There's been a lot of articles all over the net in the last 6 months about the power that the new Cell Processor (from the PS3) offers and the potential that have when linked in paralle clusters. I read one particular article where several MIT students linked 16 PS3's to form their own Supercomputer because they couldn't get enough time on the campus system. So surely the time when we'll all have supercomputers in our homes isn't far away?
Well, with NVidia's new “Tesla personal supercomputer” you can now have the serious digital horsepowerof a supercomputer at home. The beast delivers cluster computing performance and up to 250 times faster computing then a standard top of the range PC. The Tesla is powered by 4 GPU’s, each with 240 processing cores (that 4x240=960 cores working in parallel) based on the NVidia’s CUDA architecture, delivering a gigantic maximum performance of nearly 4 Teraflops!
Here are some more of its specs:
- Massively-parallel many-core architecture
- 3 or 4 Tesla C1060 Computing Processors with 4GB of dedicated memory per GPU
- 2.33 GHz+ Quad-core AMD Phenom or Opteron, — OR — Quad-core Intel Core 2 or Xeon
- Minimum system memory: 12 GB for 3 Tesla C1060s and 16 GB for 4 Tesla C1060s (at least 4GB per Tesla C1060)
- 12GB+ system memory (at least 4GB per Tesla C1060)
- 240 scalar processor cores per GPU
- Integer, single-precision and double-precision floating point operations
- Hardware Thread Execution Manager enables thousands of concurrent threads per GPU
- Parallel shared memory enables processor cores to collaborate on shared information at local cache performance
- Ultra-fast GPU memory access with 102 GB/s peak bandwidth per GPU
- IEEE 754 single-precision and double-precision floating point
- Each Tesla C1060 GPU delivers 933 GFlops Single Precision and 78 GFlops Double Precision performance
The only bad news is that you'll need to spend about $10,000 to pick up one of these monsters. If I had the money I so would!