IBM has teamed up with the Swiss Federal Institute of Technology (ETH) to build a super efficient water cooled supercomputer which could lower the system’s carbon footprint by a whopping 85 per cent.

The Aquasar system, which boasts two IBM BladeCenter servers in one rack and peak performance of 10 Teraflops, will use chip-level water-cooling as well as reusing any excess heat to warm the university’s buildings.

Each of the supercomputer’s blades will be cooled by a separate liquid cooler on each processor and will also be connected to the whole system via a series of input and output pipelines, or a “fine network of capillaries”. This will mean the water can come extremely close to the processors without damaging anything. It will need just 10 liters of water for cooling pumped at a flow rate of about 30 liters per minute.

Water temperature will be at 60 degrees C – a fair bit under the chip’s maximum 85 degrees C – and the cooling system itself will be a closed circuit. As water is heated by the chips, it will be pushed through a heat exchanger which will transfer the heat into the university heating system. The system is power efficient too, guzzling just 10KW of energy, compared with the average 257KW used by supercomputers in the top 500 list.

Dr. Bruno Michel, a top boffin at IBM’s Zurich Research lab noted “If we capture and transport the waste heat from the active components in a computer system as efficiently as possible, we can reuse it as a resource, thus saving energy and lowering carbon emissions.”

The supercomputer, which should be completed by 2010, will apparently be used for research in nanotechnology and fluid dynamics.

Incoming search terms:

• OPTERON server water cooling system
• Alphacool X-Box 360 Slim Cooler
• ban water cooling
• chilled liquid supermicro
• macbook liquid cooling
• xw9400 liquid cooling
• 23 inch rack server liquid cooling system
• Gaming PC FX-8120 8 CORE 16GB SSD HD6970 Liquid Cooled

Some pretty interesting news comes from xbit-labs, as word is that one of the most important features of DX11 will be avaiable for DX 10 and DX10.1 class graphics cards as well, compute shaders.

General-purpose processing on graphics processing units (GPGPU) is gaining popularity slowly, but surely. Unfortunately, at present there is only one standard application programming interface – OpenCL – that supports all graphics chips available today. But apparently, DirectX 11, which will enable GPGPU, physics, artificial intelligence, etc. programming via compute shaders, will also be able to take advantage of present GPUs, albeit with some limitations.

In order to popularize compute shaders (CS) among developers, the DirectX 11 includes not only compute shaders 5.0, but also compute shaders 4.0 (for Direct X10 hardware) and 4.1 (for DirectX 10.1 hardware), which are not supported by DirectX 10. Compute shaders 4.0/4.x have a number of limitations compared to version 5.0, including maximum number of threads per group (768), thread group shared memory (16KB vs. 32KB in CS 5.0), absence of atomic operations or append/consume and so on. CS 5.0 will also offer better interaction with graphics pipeline (e.g., it can output to textures), double precision and so on.

The main aim of compute shaders 4.x is to allow game developers to practice with compute shaders technology, enable GPGPU via DirectX as well as let game developers to use CS for complex rendering-related tasks instead of pixel shaders so to gain performance.

Beta version of DirectX 11 as well as drivers that support compute shaders 4.x are available from Microsoft, ATI/AMD as well as Nvidia.