Exploring CPU power scaling on the Core i9-13900K and Ryzen 9 7950X

One of the biggest rumors on social media and Reddit is how hot and power-hungry CPUs have become over the years. Once upon a time, flagship x86 CPUs didn’t even need a cooler, and now they can saturate the entire cooler.Thankfully, it’s not at the level of a nuclear reactor, as the meme says – but as the kids are now saying, it’s not either nothing burger. Designing for higher TDP and greater power consumption allows chipmakers to push the envelope in terms of performance—no easy feat in a post-Dennard world—but it certainly creates some concerns about power consumption and power consumption along the way. A new problem with heat. For better or worse, the latest flagship chips from AMD and Intel are proof of that.

But despite these general trends, it doesn’t mean that high-performance desktop CPUs also need to be power-hungry. In our review of AMD’s Ryzen 9 7950X, our tests showed that even with the current pedestrian power cap of 65 watts, the 7950X can deliver significant performance at less than half the normal power consumption.

IMHO power efficiency has been a hot topic of late as enthusiasts looking to save money on energy bills (especially in Europe) while still enjoying fast CPU performance, looking for other ways to get the most out of AMD’s Raphael and Intel’s Raptor Lake- The silicon capabilities of the S platform, in addition to filling the chip with as many joules as possible. All the while, the small form factor market remains a stalwart outpost for high-efficiency chips, where cooler chips are critical to building smaller, more compact systems that can forego the need for large cooling systems.

All of which is to say that while it’s nice to see peak performance bumped, often focusing on how an unlocked chip scales when overclocked (pushing CPU frequency and CPU VCore voltage) is just one way to look at overall CPU performance. So today we’re going to go the other way and look at the overall energy efficiency of the user — see what happens when we target the sweet spot on the voltage/frequency curve. To that end, today we’re looking at how the Intel Core i9-13900K and AMD Ryzen 9 7950X perform at different power levels and see what kind of benefits power scaling can provide compared to stock setups.

The changing definition of TDP (Thermal Design Power)

One of the most frustrating topics surrounding processors involves TDP, or more specifically, what exactly TDP is. While the term has a fixed definition—thermal design power—what it actually means with respect to product specifications varies from hardware manufacturer to hardware manufacturer. Even Intel and AMD define TDP slightly differently (despite sharing the x86 CPU market), making it a complex and sometimes confusing term.In short, TDP is no longer a firm guide to the maximum power draw of a particular processor, or even how much heat a processor can dissipate

We’ve done a number of columns on this subject over the years, so here are some previous posts on the topic of power and TDP:

Starting with Intel, while the company assigns official TDP figures to their processors, the reality is that their unlocked processors are not limited to one assigned power figure. At best, Intel provides a set of TDP values ​​based on the base frequency of their chips, with Power Level 1 (PL1) and Power Level 2 (PL2) ratings layered on top of that to define those most important power parameters (and all Advertised) turbo clock speed.

At the same time, with the release of the Ryzen 7000-series processors and accompanying AM5 platform, it has become more difficult to see how AMD defines its TDP figures. While AMD does give default TDP figures for the Ryzen 7000 series, it’s no stricter than Intel. Another element is what AMD calls Package Power Tracking, or PPT for short. The defined TDP value is assigned to its processor based on the default frequency level, while the PPT is defined as the limit that the actual socket can provide in terms of power (and the power that a flagship chip tuned for maximum performance will try to consume).

All in all, on almost any consumer board with an adequate cooler, use any Unlocked processors, including the two we’re focusing on today, will circumvent the chip’s base TDP rating to deliver the highest turbo clock speed possible, limited only by temperature and power delivery constraints. So taking Ryzen 9 7950X as an example, although the default TDP is 170W, the actual PPT wattage is defined as 230W (continuous current 160A).

However, the good news is that these power limits are easily adjustable by the user on both AMD and Intel platforms. Many motherboards for both platforms offer a plethora of adjustable values ​​to raise and lower these limits, and even in Windows there are packages like AMD’s Ryzen Master and Intel’s XTU that make the necessary changes.

Find the sweet spot between performance, power and heat

For its latest line of processors, AMD offers a new feature called ECO Mode, which reduces processors with a rated TDP of 170 W to 105 W (or 65 W for processors with a rated TDP of 105 W). When we reviewed the AMD Ryzen 9 7950X processor we found some interesting results when used at 65 W instead of its default value of 170 W.

