[ad_1]
Intel introduced its Tiger Lake platform yesterday, and the company is clearly eager to fight AMD. A royal tiger roaring somewhere off stage wouldn’t have seemed unusual. The overall positioning of the company suggests that it has something impressive on its hands.
Intel has received a lot of criticism for marching in their place when it comes to feature integrations, but Tiger Lake adds PCIe 4.0 support, fully integrated Thunderbolt 4 support, Intel’s new Xe graphics engine (debuting for the first time), tubes of additional displays, a New 6th Generation Image Processing Unit (Icelake was v4), version 2.0 of their “Gaussian Network Accelerator” (Ice Lake had v1) and support for various low-level deep learning features like AVX-512 . Throw in a massive boost in frequency, courtesy of Intel’s new 10nm “Superfin” technology, and you’ve got Tiger Lake: slimmer, meaner, and ready for a showdown with AMD’s Ryzen 3000 family.
Intel was eager to show benchmarks and discuss performance in various aspects of Tiger Lake. Demonstrations against AMD hardware were plentiful and rarely positive, though you would hardly expect more from a vendor event. Still, the strength of Intel’s presentation and the data it chose to emphasize point to the underlying strength of the company’s product.
For Intel, Tiger Lake is all about performance.
Intel likes to talk about “real world” benchmarking these days, and has shifted its messages to focus on what it considers “real world” applications. However, keep in mind that each of these tests focuses on a single single-function app (aside from Gigapixel, which actually only has one function). Also, note that Topaz products incorporate libraries that favor Intel, and while this is of secondary importance when workloads are running on the GPU, support for AMD GPUs is very early. The good news is that this support will come. I’ve been testing alpha versions of Topaz, including one compatible with the AMD Radeon 5700. It works quite well.
Chart exports are 1.3 times faster in Excel, but it doesn’t tell me what CPU is faster at calculating a series of equations. This is one of the problems with “real world” benchmarking: In an application with so many different capabilities, testing a variety of them is the only way to offer some kind of fair comparison between two products.
Intel Vehicle
Xe comes in with up to 96 EU, 1.35 GHz clock and can generate 48 texels or 24 pixels per clock, which is pretty good for an embedded solution. There’s a larger L3 cache to keep memory pressure off the main bus and a new, higher-efficiency software scoreboard. Nvidia and AMD use software scorecards, until now Intel had relied on a hardware algorithm.
Intel’s Xe is claiming some notable accelerations:
The big question with integrated graphics is always bandwidth: how efficiently will Xe be able to use its limited pool of resources? By the numbers, pretty good:
There are a few places where Xe seems poised to outperform AMD. We’ll see if Intel’s GPU driver team is ready for big things this generation, although AMD has had its own struggles in that department for a long time.
Intel even argues that its integrated graphics can trade hits with discrete cards, although the card in question is roughly the performance of a GeForce GTX 960M. On the other hand, being able to compete with a cheap discrete GPU from 4-5 years ago is truly impressive – I remember the days when Intel’s integrated graphics could barely run games that were 10 years old.
The eleventh family of CPUs, in all its glory, is below:
Intel has decided to move away from TDP as a singular heat dissipation metric and now provides data on the TDP ranges. This is probably a fairer way to capture the likely power consumption of the system at any given time, and it is practically old for AMD or Intel to come up with some new method of trying to communicate energy efficiency to customers. The TDP was an imperfect metric and this probably isn’t perfect either, but it captures range more effectively. In the first place, TDP was never intended to be a power consumption metric; refers to thermal design power – the amount of thermal energy that the cooling solution needs to dissipate to maintain full performance.
Beneath the hoopla, Tiger Lake looks like a very solid part. This may catch some AMD fanboys off guard, but it shouldn’t. Historically, mobile devices have been Intel’s strongest market and the company was serious about reducing its overall power consumption footprint years before AMD. More OEMs have become more optimized for Intel and have more experience with how it handles power. Original equipment manufacturers are more willing to invest in creating world-class Intel solutions.
We’ve already seen this situation start to change last year with the Surface Laptop 3’s debut as a major victory for AMD and other launches of the new 7nm mobile processors this year. AMD is increasing its own efforts and we hope that Zen 3 is imminent. In fact, mobile devices are shaping up to be the battlefield for companies in the imminent future. Expect Intel to try to leverage Tiger Lake’s support for AVX-512 and other deep learning capabilities at all times, while AMD is likely to avoid these comparisons in favor of other metrics.
Now read:
[ad_2]
