Sandra 2011 SP4 (17.70) Released

[Apoptosis]

Just got an e-mail from the guys at Sandra and it looks like a new patch is out! Check out the e-mail below for all the details.

London, UK, 13 July 2011 - We have released SP4 (17.70) that includes a brand-new computer-wide benchmark (Overall Score) to rate the overall performance of the computer.

It includes complete APU benchmarking (CPU+GPU) through the existing GP (General Purpose) benchmarks (GP Arithmetic, GP Cryptography, GP Memory Bandwidth) for the newly released "Fusion"/"Llano" AMD platforms.

Sandra can benchmark APU units (CPU+GPU) simultaneously by harnessing the flexibility and power of OpenCL and thus show the performance improvements APUs bring compared to legacy CPU and GPU-only solutions.

We are launching SP4 for Sandra 2011 with various enhancements based on telemetry as well as optimisations based on further tests on RTM hardware.


* General-Purpose (GP) Computing benchmarks - CPU vs. GP(GPU) vs. APU
---------------------------------------------------------------------

Sandra's GP (former GPGPU) benchmarks may still be the only ones that allow full APU performance measurement against CPUs or even GP(GPU)s (through OpenCL) as they use the *same workload* as the native as well as the software VM (.Net/Java) counterparts allowing apples-to-apples comparisons:

- GP Performance (OpenCL / DX CS / CUDA) = CPU Multi-Media / .Net Multi-Media / Java Multi-Media / Video Shading
- GP Cryptography (OpenCL / DX CS / CUDA) = CPU Cryptography / .Net Cryptography / Java Cryptography
- GP Bandwidth (OpenCL / DX CS / CUDA) = Memory Bandwidth / Video Bandwidth

As a user, you would not care if a program uses native CPU instructions, the GP(GPU) or even your APU (CPU+GPU) to get your work done faster.

The point is you are benchmarking CPU+GPU together and not just CPU or GPU individually; both resources (thus the whole APU) is used to perform computations better which is the whole point.

As we support OpenCL, DirectX ComputeShader and CUDA (even STREAM for legacy devices) just about all but the simpletest GPUs are supported.

As long as the work-load is the same (which it is for Sandra) then the comparison is valid.


* Transcoding benchmark - CPU vs. GP(GPU) vs. APU
-------------------------------------------------

The key advantage of Sandra's benchmark is WMF (Windows Media Foundation): it can use either software (CPU) transcoding, GP(GPU) or APU transcoding depending on the encoders/decoders installed. So you can benchmark CPU vs. GP(GPU) or APU using the *same workload*.

Other benchmarks may use only software decoders/encoders which means you only test CPU performance and ignore GP(GPU) or APU performance entirely. Only by using the hardware accelerated decoders/encoders you can harness the power of GP(GPU) and APU.


* System Overall benchmark - GP(GPU/APU) support
------------------------------------------------

We have decided to use the GP benchmarks instead of the Video/GPU benchmarks where GP(GPU/APU) support is available with fall-back to Video/GPU benchmarks where GP support is not available.

This allows the score to better reflect the power of systems with multiple GPUs (that may not be fully used by SLI/Crossfire) and APUs while maintaining compatibility with previous scores.

The score is now an average of the current benchmarks:

- Native CPU performance: CPU Arithmetic, Multimedia
- Software VM performance: .Net or Java Arithmetic, Multimedia
- Native Memory & Cache performance
- Storage performance
- GP performance: Arithmetic, Memory


Complete systems are now ranked not just their individual components:

http://www.sisoftware.co.uk/rank2011d/t ... d4f281bc8c


* Power Management states for both CPU & GPU
--------------------------------------------

For each power management level supported (i.e. frequency) Sandra displays as much information as there is available, e.g. voltage, current and naturally power. The amount of information naturally depends on the actual hardware tested.


-----------------------------------------------------------------------------