GPU Performance: Nvidia Tegra 2
As we’ve mentioned in the past, mobile devices like smartphones and tablets use what’s known as a system-on-chip (SoC). This integrates the processor, GPU, RAM, and several other subsystems onto single device. Since all of those components sit next to each other on the same chip, there is greater efficiency in data transfers, while reducing the amount of space consumed on the PCB.
| Header Cell - Column 0 | Apple A4 (iPad) | Apple A5 (iPad 2) | Tegra 2 (Xoom/Iconia A500/Eee Pad Transformer) |
|---|---|---|---|
| Processor | 1 GHz ARM Cortex-A8 (single-core) | 1 GHz ARM Cortex-A9 (dual-core) | 1 GHz ARM Cortex-A9 (dual-core) |
| Memory | 256 MB 333 MHz LP-DDR (single-channel) | 512 MB 1066 MHz LP-DDR2 (dual-channel) | 1 GB 667 MHz LP-DDR2 (single-channel) |
| Graphics | PowerVR SGX535 (single-core) | PowerVR SGX545MP2 (dual-core) | ULP GeForce (single-core) |
| L1 Cache(Instruction/Data) | 32 KB / 32 KB | 32 KB / 32 KB | 32 KB / 32 KB |
| L2 Cache | 640 KB | 1 MB | 1 MB |
Tegra is Nvidia’s SoC brand, and it represents the company’s effort to tap into the mobile market beyond its desktop-derived GeForce graphics processors. For those unfamiliar with Tegra 2, read page eight of our Motorola Xoom review for a full discussion on GPU architecture. On the CPU side, Tegra 2 shares the same dual-core Cortex-A9, which is a substantial boost from Cortex-A8 seen in the first-generation iPad. Check out Apple's iPad 2 Review: Tom's Goes Down The Tablet Rabbit Hole for a full discussion of Cortex-A9 performance.
While we've already covered Tegra 2, it's important to examine the graphics performance of each tablet. Sy Choudhury, director of product management at Qualcomm, states, "There is a misconception that the same processor and operating system gives the same performance." We run into that issue every time we review a tablet. Why is it wrong to expect the same performance from all Honeycomb-based Tegra 2 tablets? Hardware vendors get access to the same optimizations from the SoC developer, but not all of the software-based tweaks are enabled by OEMs.
| GPU (System-on-Chip) | PowerVR SGX 535(Apple A4) | PowerVR SGX 543(Apple A5) | ULP GeForce (Tegra 2) |
|---|---|---|---|
| SIMD | USSE | USSE2 | Core |
| Pipelines | 2 (unified) | 4 (unified) | 8 (4 pixel / 4 vertex) |
| TMUs | 2 | 2 | 2 |
| Bus Width (bit) | 64 | 64 | 32 |
| Triangle rate @ 200 MHz | 14 MTriangles/s | 35 MTriangles/s | ? |
The ULP GeForce has a maximum operating frequency of 300 MHz, but device vendors can tweak this setting to save on power. Nvidia provides less information on the Tegra 2 than it does for its desktop GPUs, so it’s best to move on to benchmarks. As in our iPad 2 review, we're turning to GLBenchmark 2.0.
In terms of frames rendered in a set period of time, the Xoom, Iconia A500, and Transformer offer more performance than the original iPad. But all three Tegra 2-based tablets still fall short of the iPad 2. Interestingly, the Transformer performs worse than the A500 with Honeycomb 3.1. Once we update to 3.2, Asus' tablet jumps ahead by more than 20%.
There is a direct relationship between memory bandwidth and triangle rates in the A4 and A5's PowerVR GPUs, due to their tile-based deferred rendering architecture. But PowerVR GPUs operate differently than what we're used to seeing on the desktop. Tegra 2, on the other hand, employs a more traditional z-buffered rendering architecture. As such, it's only possible to compare the triangle rates of competing Tegra 2-based tablets.
If you want to go into more depth than what is provided here, the full results of GLBenchmark 2.0 have been posted on Google Docs.
