|
Hi Jeff- The secretive ways of TI's doc center! Dont you think this piece of info is a prime candidate for addition as foot note to the IDLE instructions's description in the CPU and instn set reference guide? And my reference to of "branch to itlsef" being like IDLE is apparently flat out wrong! After all since both IDLE and NOP *dont get executed* by any functional unit and IDLE just being a version multicycle NOP's,it does seem like it will be definitely less power consuming. (But still IDLE is no *stop*!) And also since IDLE is just a function of the NOP being accessed from memory as you rightly point out, IDLE could be the best choice. So, yeah I guess it makes sense to me now (duh!). Thanks for pointing to the doc. --Bhooshan On Sun, 20 Feb 2005 00:18:00 -0600, Jeff Brower <> wrote: > Hi Bhooshan- > > > Isnt IDLE like an infinite "branch to itself", if so does it really > > help save power? Does TI mention anywhere about power savings due to > > IDLE? > > It does NOPs, which are supposed to save power. Did you know there is a "TMS320C6000 > DSP Power-Down Logic and Modes Reference Guide" document? I did not, but there is, > seems like it came out in Aug 2004 and was revised Jan 2005: > > http://focus.ti.com/lit/ug/spru728b/spru728b.pdf > > On pg 9: > > "In addition to power-down modes described in this document, > the IDLE instruction provides lower CPU power consumption > by executing continuous NOPs. The IDLE instruction terminates > only upon servicing an interrupt." > > How much power is saved by hard-coded NOPs vs. a code loop that does the same thing? > That's not quantified, but I suppose a bit might be saved by not executing NOPs from > P cache, which is a piece of SRAM. > > -Jeff > > On Sat, 19 Feb 2005 10:41:41 -0600, Jeff Brower <> wrote: > > > > > > Mike- > > > > > > > 1. there is no functional reason to halt [or try to halt] a c6000 DSP. > > > > > > Is reducing power consumption a valid reason to make frequent use of IDLE/interrupt > > > combination? > > > . > > > . > > > . > > > > > > > Of course, if we all add bullet proof hardware and software, we would > > > > not be thinking about reset after power up. > > > > > > There is always the cosmic ray with the name of your product on it :-) I know that's > > > truly nanoscopic writing not yet developed by mankind, but it's there. > > > > > > -Jeff > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > ------------------------------- > > "I've missed more than 9000 shots in my career. > > I've lost almost 300 games. 26 times I've been trusted to take the > > game winning shot and missed. > > I've failed over and over again in my life. > > And that is why I succeed." > > -- Michael Jordan > > -------------------------------- > -- ------------------------------- "I've missed more than 9000 shots in my career. I've lost almost 300 games. 26 times I've been trusted to take the game winning shot and missed. I've failed over and over again in my life. And that is why I succeed." -- Michael Jordan -------------------------------- |
Re: Re: Halting the DSP
Started by ●February 20, 2005
Reply by ●February 20, 20052005-02-20
|
If the real reason you want to halt the DSP, is to save power, then a better way to do it would be to use the power down modes. Regds JS -----Original Message----- From: Jeff Brower [mailto:] Sent: Sunday, February 20, 2005 12:18 AM To: Bhooshan Iyer Cc: Subject: Re: [c6x] Re: Halting the DSP Hi Bhooshan- > Isnt IDLE like an infinite "branch to itself", if so does it really > help save power? Does TI mention anywhere about power savings due to > IDLE? It does NOPs, which are supposed to save power. Did you know there is a "TMS320C6000 DSP Power-Down Logic and Modes Reference Guide" document? I did not, but there is, seems like it came out in Aug 2004 and was revised Jan 2005: http://focus.ti.com/lit/ug/spru728b/spru728b.pdf On pg 9: "In addition to power-down modes described in this document, the IDLE instruction provides lower CPU power consumption by executing continuous NOPs. The IDLE instruction terminates only upon servicing an interrupt." How much power is saved by hard-coded NOPs vs. a code loop that does the same thing? That's not quantified, but I suppose a bit might be saved by not executing NOPs from P cache, which is a piece of SRAM. -Jeff > On Sat, 19 Feb 2005 10:41:41 -0600, Jeff Brower <> wrote: > > > > Mike- > > > > > 1. there is no functional reason to halt [or try to halt] a c6000 DSP. > > > > Is reducing power consumption a valid reason to make frequent use of IDLE/interrupt > > combination? > > . > > . > > . > > > > > Of course, if we all add bullet proof hardware and software, we would > > > not be thinking about reset after power up. > > > > There is always the cosmic ray with the name of your product on it :-) I know that's > > truly nanoscopic writing not yet developed by mankind, but it's there. > > > > -Jeff > > > > > > > > > > > > > > > > > > -- > ------------------------------- > "I've missed more than 9000 shots in my career. > I've lost almost 300 games. 26 times I've been trusted to take the > game winning shot and missed. > I've failed over and over again in my life. > And that is why I succeed." > -- Michael Jordan > -------------------------------- |
