Basemark’s Cross-Platform GPU Test: iPhone 14 Pro Beats Steam Deck
base mark introduced it is new GPUScore Holy Path A cross-platform benchmark that allows users to compare completely different GPU architectures running on different operating systems. Our GPU benchmark tier is limited to Windows 11 testing, but what if you want to know how the best graphics cards compare to smartphones?
To name a few, Sacred Path shows how Apple’s A16 GPU compares to Steam Deck’s GPU and Samsung’s Galaxy S22+ GPU, and how Nvidia’s GeForce RTX 3090’s performance compares to Intel’s Arc A770 and It shows how it relates to the performance of Apple’s M1 Max. Of course, there are caveats for benchmarks that need to work across a wide range of hardware capabilities. But let’s see what the numbers say.
Cross-platform, cross-device benchmark
There are many leading hardware graphics platforms for PCs and mobile devices. These include AMD’s integrated and discrete GPUs, Apple’s PowerVR-derived integrated graphics processors, Arm’s Mali graphics solutions, Imagination’s PowerVR-based graphics solutions, Intel’s standalone Arc Alchemist and integrated Xe-LP GPUs, Nvidia’s ‘s discrete graphics processor, including Qualcomm’s Adreno graphics for mobile applications. , Samsung RDNA 2-based embedded GPU cores, and numerous other platforms.
Officially, all of these GPUs are designed to enable high-performance gaming, but high-end mobiles and high-end PCs are obviously very different targets. All architectures have their strengths and weaknesses and often run on different operating systems with different APIs (Application Programming Interfaces). So it’s very difficult to tell how one GPU stacks up with another. Because it’s almost impossible to make an apple-to-apple or apple-to-desktop comparison.
Basemark’s cross-platform Sacred Path is designed to clear up some of the confusion as it runs the exact same workload on all the various GPUs (tests were run using the official native resolution of 2560×1440). ). Various performance optimization techniques are supported, including variable rate shading, advanced custom shaders, ground truth ambient occlusion (GTAO), and temporal anti-aliasing to make the benchmarks more relevant for different devices. I’m here.
Reproducible but variable
After several hours of running the Sacred Path benchmark on several Apple devices, we found that the program generally produced consistent, reproducible results that correlated with the GPU capabilities of the devices, although sometimes the results fluctuated wildly. understood. I’m talking more than 50% here. Apple’s mobile platform may adapt to the workload and improve performance, but smartphones and tablets aren’t in our wheelhouse, so I’ll leave the proper research to others.
PowerBoard 4.0 database from Basemark There are already enough submissions at the official rendering resolution of 2560×2440 to allow comparison across many different platforms. To give you a better perspective, we’ve compared the best results achieved by each graphics processor, but CPUs and APIs aren’t taken into account.
Unfortunately, these results can vary considerably, so PowerBoard 4.0 submissions should be salted to the appropriate level. Also note that results from smartphones, notebooks and tablets may vary significantly due to differences in cooling and ambient temperature. All of this was done under test lab conditions by no stretch of the imagination.
high end gaming GPU
Basemark’s Sacred Path benchmark isn’t particularly impressive compared to modern games, but it does take a lot of resources to run at decent speeds. (for example, 1840 points means 18.4 FPS). Luckily, even existing GPUs from AMD and Nvidia can beat this benchmark very easily, delivering 200-280 FPS. This also means that GPU utilization may not be close to 100%. This is a problem when trying to create a “universal” GPU benchmark.
GPUs | Score | device | APIs | OS |
GeForce RTX 3090 | 28,270 | Desktop, AMD Ryzen 9 5900X | Direct X 12 | windows 11 pro |
GeForce RTX 3080 Ti | 25,050 | Desktop, AMD Ryzen 7950X | Vulcan 1.3 | windows 11 pro |
Radeon RX 6900 XT | 24,244 | Desktop, Intel Core i5-12600KF | Vulcan 1.3 | windows 10 pro |
GeForce RTX 3080 10GB | 22,489 | Desktop, AMD Ryzen 9 5900X | Vulcan 1.3 | windows 10 pro |
Radeon RX 6800 XT | 20,569 | Desktop, AMD Ryzen 5800X | Vulcan 1.3 | windows 10 pro |
Judging by our proven results on high-end gaming graphics cards, we can conclude that Basemark’s Sacred Path likes raw compute performance. The more FP32 teraflops, the better the results. AMD’s Radeon RX 6900 XT is a close match to the GeForce RTX 3080 Ti, but the GeForce RTX 3090 is far behind.
Unfortunately, there are no RTX 4090 submissions yet. Jarred did that, but it failed to submit (perhaps the system has a whitelist of known GPUs and 4090 is not on that list). A manual (custom) test using official settings gave him a result of 48,709 in about 4 seconds of running. This seems to be at least in the correct range.At the same time, the test Done in seconds This means that the GPU was not remotely warmed up before the test finished. (Where’s my salt shaker?)
