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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Petz who wrote (17556)	11/4/2000 4:10:11 PM
From: fyodor_	Read Replies (1) \| Respond to of 275872

<Petz: in my opinion they should allow use of assembly language in the SPEC benchmarks. I've seen it alleged that some of the "optimizing compilers" actually recognize when they are compiliing SPEC code and substitute hand coded assembly language.> I would really like to see two numbers: PEAK and BASE. Not with SPEC's current rules, but rather with much looser restrictions for PEAK scores (i.e. hand-coded assembly or whatever), but with much tighter restrictions for the BASE scores. It's always interesting to see what an architecture is capable of if everything is optimized, but it's equally interesting to see what can be achieved using _normal_ compilers. I would actually like to see the BASE scores reported with "the most commonly used compiler" for that platform. This would, imho, have two distinct advantages. Firstly, the base scores would reflect actual performance to a much closer degree. Secondly, it would encourage the CPU makers to help/force/convince the compiler maker to optimize their compilers to a much larger degree. Alternatively, the CPU maker could push their own compilers much more aggressively (while making them able to actually compile more code correctly). Either way, end users would benefit from better compilers and more relevant benchmarks. -fyo