tag:blogger.com,1999:blog-664175667937540078.post279964352680792269..comments2024-03-28T21:26:11.788+01:00Comments on Tomeu Vizoso: Dynamic scaling of the memory busTomeu Vizosohttp://www.blogger.com/profile/16626407169435386757noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-664175667937540078.post-64104334334107911242014-08-14T08:40:59.706+02:002014-08-14T08:40:59.706+02:00Yeah, but that's true of any other resource in...Yeah, but that's true of any other resource in the critical path (including the CPU itself).<br /><br />I think the API proposed is sufficiently low-level to be useful no matter what concrete strategy is followed, which is something that I hope will be tuned with the help of the vendor's QA teams.Tomeu Vizosohttps://www.blogger.com/profile/16626407169435386757noreply@blogger.comtag:blogger.com,1999:blog-664175667937540078.post-75870631470595448542014-08-13T22:36:10.241+02:002014-08-13T22:36:10.241+02:00Fun project!
However, beware, slowing down the me...Fun project!<br /><br />However, beware, slowing down the memory bus could actually force the CPU to be active for a longer period of time and result in a greater power usage than if memory requests were answered at full speed before drastically downclocking memory.<br /><br />That is something that performance counters may not be able to capture if polled at a period slower than these burst. Good thing SoCs tend to have PMUs nowadays!<br /><br />Thanks for working on this project :)mupufhttps://www.blogger.com/profile/17157987837111564920noreply@blogger.com