IDF: Intel reclaims floating point superiority

Floating point performance is one of the main strengths of AMD's Opteron processor over Intel's chips. The metric represents a workload that is common in applications such as financial analysis or medical modelling.

Pat Gelsinger, general manager for Intel's digital enterprise group, proclaimed that its forthcoming 45nm processor will enable "the fastest machine on the planet".

AMD currently boasts the fastest floating point performance, claiming that its 2.5GHz quad-core 'Barcelona' beats the fastest available Intel chip by 35 per cent.

AMD will release a faster Opteron later this year running at 2.5GHz, and projects a score of 86.3 on the SPECfp_rate2006 benchmark.

Gelsinger revealed that the forthcoming 3.2GHz model of Intel's 45nm Xeon processor will achieve a score of 89.8. Marking a four percent lead over AMD, the Xeon will ship on 12 November.

Intel's projected victory is mostly a morale booster. The average enterprise is rarely confronted with floating point workloads, but AMD's strength in floating point performance suggests that Intel will pull away even further on other benchmarks.

Copyright ©v3.co.uk

floatinghardwareidfintelpoint