eras:
    2015-era, 16nm, Zen1, Vega Graphics
    2018-era, 8nm, Zen2, RDNA-1 graphics
    2021-era, 6nm, Zen3, RDNA-2 graphics
    2024-era, 4nm, Zen4, RDNA-3 graphics
    ...roughly speaking (obviously years don't align)

    And someone earlier asked how do these different GPU architectures compare. It's hard to do a true Potatoes-to-Potatoes comparison. However, from my understanding of the latest options it goes:

    Qualcomm Adreno 7th-gen > Apple Graphics M1 > PowerVr IMG B-series > AMD RDNA-2 > Nvidia Ampere > ARM Mali Valhall 4th-gen > Intel Xe 1st-gen.

    ...obviously even the latest 4nm Adreno 730 when maxed out at 10W TDP, is no match against an older 8nm RTX 3050 that is thermal limited to only 100W TDP.
  • cbutters - Tuesday, March 1, 2022 - link

    I hate to say it, but I agree.... this is literally one of the biggest step increases in iGPU performance; EVER, and nobody is talking about it. Why? Does Intel have input on how these articles are written? Or does it contribute to the ad revenue and people are wary of disrupting that? Its literally the MOST interesting thing about this CPU.
  • JasonMZW20 - Tuesday, March 1, 2022 - link

    Mostly because these systems ship with dGPUs. The iGPU in mobile Ryzen 6000 is a nice upgrade and simply demolishes anything Intel offers, currently.

    The mainstream 15-28W article should focus on iGPU, as these won't ship with dGPUs, usually. This is the meat of the market, and a good iGPU is critical to a good experience.

    AMD's mobile strategy seems to be a quick-iterative design. Renoir and Cezanne were nearly on top of each other, as Cezanne came back from the fab just as Renoir shipped. So, with mobile Ryzen 6000/Rembrandt, AMD offers a new iGPU+(LP)DDR5 rather than new CPU cores, plus SoC optimizations overall.

    The one thing that bothers me about the RDNA2 iGPU is that AMD disables an entire shader array in the 6600U instead of simply turning off 2 WGPs (4 CUs). So, there's a sharp performance drop between the two models and 6600U will be the primary volume seller, I think. 12 CUs to 6 CUs, instead of 12 -> 8 -> 6.
  • DannyH246 - Tuesday, March 1, 2022 - link

    Honestly it’s been like this for a while now. Just go to the home page and count how many Intel marketing articles we’ve had over the last couple months. Now we get a half arsed joke of a review like this on AMD hardware. Obviously in Intel’s pocket.
  • 29a - Tuesday, March 1, 2022 - link

    Typical half assed AMD article.
  • DannyH246 - Tuesday, March 1, 2022 - link

    Next headline article on www.IntelTech.com….
    Breaking news Intel CEO Gelsinger breaks wind.
  • Qasar - Wednesday, March 2, 2022 - link

    DannyH246 i see you cry about this all the time, if anandtech is that bad, WHY do you keep coming here ? is it just to whine and cry?? im sure you will just reply with some sort of BS, but it HAS been stated before, there are times when most articles are intel, and others are AMD, its just the way the cycles go.

Log in

Don't have an account? Sign up now