Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Jim McMannis who wrote (114504)	10/19/2000 3:16:31 AM
From: fp_scientist	Respond to of 186894

Jim, Re: SPEC scores are mainly relevant to server salesmen aren't they? Or who cares? SPEC scores are extensively used by people who do intensive floating point (fp) to decide which particular platform to buy. Granted that intensive fp users care mostly about 64 bit, Alpha, RISC, etc, machines. They are also a rather small bunch compared to the consumer or commercial market of PCs. The SPECfp benchmark was developed by a consortium with the intention of avoiding the manipulation that computer vendors were doing with Linpack, which was used as a benchmark before. There are many stories about FORTRAN compilers that could recognize specific portions of Linpack code and produce highly optimized executable. SPECfp is a combination of several codes in the public domain (from several fields like molecular dynamics, quantum Monte Carlo, particle physics, etc) and is pretty balanced between things that are better executed in serial, parallel and/or vector machines. It's mostly FORTRAN77 and 90 but there is some C too. No benchmark is perfect, but SPECfp is much better than anything else out there. It contains a good mixture of typical things that fp intensive users do. X86 machines are making huge inroads in the intensive fp community because of their price/performance ratio (not just performance!). The 1.5 GHz P4 SPECfp number of 524, if correct, is very good. The Tbird 1.1GHz scores 311 and 331 for base and peak, respectively. If you scale these numbers by 1.5/1.1 (ratio of clock speeds), you get 424 and 451. The P4 score of 524 would be 16% faster than the peak value for the Tbird at similar clock frequency. You have to give Intel some credit on this one. For other SPECfp benchmarks, check spec.org Regards, fp