Mid-range and high-end mobile gaming GPUs
GPUs | Score | device | APIs | OS |
GeForce RTX 3070 Ti | 17980 | Desktop PC Core i7-12700F | Vulcan 1.3 | windows 11 pro |
GeForce RTX 3070 | 15967 | Desktop, Core i7-11700K | Vulcan 1.3 | windows 11 pro |
Radeon RX 6700 XT | 15514 | Desktop, Ryzen 7 3700X | Vulcan 1.3 | Windows 10 enterprise |
GeForce RTX 3060 Ti | 14010 | Desktop, Core i3-10105F | Vulcan 1.3 | windows 11 pro |
GeForce RTX 3080 Laptop | 12277 | Acer Nitro AN515-45, Ryzen 9 5900HX | Vulcan 1.3 | windows 11 home |
GeForce RTX 3070 Laptop | 12524 | ASUS ROG Strix G533QR, Ryzen 9 5900HX | Direct X 12 | windows 11 home |
Mid-range and high-end mobile gaming graphics processors are not as powerful as high-end desktop cards. Nevertheless, AMD’s Navi 22, Intel’s ACM-G10, and Nvidia’s GA104 all work very well. On another note, Intel’s Arc A770 lags Nvidia’s GeForce RTX 3060 Ti by 11%.
On the other hand, assuming Apple’s M1 Ultra scales graphics performance very well (the M1 Ultra consists of two M1 Max SoCs connected together), its performance is on par with Nvidia’s GeForce RTX 3070. expected. Integrated graphics processor.
entry level GPU
GPUs | Score | device | APIs | OS |
Apple M1 Max | 8939 | MacBook Pro 18.4 | metal | macOS 13.0 |
GeForce RTX 3050 | 8117 | Desktop PC, Ryzen 5 5600X | Vulcan 1.3 | windows 11 pro |
GeForce RTX 3050 Laptop | 6110 | Mecha Revo Z3 Air Series GM5TGEO, Intel Core i7-11800H | Vulcan 1.3 | windows 10 home |
Intel Arc A380 | 4801 | Desktop PC Core i5-10400 | Direct X 12 | windows 11 pro |
Apple M1 Pro | 4698 | MacBook Pro 14 2021 | metal | macOS 12.6 |
Entry-level GPU owners haven’t submitted, or at least haven’t shown, many results so far. Basically, the more teraflops your GPU has, the better it performs on Sacred Path.
The GA106 in the form of Apple’s M1 Max (iGPU) and Nvidia’s desktop GeForce RTX 3050 dominates here. Meanwhile, Intel’s standalone Arc A380 can barely keep up with Apple’s M1 Pro integrated GPU. Even his Nvidia GeForce RTX 3050 for laptops is significantly faster than the Intel Arc A380.
handheld, ultraportable, integrated
GPUs | Score | device | APIs | OS |
Apple M2 | 3124 | MacBook Air 2022 | metal | macOS 13 |
Apple M1 | 2477 | mac mini 2020 | metal | macOS 12.5.1 |
Apple M1 | 2423 | iPad Pro 11 2021 | metal | iPad OS 15.6.1 |
Apple A16 | 1837 | iPhone 14 Pro | metal | iOS 16.0.3 |
AMD Van Gogh | 1828 | steam deck | Vulcan 1.2 | Steam OS |
AMD Ryzen 9 5900HX | 1744 | Acer Nitro AN515-45 | Vulcan 1.3 | Windows 11 |
Apple A15 | 1595 | iPhone 13 Pro Max | metal | iOS 16.0.2 |
Qualcomm Adreno 730 (Snapdragon 8 Gen 1 — SM8450) | 1365 | Oppo Find X5 Pro | Vulcan 1.1 | Android 12 |
Samsung Xclipse 920 | 1265 | Galaxy S22+ | Vulcan 1.1 | Android 12 |
Arm Mali-G710 MC10 (MediaTek Dimensity 9000) | 899 | Xiaomi Redmi K50 Pro | Vulcan 1.1 | Android 12 |
When it comes to cross-platform benchmarks, handhelds, ultraportables, and iGPUs are arguably the most interesting categories as they include multiple hardware and software architectures and various form factors.
Apple uses the M1 and M2 system-on-chip in their tablets, so it makes sense to include them in this comparison, even though they can power full-fledged PCs like the MacBook Air/MacBook Pro. increase. After all, AMD’s Ryzen SoCs with high-end GPUs are also designed for exactly this purpose. Meanwhile, Apple’s M1 and M2 iGPUs are 40% to 80% faster than AMD’s Cezanne iGPUs.
When it comes to SoCs for handheld devices, Apple’s A16 can even beat out the RDNA 2-based AMD Van Gogh iGPU used in the Steam Deck game console, while the iPhone 14 Pro has something like 2200 points. You have to run the test a few times to get to the stuff. Qualcomm’s Adreno 730, which is used in the Snapdragon 8 Gen 1 (SM8450), performs well, but still lags behind Apple’s previous generation A15 SoC. On the other hand, the RDNA 2-based Samsung Xclipse 920 looks somewhat disappointing when compared to its rivals.
Overview
Basemark’s Sacred Path benchmark can be used to test all kinds of graphics processors, from modest GPUs for smartphones to huge graphics cards that can be used not only for games but also as space heaters and AI supercomputers. .
In most cases, the software produces consistent and reproducible results, which are highly dependent on the computational performance of modern GPUs. While this may be somewhat useful as a performance metric, the 1.15 GB benchmark, which can be completed in seconds on the fastest GPUs, does not adequately show the difference between 1 GB and 24 GB (or more) GPUs. Please be careful. ) VRAM. Raw teraflops beat the game in this test, although the actual game title also depends on graphics memory capacity, memory bandwidth, and other factors.
Also, this benchmark doesn’t use ray tracing to work with all graphics processors from the last 10+ years, so we can’t show all the benefits that leading PC GPUs support. There are also fluctuation-related issues on some platforms, but it’s enough to see the quirks of cross-platform “universal” GPU testing.