首页 | 最新需求 | 最新现货 | IC库存 | 供应商 | IC英文资料库 | IC中文资料库 | IC价格 | 电路图 | 应用资料 | 技术资料
 IC型号:
您现在的位置:首页 >  IC英文资料库 进入手机版 
 
资料编号:148742
 
资料名称:TMS320DM642AZDK5
 
文件大小: 1669.24K
   
说明
 
介绍:
Video/Imaging Fixed-Point Digital Signal Processor
 
 


: 点此下载
  浏览型号TMS320DM642AZDK5的Datasheet PDF文件第3页
3
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第4页
4
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第5页
5
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第6页
6

7
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第8页
8
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第9页
9
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第10页
10
浏览型号TMS320DM642AZDK5的Datasheet PDF文件第11页
11
 
本平台电子爱好着纯手工中文简译:截至2020/5/17日,支持英文词汇500个
www.德州仪器.com
2.2cpu(dspcore)描述
TMS320DM642
video/imagingfixed-pointdigitalsignalprocessor
SPRS200J–JULY2002–REVISEDAUGUST2005
thecpufetchesvelociti™advancedvery-longinstructionwords(vliws)(256bitswide)tosupplyupto
eight32-bitinstructionstotheeightfunctionalunitsduringeveryclockcycle.thevelociti™vliw
architecturefeaturescontrolsbywhichalleightunitsdonothavetobesuppliedwithinstructionsifthey
arenotreadytoexecute.thefirstbitofevery32-bitinstructiondeterminesifthenextinstructionbelongs
tothesameexecutepacketasthepreviousinstruction,orwhetheritshouldbeexecutedinthefollowing
clockasapartofthenextexecutepacket.fetchpacketsarealways256bitswide;不管怎样,theexecute
packetscanvaryinsize.thevariable-lengthexecutepacketsareakeymemory-savingfeature,
distinguishingthec64xcpusfromothervliwarchitectures.thec64x™velociti.2™extensionsadd
enhancementstothetms320c62x™dspvelociti™architecture.theseenhancementsinclude:
Registerfileenhancements
Datapathextensions
quad8-bitanddual16-bitextensionswithdataflowenhancements
Additionalfunctionalunithardware
Increasedorthogonalityoftheinstructionset
Additionalinstructionsthatreducecodesizeandincreaseregisterflexibility
thecpufeaturestwosetsoffunctionalunits.eachsetcontainsfourunitsandaregisterfile.oneset
containsfunctionalunits.l1,.s1,.m1,和.d1;theothersetcontainsunits.d2,.m2,.s2,and.l2.the
tworegisterfileseachcontain3232-bitregistersforatotalof64general-purposeregisters.inadditionto
supportingthepacked16-bitand32-/40-bitfixed-pointdatatypesfoundinthec62x™velociti™vliw
architecture,thec64x™registerfilesalsosupportpacked8-bitdataand64-bitfixed-pointdatatypes.这
twosetsoffunctionalunits,alongwithtworegisterfiles,composesidesaandbofthecpu[seethe
functionalblockandcpu(dspcore)图解,和figure2-1].thefourfunctionalunitsoneachsideof
thecpucanfreelysharethe32registersbelongingtothatside.additionally,eachsidefeaturesa"数据
crosspath"—asingledatabusconnectedtoalltheregistersontheotherside,bywhichthetwosetsof
functionalunitscanaccessdatafromtheregisterfilesontheoppositeside.thec64xcpupipelines
数据-交叉-pathaccessesovermultipleclockcycles.thisallowsthesameregistertobeusedasa
数据-交叉-pathoperandbymultiplefunctionalunitsinthesameexecutepacket.allfunctionalunitsinthe
c64xcpucanaccessoperandsviathedatacrosspath.registeraccessbyfunctionalunitsonthesame
sideofthecpuastheregisterfilecanservicealltheunitsinasingleclockcycle.onthec64xcpu,一个
delayclockisintroducedwheneveraninstructionattemptstoreadaregisterviaadatacrosspathifthat
registerwasupdatedinthepreviousclockcycle.
inadditiontothec62x™dspfixed-pointinstructions,thec64x™dspincludesacomprehensive
collectionofquad8-bitanddual16-bitinstructionsetextensions.thesevelociti.2™extensionsallowthe
C64xCPUtooperatedirectlyonpackeddatatostreamlinedataflowandincreaseinstructionset
效率.thisisakeyfactorforvideoandimagingapplications.
anotherkeyfeatureofthec64xcpuistheload/storearchitecture,whereallinstructionsoperateon
寄存器(asopposedtodatainmemory).twosetsofdata-addressingunits(.d1and.d2)是
responsibleforalldatatransfersbetweentheregisterfilesandthememory.thedataaddressdrivenby
这.dunitsallowsdataaddressesgeneratedfromoneregisterfiletobeusedtoloadorstoredatatoor
fromtheotherregisterfile.thec64x.dunitscanloadandstorebytes(8bits),half-words(16bits),和
words(32bits)withasingleinstruction.andwiththenewdatapathextensions,thec64x.dunitcanload
andstoredoublewords(64bits)withasingleinstruction.此外,thenon-alignedloadandstore
instructionsallowthe.dunitstoaccesswordsanddoublewordsonanybyteboundary.thec64xcpu
supportsavarietyofindirectaddressingmodesusingeitherlinear-orcircular-addressingwith5-or15-位
补偿.allinstructionsareconditional,andmostcanaccessanyoneofthe64registers.someregisters,
不管怎样,aresingledouttosupportspecificaddressingmodesortoholdtheconditionforconditional
说明(iftheconditionisnotautomatically"真实").
DeviceOverview
7
资料评论区:
点击回复标题作者最后回复时间

标 题:
内 容:
用户名:
手机号:    (*未登录用户需填写手机号,手机号不公开,可用于网站积分.)
      
关于我们 | 联系我们
电    话13410210660             QQ : 84325569   点击这里与集成电路资料查询网联系
联系方式: E-mail:CaiZH01@163.com