You’ll be hearing a lot about Meteor Lake over the next few days and weeks; These new chips are some of the most exciting chips from Intel in some time. I’d say that despite their intended use as low power processors in ultra-thin laptops, not even in gaming PCs. Some of the new features packed into these different chips could come in handy for the next generation of gaming processors.

For starters, the same separation that is central to the design of Meteor Lake. The most notable change to Meteor Lake versus, say, a Raptor Lake chip (or really any of Intel’s client processors ever) is how its silicon is divided up into different components.

The four (of five types) tiles that make up each Meteor Lake processor are:

  • count tile – Consisting of both P-Core (Redwood Cove) and E-Core (Crestmont). Built on Intel 4 process node.
  • SoC tile – Consisting of traditional Uncore components, a new AI engine called an NPU, a media block, memory controller and low power island – more on this in a moment. Manufactured by TSMC.
  • graphics tile – Intel Xe-LPG GPU with 8 Xe-cores, including ray tracing support and many shared features with existing Xe-HPG graphics cards (Alchemist). Built on TSMC N5 process node.
  • io tile – Thunderbolt and PCIe Gen 5 connectivity. Manufactured by TSMC.
  • Base Tile/Die/Interposer – Intel doesn’t always refer to it as one of the tiles of Meteor Lake – it’s an interposer – but it’s absolutely critical to putting this chip together.

Let’s look at the Compute tile in more detail, and I’ll start with the least exciting information for PC gamers. In terms of raw p-core performance, Redwood Cove does not offer much improvement over the instructions per clock (IPC) of the existing Golden Cove cores found in today’s 13th generation processors. However, Intel suggests there have been some improvements, including a larger L2 cache, increased bandwidth per core, better efficiency, and better utilization via the thread director, but I’m not expecting any major improvements in performance here. .

Compute Tile is built on an Intel 4 process node. Intel is talking up the new node’s 2x high-performance library scaling, EUV lithography, more robust metal stack, better transistor design, and a number of improvements designed to ensure that this node runs without a hitch. One thing that’s important to note for gamers anyway is that the main focus of this initial Intel 4 tile has been on efficiency.

I asked Bill Grimm, Intel VP of logic technology development, if we might see some physical impact on clock speeds from the step-up to the new process node, to which he responded “We see power efficiency as a priority going forward.” Have been.” This doesn’t rule out more performance-based tiles on Intel 4, but there’s certainly not a focus on mobile chip speed for it.

(Image credit: Intel)

Since the P-cores won’t surprise gamers with their impressive performance, the new E-cores, known as Crestmont, may be of interest instead. Intel says the biggest IPC improvements have been seen in the e-Core, though it won’t say how much to expect. The company says that enhanced branch prediction, feedback from the thread director, and AI acceleration are all in the mix for these compact cores, which should mean more workloads will remain on these lower-powered cores for longer.

One of the more fascinating parts of Meteor Lake is actually where Intel is placing these new E-Cores – the two appear within the new Low Power Island on the SoC tile.

So let’s talk about that low power island. It’s a simple idea but could prove to be a very practical solution for power-thirsty laptops. It is a block that contains everything a chip needs for lightweight workloads: two low-power e-cores, memory subsystem, power controller (PUNIT), IPU, media engine, display engine, and AI acceleration block. All are connected through the Scalable Fabric. Since this low power island can run independently of the compute tile, it can be turned off completely when not needed, thus, in theory, saving considerable power. Again, Intel won’t say how much, or provide any information about battery life.

Close up of Intel Meteor Lake assembly.

(Image credit: Intel)

Although you won’t get any gaming with a low-power e-Core, I think the idea of ​​a low power island could come in handy for extending laptop battery life, even for gamers. I’m just hoping it reduces power consumption as much as a real laptop would when loaded.

While Intel has been skimpy on actual processor details, from recent leaks and rumors we can fairly safely assume that the top Meteor Lake chip will be a 6+8(+2) chip. To put it plainly, the top Meteor Lake chip is expected to have six P-cores, eight E-cores (compute tile) and 2 E-cores (SOC tile).

The key to linking all those cores together are two fabrics – the connections that run between tiles and the individual blocks within those tiles – called the NoC (network-on-a-chip) and the IO fabric.

“This network-on-a-chip fabric is meant to meet our demands for higher-performance devices as well as provide efficient access to memory for those lower power SoC cores,” says Intel Fellow Mikal Hunsaker.

“Our solution for better IO efficiency was to take our graphics block and media block and take them out of the ring. We put the graphics on a tile and it’s connected to the NoC. We took the media out of the graphics, we put it on the NoC. And our other devices that were on our IO fabric that are now on the NOC include imaging, displays, and our newly introduced NPUs.”

The NoC connects to the compute tile and graphics tile, although they do not need to be turned on for it to work. With those low-power e-cores on the SoC tile and taking media, display, and imaging away from the graphics tile, Intel can just run the SoC tile and be left with a functioning CPU on its own.

Meanwhile the IO fabric includes blocks like Wi-Fi and Bluetooth, security, Ethernet, audio, PCIe, SATA, USB, and is responsible for hooking into the IO tile.

Intel Meteor Lake tile diagram with both NoC and IO fabrics noted.

(Image credit: Intel)

The thread director is also important in keeping Meteor Lake active. Core Utilization technology, first introduced in Alder Lake and today’s Raptor Lake desktop processors, is what Intel calls a “hinting mechanism” for Windows, as it is the OS that actually makes all the calls about which threads are running which cores. But let’s go. Intel has improved the thread director with Meteor Lake to better allocate work between all three core types (P-core, e-core, LP e-core).

Intel gave this simple idea of ​​how the thread director has changed.

  • Meteor Lake: Try to include threads on the SoC e-core -> if threads cannot be included then proceed to calculate e-cores -> if threads can benefit then proceed to calculate p-cores. proceed to
  • Raptor Lake: High QoS threads move to P-core -> Low QoS threads move to E-core -> periodically move and reclassify E-core threads to P-core

Moving away from the core configuration, a more direct benefit for gamers with these new Meteor Lake chips is the new Xe graphics tile. Built on the Xe-LPG architecture, and technically brand new to these mobile chips, Intel’s graphics guru Tom Peterson describes it as like the With power-savvy design features. ,

Long story short, this is more or less the architecture we already know for gaming. This means that driver packages for both discrete and integrated graphics can be very similar, if not identical.

Peterson tells me this means that all the work Intel has done to overhaul its Arch driver package for gaming will also be copied over to Meteor Lake’s iGPUs, including new APIs and older APIs like DX9. Are. Although Peterson admits that these drivers will likely release on different schedules – Meteor Lake will likely not get support for day-zero or day-one games.

For us gamers, an iGPU on a low-power mobile part is very exciting, especially one with only eight Xe-cores. I recently tried out Meteor Lake in Dying Light 2 at an event, and there’s something to be said for its performance compared to previous iGPUs. Still, this type of integrated chip paves the way for bigger, better iGPUs on future processors.

“We can use different processes on different sized chiplets in the product stack… If you think about it, and I’m not talking about SKUs, but you can look at different variations of GPUs and CPU complexes. “Can imagine different forms that will address different segments,” Peterson tells me.

Now, this is an attractive idea: a gaming specific mobile chip with a budget or mid-range GPU loaded on top of it. Peterson says it’s the kind of budget GPU replacement that Intel is also trying to integrate.

“As we integrate, and as we get faster and faster, you disrupt the low end of the discrete market. And that’s absolutely intentional.”

How you ensure that any large iGPU has access to plenty of fast memory is another challenge, but there seems to be scope for all kinds of configurations due to the separation.

Close up of Intel Meteor Lake assembly.

(Image credit: Intel)

The separation in Meteor Lake is made possible by Foveros, Intel’s 3D chip packaging technology. If you’re not already familiar, I published a Fovros explainer video about it on the PC Gamer TikTok to give you an idea of ​​how it works.

“I think this is our most advanced package that we’ve ever introduced in the history of Intel. I think it’s probably, arguably, one of the most complex that the industry has ever seen,” said Fovaros, director of technology development programs. Director Patrick Stover said, says.

@pcgamer_mag ♬ Original Sound – PC Gamer

The main thing to note is that there is something called Base Wafer on Meteor Lake that completes the package. This is known as the base tile/die – so when Intel says there are only four tiles, it actually amounts to five. Perhaps the reason it is not always mentioned is that it is not full of complex logic; The base tile is the glue that holds the other chips together, and provides all the critical features through silicon vias or TSVs to provide power to the rest of the chip from the package.

The good thing to know about this base die is that its configurable size and layout is what makes Fovros interesting for gaming processors. In theory, any kind of shaped tile could be placed on top of it, whether Intel wants to produce it or buy it from an outside foundry.

However Fovros comes with its own challenges, including more limited perimeter available to break external IO connections and additional latency between tiles. Still, “from the SoC architect’s perspective,” says Stover, “the benefits of Faveros far outweigh the challenges.”

We’ll likely see this type of stacked chip push the boat out in terms of gaming CPU performance and features. More cores and more cache can be accommodated depending on the need. Cache will have to be addressed with future gaming CPUs, as according to Intel, Meteor Lake moves the needle a bit, but with future Fovros processors on track to match AMD’s 3D V-Cache chips in terms of gaming performance. There is scope.

So while you’re unlikely to buy a Meteor Lake laptop specifically for gaming – even Peterson admits that many Intel-powered laptops will still have Nvidia GPUs fitted – at least there’s plenty here that will. Intel’s processors and iGPUs are promising for the future of gaming.

Source