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: fingolfen who wrote (142387)	8/29/2001 4:18:50 PM
From: pgerassi	Read Replies (1) \| Respond to of 186894

Wanna_bmw: Look at those test that use real world data, Anand's database server test. The data was pulled from the transaction logs for a period of actual server use. IE, how fast does a system perform the various transactions in the order received from a real period in a real environment? The Scimark routines use the best currently available algorithms, libraries and working compilers optimized for each processor type. These fit in your purported "good benchmarks rule". Yet, you do not like them because P4 does more poorly than Athlon at them. The Quake demos and most of the game demos just do a script that does not reflect how a typical hot shot plays on a real game. AI is also shut off as only the results are captured and not the inputs to the game. This over estimates the performance of the game especially for highly pipelined CPUs as the AI code has very many unpredictable branches. Descent 3 has a demo mode that scripts the user input and different things happen for each run. AI is not turned off. These come closer to using real world game benchmarking and the results comes closer to actual performance at the cost of repeatability. Game demos are good for benchmarking video performance of the graphics cards but, not CPU or system performance. BAPCO and the other synthetic office, multimedia and internet benchmarks all purport to simulate real world usage. The mere fact that small rotates are not used in the Photoshop tests but some very esoteric filters are, shows me that they not even close to truly real world usage. Check Dean's comments about sort testing on IBM mainframes at RWT. Real world tests are done by customers when millions of dollars are at stake. Not the synthetic ones. The real world tests show how the system does for the things they do and the way they do them. Pete