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 (68737)	1/24/2002 2:06:34 AM
From: Joe NYC	Respond to of 275872

John, I think this report also gives proof that Sysmark 2001 is totally unreliable in measuring system performance. it showed a 92% speedup with a 36% increase in CPU clock. Good catch. Either that or there is something fishy with their testing procedure. Another possibility is that DDR helps a lot in this test, since the other tests are with RDRAM. Unfortunately, this article with it's water cooling to 0 degC also says nothing about the overclockability of the "typical" NW 2200. I totally discount the report of the super-special-edition NW 2200 that could overclock to 3000 on air cooling, since no specific benchmarks are assigned to this CPU. Nobody other than the infamous 3 know what happened during that all nighter. We can only speculate. My guess is that the test was supposed to be that of handpicked NW running at 3 GHz with RDRAM (like all the other ones), air cooling. It didn't quite work out, so they had resort switching to DDR motherboard first, then watercooling this handpicked CPU, and finally adding ice in the water, in order to make it through all the benchmarks. Joe

To: Petz who wrote (68737)	1/24/2002 1:18:08 PM
From: Ali Chen	Read Replies (4) \| Respond to of 275872

"I think this report also gives proof that Sysmark 2001 is totally unreliable in measuring system performance. it showed a 92% speedup with a 36% increase in CPU clock." The Sysmark does what it is supposed to do - to measure the _system_ performance. The result simply says that the performance of peripherals and memory is more important than the raw CPU speed on the Sysmark-selected set of applications. "This is, of course, physically impossible unless the measurement is faulty." If in your workload the CPU mostly waits for data from memory or for completions of disk transfers (or finishing polygons drawings), and you increase the overall system clock by 2x, your performance will also jump by about 2x regardless of CPU speed. Regards, - Ali