The most important thing we observed was that the Ryzen 9 7950X processor outperformed the Intel Core i9-12900K at 65 W on default settings in the CineBench R23 multi-threaded test, which is simply impressive. Since then, Intel has released the 13th generation Core series code-named Raptor Lake, so the overall playing field has been leveled out a bit. But it still shows the potential value of reducing the power consumption of flagship CPUs, as the 7950X can deliver 80% of its peak performance at less than half the power consumption.

These findings, in turn, inspired the argument behind this article: we wanted to see how the Core i9-13900K and Ryzen 9 7950X performed when used with various power limits applied. So for this article, we’ll continue testing this scenario, running AMD and Intel’s flagship CPUs at various power levels.

We selected several power limits/PPTs for testing, including the following:

  • Intel Core i9-13900K at default settings (125 W base, 253 W turbo)
  • AMD Ryzen 9 7950X at default settings (170 W base, 230 W PPT)
  • Both are 125 W
  • Both are 105 W
  • Both are 65 W
  • Both are 35 W

In all of this, it’s important to note that there is one key architectural difference between the Intel Core i9-13900K and the AMD Ryzen 9 7950X: the chiplet. The Core i9-13900K is a monolithic design, with the cache, core, memory controller (IMC), and iGPU all built into a single die, while AMD’s Ryzen 9 7950X is built around a chiplet design. As such, the 7950X uses three different chiplets for all of its major functions, two octa-core core duo chips (CCDs) and an IO chip (IOD) that contains the memory controller, PCIe, and graphics.

Screenshot of AMD Ryzen Master overclocking software test

This means that when setting a defined power limit with the Core i9-13900K, it limits the power for all mentioned variables. And doing the same thing on the Ryzen 9 7950X won’t have the exact same effect, since the IOD and CCD run semi-independently, despite being interconnected via AMD’s Infinity Fabric. To properly power limit the Ryzen 9 7950X, the package power tracking limit needs to be applied to properly limit the power, eg setting a value of 125 W on the PPT, not just on the CPU cores.

Test bench and setup

Since Intel and AMD use different platforms and sockets, we used MSI MPG Z790 Carbon WIFI for Core i9-13900K and GIGABYTE X670E Aorus Master for Ryzen 9 7950X. To try to keep the playing field as level as possible, we used an SK Hynix Platinum P41 2TB NVMe storage drive and an AMD Radeon RX 6950 XT graphics card.

Normally we would rely on the JEDEC memory settings defined for each platform, but in this case we chose to test both platforms using the same SK Hynix DDR5-5600B CL46 memory configuration. This is the memory we used for our Intel Core i9-13900K review, but it’s faster than the highest official JEDEC setting supported by AMD’s Ryzen 7000-series platform (DDR5-5200). Either way, we wanted to do a like-for-like comparison as much as possible, including keeping the memory at a constant performance and power level, to investigate the benefits, if any, of limiting system power consumption.

Power Scaling Test Bench (DDR5)
Intel and AMD
processor Core i9-13900K ($589)
24 cores, 32 threads
125W base power, 253W turbo

Ryzen 9 7950X ($699)
16 cores, 32 threads
170W Pedestal, 230W PPT

motherboard MSI MPG Z790 Carbon Fiber WIFI (13900K)
Gigabyte X670E Aorus Master (7950X)
memory SK Hynix
DDR5-5600B CL46
cool down EKWB EK-AIO Elite 360 ​​D-RGB 360mm
storage SK Hynix Platinum P41 2TB PCIe 4.0 x4
power supply Corsair HX1000
graphics card AMD Radeon RX 6950 XT, 31.0.12019
operating system Windows 11 22H2

Regarding benchmarks, we chose a small subset of rendering and encoding tests because we believe these will highlight differences in processor performance at different power levels. For games, we chose Total War: Warhammer 3, as it’s a notoriously CPU-heavy game, while Borderlands 3 is more graphically demanding. We think it’s wise to test each one, though if anyone wants to request more titles from our CPU 2023 benchmark suite, feel free, we’ll add data as we find time to collect and organize it.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *