How will the collaboration between Intel and AMD impact gamers? – article

After years of predictable and often repetitive roadmaps in technology leaps, the past 12 months have marked radical leaps in the capabilities of PC hardware, but few could have foreseen the announcement the other day regarding the joining of forces between Intel and AMD to battle Nvidia in the notebook sector. A new series of Kaby Lake G processors will arrive in the coming months, and will combine Intel Core i7 processors with semi-custom GPUs with HMM2 memory. So why did this collaboration come about? And what will it mean for the players?

First of all, let’s not underestimate the GeForce brand’s dominance in notebook graphics: from the GTX 1050, up to the top of the range GTX 1080, the Pascal mobile GPU lineup is remarkably good. These are GPUs on par with their desktop counterparts, or at least just 10 percent slower, but the gaming experience remains pretty much the same as on desktop. Low in terms of energy consumption, Nvidia’s technology paves the way for products like the Razer Blade 14, a notebook practically as powerful as a destkop equipped with quad-core i5 and GTX 1060. It’s a good enough machine to handle gaming 1080p and 1440p, all in a 14-inch laptop. The Blade is arguably the product that offers the best miniaturization of a mid-range Pascal chip, but regardless of that, Intel and AMD are not fully capable of competing with these offerings.

For Intel, the collaboration with AMD means that the new Kaby Lake G line will replace its top-of-the-line Iris / Iris Pro graphics solutions. These offered powerful GPUs paired with 128MB of cache memory – performance was increased, but was still poor in comparison. to Nvidia standards, and drivers have always been a problem. For AMD, the collaboration with AMD offers another victory from the point of view of semiconductor design (nothing comparable to the production of chips for PS4 and Xbox One, but still we talk about important revenues and also the possibility of broadening the horizons with projects future if this proves successful).

The Kaby Lage G processors are certainly fascinating, being equipped with what we suspect is a standard 7th generation i7 processor and a Radeon GPU that uses a new type of custom interconnect to ensure communication between the two chips. Memory bandwidth has always been a problem in integrated graphics, but the use of HBM2 eliminates the question at source: the slides provided by Intel suggest the use of a single module, which should offer a theoretical bandwidth limit of 256GB / s. We expect a more conservative design to be used, but here we are on par with the Radeon RX 580. Considering the specs and price of current HBM2 memory, we think the built-in amount could be 4GB.

The leaked specs that have appeared elsewhere give us an idea of ​​the technical side of the Radeon GPU. Important parameters such as amount of shaders and clock speed indicate a processor with 1536 stream processors and 24 compute units (CU). Operating frequencies, which may be subject to change, are in the 1GHz-1.1GHz range. These are only speculations at the moment, but they are elements that make us presume it is a GPU with a maximum computational power of 3.3 teraflops. From the point of view of quantifying the power, it would be a GPU that is halfway between PlayStation 4 and PlayStation 4 Pro.

But we wonder how much these specs can challenge mobile GeForce solutions. The GTX 1060 delivers excellent 1080p performance in modern titles set at ultra details, and its desktop counterpart, which is slightly faster, plays with the Radeon RX 570 and RX 580, 5.1TF and 6.1TF GPUs respectively. . Granted that these specs are correct, the new chip from the collaboration between Intel and AMD would be more geared towards the performance of the Nvidia GTX 1050 Ti, or just above it. We are of the opinion that Nvidia will not feel threatened in the least for the moment, and its strong earnings thanks to its large share in the mobile sector are likely to remain unchanged.

And so, if it can’t win from a performance standpoint, what is the benefit of the product achieved through the joint efforts of Intel and AMD? The advantage lies in the form factor. In its PR material, Intel notes the space savings guaranteed by Kaby Lake G compared to a standard solution consisting of CPU, GPU and the GDDR5 memory cluster. The documents also discuss the Z dimension: the height of the processor, which the company emphasizes to be really contained. We can say without any doubt that this is the only factor that can push the sales of the Kaby Lake G, the differentiating aspect from all the other current proposals. In practice, Kaby Lake G will allow you to play the latest titles at 1080p and medium-high details, with a very thin and light machine.

There are still many unknowns regarding these processors, specifically to what extent the Radeon GPU has been modified in collaboration with Intel, and also on the underlying technology used. It is almost assumed that this is the recent Vega line, considering that it is the technology that will be used in the next generation AMD APUs, codenamed Raven Ridge. But on the other hand, Vega required a large number of transistors to achieve high clock rates, and this would make the old Polaris a more suitable technology for pushing a GPU that has to run at roughly 1GHz. Time will give us the answers to these questions, but as AMD is working on a contract with Intel, Intel will reveal more details in the coming weeks.

At the moment, it looks like the specs aren’t as exciting as the underlying concept, which is the simple idea that bitter rivals like AMD and Intel are teaming up to create a project together. But our feeling is that the breakthrough of this chip lies in the form factor. As things stand, gaming notebooks like the Razer Blade 14 are an exception, not the norm. Intel’s solution certainly won’t be cheap, but the fact that more compact gaming machines, MacBook-style, are becoming more popular may only be a welcome possibility.