The latest gaming tests are in, and it’s time to see how the new generation of Intel CPUs and AMD APUs perform. Join as we benchmark two competing mid-range processors with an integrated GPU against each other.
Introduction
Intel Alder Lake against AMD Cezanne, AMD and Intel integrated graphics performance, two CPUs with integrated graphics compete to discover who gives the greatest value in gaming without a separate video card, is the battle of 2022 in the integrated graphics CPU arena. In order to evaluate gaming performance and computation performance, we have selected the finest and fastest integrated graphics from each camp.
The Intel Core i9-12900K CPU from Intel comes with an integrated UHD Graphics 770 from Intel GPU. The AMD Ryzen 7 5700G APU from AMD comes with AMD Radeon RX Vega 8 graphics built right in. On the quickest integrated graphics desktop CPU and desktop APU in 2022, how well can you game without a separate video card? We shall learn the answer to it.
Specifications of AMD Radeon RX Vega 8 vs. UHD Graphics 770 from Intel
Before delving into each integrated graphics in more detail, we’ll put them into a table to highlight the key differences as clearly as possible. As a result of the differences in how Intel and AMD describe their designs and configurations, it is impossible to compare their architectures directly. Nevertheless, we will try our best to align the relevant comparisons below. You will then have a general understanding of what distinguishes them on a hardware level.
| Specification | UHD Graphics 770 from Intel (12900K) | Radeon RX Vega 8 by AMD (5700G) |
|---|---|---|
| Structure Name | Alder Lake GT1 and Gen 12.2 Intel Xe LP | Vega Raven Ridge | Vega 8 | GCN 5.0 |
| Units of Execution | 32 | 8 |
| Shader Cores/Units | 256 | 512 |
| ROPs | 8 | 8 |
| TMUs | 16 | 32 |
| GPU Cycle Rate | 1550MHz | 2000MHz |
| Image Fill Rate | 1.24 GPixel/s | 16 Gpix/s |
| Fillrate Texture | 24.8 GTex/s | 128 GTexel/s |
| Feature Level for DirectX | 12.1 | 12.1 |
| Model Shader | 6.6 | 6.6 |
| OpenCL | 3.0 OpenCL | 2.1 OpenCL |
| Vulkan | 1.2.203 | 1.3.203 |
| Encoder hardware | Version 8 of Intel QuickSync (QSV) | 2.2 AMD VCN |
UHD Graphics 770 on the Intel Core i9-12900K
The two CPUs that we are examining today are both relatively new additions to the CPU market. Let’s first discuss the Intel Core i9-12900K CPU, which is the 12th generation Intel Core desktop processor. Alder Lake-S 881 is the codename for the 12900K, and its SSPEC is SRL4H. On the socket V LGA1700 Intel platform, based on the Intel 7 manufacturing process, this hybrid CPU design was introduced at the end of October 2021. For this CPU, the launch RCP Pricing in USD $1K is $589.
This CPU is distinctive in that it introduces a brand-new design architecture known as a hybrid architecture that combines Performance Cores (P-core) and Efficiency Cores (E-core) into a single design. Intel Hyperthreading is supported by the P-cores. The architecture is covered in our first evaluation. Eight P-cores and eight E-cores make up the Intel Core i9-12900K processor. This indicates that it has a total of 24 threads and 16 “cores.” The P-Core Max Turbo frequency is 5.1GHz, the E-Core Max Turbo frequency is 3.9GHz, and the Intel Turbo Boost Max 3.0 frequency is 5.2GHz. The maximum turbo power is 241W, while the base power is 125W.
UHD Graphics 770 from Intel
We have reviewed this CPU for its CPU performance, but today is the first time we are focused purely on its integrated graphics (iGPU) performance. The CPU utilizes the UHD Graphics 770 from Intel onboard. The UHD Graphics 770 from Intel is based on Intel’s Xe LP Architecture of the 12th Generation called Gen12 for Alder Lake-S, formerly called 10ESF. Though the UHD 770 in Alder Lake actually goes slightly beyond that and is based on the Generation 12.2 architecture, versus the 12.1 architecture in Rocket Lake CPUs.
The LP in the Structure Name stands for low power and that means this is the low power variant of the Intel Xe GPU architecture, 15W TDP in the 12900K. The Intel Xe GPU architecture should not be confused with Intel’s upcoming dedicated GPU line known as Alchemist, which is a completely new and different beast. The Intel Xe GPU architecture is related to Intel’s Iris Xe Graphics in the mobile space.
There are some new abilities added in the current Generation 12.2 Alder Lake version of the Intel Xe architecture, such as Sampler Feedback, Dual Queue Support, DirectX 12 View Instancing Tier2, and AV1 8-bit and 10-bit fixed-function hardware decoding. It also supports Variable Rate Shading Tier 1. The UHD Graphics 770 from Intel is the top-end variant on Alder Lake CPUs, there is a UHD 730 and UHD 710 versions available, but those are on the earlier Rocket Lake Gen 12.1 architecture. The 12900K, 12700K, and 12600K use the newer Generation 12.2 architecture, and the 12900K itself has the fastest clock speed implementation of the UHD 770.
