Chipmakers admit: Your power may vary

TalkBack 1 of 11:

Next »

• Thread View
• Flat View

Flops/Watt and IntOps/Watt ?

`` Parsing the differences
But TDP is somewhat unrealistic, in that servers and PCs aren't usually running anywhere near full bore for extended periods of time. "This is like going around your house and counting light bulbs to determine the monthly power bill," Skaugen said.''

So what if system A and B, having identical amounts of DRAM, sufficient so that NO PAGING is necessary, execute exactly the SAME set of benchmarks.

Then measure and compare the TOTAL power consumption for them EACH to solve the SAME problems.

If problem X requires T_yx total time and F_yx and I_yx FPOps and IntOps, respectively, and WT_yx total Watts for system y, then for each system compute

Flops_y = F_yx / T_yx;
Ints_y = I_yx / T_yx;

FPops_per_Watt_y = F_yx/WT_yx
IntOpes_per_Watt_y = I_yx/WT_yx


Then you can compare average performance (in say gigaflop/sec) and average power (in say gigaflop/Watt) per system.

Posted by: michael_t   Posted on: 06/09/06 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

SponsoredWhite Papers, Webcasts, and Downloads

• Using Red Hat Enterprise Linux AS to Achieve Highly Available, Load-Balanced Clusters Dell The Red Hat Enterprise Linux AS operating system integrates Cluster ... Download Now
• The True Costs of Virtual Server Solutions VMware In an economic environment that is repeatedly heralding the message "do ... Download Now
• VMware Infrastructure: A Guide to Bottom-Line Benefits VMware Frustrated by the costs of maintain ever larger data centers?or building ... Download Now