The most important news on technology, developments and trends with insightful analysis. Coverage includes hardware, software, networking, wireless computing, personal technology, security and cutting-edge technology from the developer world to the consumer world.
A ton of great new SoC developments have been introduced as of late. Qualcomm’s S4 is an exciting chipset yet to be publicly available in a product, NVIDIA’s Tegra 3 with its 4+1 quad-core chipset and 12 core GPU has been established and is set to make a big splash this year, and Apple claims their A5X inside the new iPad is way better than that (yet to be proven). But Huawei could have something even more powerful than all of those.
Sweden-based website Nordic Hardware uncovered a bit more information about the chipset inside the Huawei Ascend D Quad XL. It’s an in-house processor called the Huawei K3V2 and is a quad core processor which comes in either 1.2GHz or 1.5GHz (depending on configuration) and has a 16 core GPU, 4 more than what NVIDIA’s GeForce has inside Tegra 3. They figured they’d put the device to the test to see if it really was that much faster than all the rest.
One benchmark they used was from Powerboard.com and it tested megapixels per second in each device’s native resolution. The D Quad outscored Samsung’s Exynos, NVIDIA’s Tegra 3 and Qualcomm’s MSM8260 overclocked to 2GHz. The device houses an HD display so the results aren’t as skewered as it would on, say, a WVGA or qHD device. We’re not sure if screen size differences are a factor in determining results.
They moved on to the ever familiar Nenamark 2 benchmark. This benchmark tests OpenGL ES 2.0 framerate. The Ascend D Quad XL beat out everyone with a frame rate of 62.75. The next closest was Qualcomm’s S4 with Adreno 225, NVIDIA’s GeForce with ULP2 and all the other dual-core options you’d expect it to be. (The S4 is the only dual-core chipset that could hang with it.) Four of the five top results are devices with at least 720p displays, with the Tegra 3-powered ASUS Transformer Prime having a slightly higher resolution (which could skewer results, but not by much).
Benchmarks are never an end-all, be-all method of determining which chipsets are faster and better than one another. Some benchmarks may put more emphasis on certain features of a chipset than others, and some may just be bad benchmarks period.
Another thing to consider is that benchmarks aim to push devices to their limit to see how much they’re capable of, but users often won’t notice differences in the real world. Games are still being developed for older 1GHz single-core processors so you won’t see a huge difference from platform to platform unless games are optimized like they are for Tegra 3. Also, single-core processors utilize Android quite well, so OS speed gains are usually only marginal at best. What do you guys think? [Thanks Anton!]