Why? Most devices cannot play video recorded by a camcorder. Such media would be most likely be edited first and then transcoded (possibly transparently) when copied to a device for playback. With current high-definition media such an operation can take a long time, thus any performance improvements can save a lot of time.

Using the new Media Foundation in Windows 7, we measure WMV (e.g. Movie Maker) > MP4/H264 and MP4/H264 (e.g. phone video) > MP4/H264 using various profiles (HD TV 720p, SD TV 480p, tablet, phone, etc.).

The benchmark can compare the performance of hardware accelerated media transcoders (e.g. GPUs) with multi-threaded software (CPU) transcoders.

New Benchmark: GPGPU Cryptography

Measures the encryption, decryption and hashing bandwidth of GPGPUs

Why? While modern CPUs include special instruction sets (Intel AES, VIA PadLock) to accelerate cryptographic operations, are they a match for modern GPGPUs? The benchmark uses the same workload as the CPU Cryptography benchmark allowing direct comparison of modern CPUs and GPGPUs of the most common algorithms in use today (AES256, AES128, SHA256, SHA1).

OpenCL 1.x, DirectX 11 Compute Shader and CUDA3 3.x are supported. Multiple video cards are supported and CPUs can also be used in parallel in OpenCL. The benchmark allocates workload based on the respective processing power of GPGPUs and CPUs.

New Benchmark: Blu-ray

Measures the performance of Blu-ray optical drives

Why? While Blu-ray has won the high-definition war some time ago, it has finally become affordable for computer users as a replacement for DVD. We have thus added a brand new Blu-ray Test Disk to measure the performance of the latest drives on the market, adding to the existing DVD and CD Test Disks which are still required to measure DVD and CD disk performance.

AVX/FMA instruction set support for future CPUs Processor Multi-Media, Processor Cryptography, Memory Bandwidth, Cache and Memory Benchmarks

Using 256-bit register width (instead of 128-bit of SSE/2/3/4) yields further performance gains through greater parallelism in most algorithms. Combined with the increase in processor cores and threads we will soon have CPUs rivaling GPGPUs in performance.

Measure each operation report min-max score as well as average

Why? The maximum time an operation can take is just as important as the average time; if the measured performance varies greatly from test to test, the response time is not consistent and thus the performance unreliable. If a minimum performance must be guaranteed, then a slower but consistent device would be preferable to a faster but inconsistent one.

Using the submitted scores to the Ranker, we further measure the consistency of the performance of a device; we can also verify whether the measured performance falls between the range of performance expected for the tested device.

Geometric mean for aggregate scores

Why? When combining components of very different magnitudes, a geometric mean is more meaningful than an arithmetic one. This makes comparing the relative performance of very different devices much fairer, especially between platforms (e.g. desktop to mobile or to server).