The UHD Graphics 770 from Intel is naturally also manufactured on the Intel 7 process. The UHD 770 is made up of 32 Execution units and 256 Shading units as Intel defines them. It has a max boost clock in the range of 1450-1550MHz, on the 12900K the boost is the full 1550MHz and this is the highest it will boost at. It has 8 ROPs but only 16 Texture Mapping Units, therefore its Fillrate Texture is a bit low, and the Image Fill Rate is low because of the lower frequency. The architecture does work with AMD FidelityFX Super Resolution (FSR) technology, though it isn’t optimized for it, it works, to varying degrees since FSR doesn’t require vendor support.
One good thing it has going for it is the inclusion of Intel’s latest Version 8 QuickSync hardware decoder/encoder. It adds VP9 12-bit and 12-bit 4:4:4 hardware decoding and HEVC 12-bit 4:2:0, 4:2:2, and 4:4:4 hardware decoding. It also fully supports AV1 decoding. The encoding performance is actually very good, industry-leading in fact. From a software side, it supports DirectX 12.1 and Model Shader 6.6 and is also using 3.0 OpenCL.
Radeon RX Vega 8 AMD Ryzen 7 5700G
The new AMD Ryzen 7 5700G APU, which was just released, is the next child on the block. It contains integrated graphics dubbed AMD Radeon Graphics, much as the 12900K. The AMD Ryzen 7 5700G uses AMD’s most recent architecture, Zen 3, and is a conventionally built CPU. On the AM4 platform, the Ryzen 7 5700G is known as the codename Cezanne. AMD ultimately made this APU available for customer purchase later in 2021 after it was first unveiled in the spring of that year. The APU’s MSRP at launch was $359 and it is based on the TSMC N7 node.
The AMD Ryzen 7 5700G, a conventional CPU, has 8 Cores and 16 Threads, making it comparable to the Ryzen 7 5800X. In reality, it has many similarities with the Ryzen 7 5800X; consider it to be a lower-end version of the 5800X with built-in AMD Radeon Graphics. This CPU has a 4.6GHz maximum turbo or boost frequency. 65W is the TDP.
The AMD Ryzen 5000 series CPUs have the CCD and CCX layouts built into them, and since the CPU is built on the most recent Zen 3 architecture, it also has all of these characteristics. The Ryzen 7 5700G only supports PCI-Express Gen 3.0, not 4.0, which is a significant difference. The performance of the CPU or integrated graphics won’t be affected by this.
Radeon RX Vega 8 by AMD (5000 Series Variant)
The AMD Radeon RX Vega 8 Radeon Vega Graphics are built inside the AMD Ryzen 7 5700G. Unfortunately, the AMD Ryzen 7 5700G does not use the more recent RDNA or even RDNA2 design. Instead, it is still using RDNA from an earlier generation, which is rather dated. We are discussing the Graphics Core Next (GCN) architecture in its entirety, more especially the GCN 5th Generation or GCN 5.0. Unfortunately, the Raven Ridge Vega generation, on which the Radeon RX Vega 8 is based, was only released in 2017.
As the name suggests, AMD Radeon RX Vega 8 is made from 8 Compute Units giving it 512 Streaming Processors or shader cores. It has 8 ROPs and 32 Texture Mapping Units. On the Ryzen 7 5700G the Vega 8 Graphics is clocked at a high 2000MHz (2GHz) clock speed. This gives us a good pixel and Fillrate Texture number, much greater than the UHD Graphics 770 from Intel. The architecture does not support Ray Tracing, and it also does not support Variable Rate Shading at all, whereas the UHD Graphics 770 from Intel does support Tier1 VRS. It does however work just fine with AMD FidelityFX Super Resolution (FSR) technology.
AMD Radeon RX Vega 8 utilizes Video Core Next (VCN) version 2.2. This is actually a step-up from NAVI/RDNA1 which used VCN 2.0. However, it is nowhere near the latest version of VCN which with RDNA2 is VCN 3.0+. VCN 2.2 on Vega 8 does not support AVI1 decoding, so UHD Graphics 770 from Intel has a leg-up in that department.
From a software perspective, the Radeon RX Vega 8 supports DirectX 12.1 and Model Shader 6.6. Interestingly, it does only support 2.1 OpenCL at this time, even with the latest drivers installed 2.1 OpenCL is all we had available to us, therefor the Intel Graphics UHD 770 does support 3.0 OpenCL right now, which you will see in our compute tests makes a difference. The Vulkan version, however, is higher with Vega 8 versus UHD 770.
Naturally, the system memory capacity and bandwidth are used as the main memory source for both integrated graphics solutions. Therefore, the amount of available bandwidth for these graphics processors to use depends entirely on the size and frequency of the RAM that is installed in your system.
