Mene ne Kama data da caching?

Kwamfuta, alas, ba nan da nan executes da dokokin, wanda ake samu daga cikin mutane. Don hanzarta wannan tsari yana amfani da wani iri-iri na dabaru, da kuma girman kai na wurin, daga gare su nasa ne caching. Mene ne wannan? Mene ne Kama data? Ta yaya wannan tsari zahiri faruwa? Abin da aka adana bayanai a wani smartphone "Samsung", misali, da kuma sun bambanta da ɗan daga waɗanda suke a cikin kwamfuta? Bari mu samu saukar zuwa samun amsoshin wadannan tambayoyi.

Mene ne cache?

Don haka ya kira matsakaici buffer, wanda ya samar da dama mai sauri zuwa bayanai, yiwuwar cewa query a sama da dukan. All data dauke a cikin ta. An muhimmin amfani ne cewa cire dole bayanai daga cache iya zama da muhimmanci da sauri fiye da daga asali store. Amma akwai wani gagarumin drawback - size. Kama data aka yi amfani da bincike, da wuya faifai, CPU, da sabar yanar gizo, ya lashe, kuma DNS sabis. Dalili na tsarin ne recordsets. Kowace daga cikinsu ake dangantawa da wani kashi ko data naúrar cewa protrude kwafa abin da yake a cikin babban memory. Entries da wani ganowa (sawa), ta hanyar da aka ƙaddara da matching. Bari mu dubi kadan daban-daban ra'ayi: abin da aka adana bayanai a wayarka "Samsung" ko wani manufacturer? Shin, sũ ne daban-daban daga wadanda suke a cikin kwamfuta? Daga wani muhimman hakkokin ra'ayi - ba kawai bambanci a size.

A tsari na yin amfani da

Lokacin da wani abokin ciniki (su kana aka jera a sama) buƙatun data, abu na farko da cewa sa a kwamfuta - nazarin cache. Idan shi ne ake so shigarwa, to, shi ne amfani. A cikin wadannan lokuta akwai wani hit. Lokaci-lokaci da data kofe daga cikin cache zuwa babban memory. Amma idan bukata rikodin ba a samu, akwai wani search abun ciki a cikin tushe ajiya. All dauki da bayanin da aka canjawa wuri zuwa ga cache, saboda haka nan za a iya isa ga mafi sauri. Yawan buƙatun lokacin lashe tare da nasara, da ake kira matakin, ko buga rabo.

Ana ingantawa data

Lokacin amfani, misali, a yanar gizo browser da aka bari na gida cache a sami wani kwafi na page. Ganin gazawar wannan memory type, wani kuskure da aka yanke shawarar zuwa jefar da bayanai yin sarari. Don shirya daidai da abin da za a iya maye gurbinsu, yi amfani da daban-daban Algorithms hijira. Af, idan mun yi magana game da gaskiyar cewa wannan Kama data ga "Android", da cewa ga mafi part su ana amfani da su yi aiki tare da hotuna, da kuma aikace-aikace bayanai.

rubuta Policy

A lokacin gyare-gyaren da abinda ke ciki na cache da kuma sabunta data a cikin babban memory. A lokacin da bata lokaci da ya wuce tsakanin aikace-aikace bayanai dangane da rikodi manufofin. Akwai biyu main iri:

1. Instant rikodi. Kowane canji da aka rubuta synchronously zuwa babban memory.
2. Jinkirin ko writeback. Update data gudanar lokaci zuwa lokaci, ko a lokacin da nema da abokin ciniki. Don kiyaye hanya ko shi da aka gyara don amfani da wata ãyã tare da jihohin biyu: "datti" ko modified. A cikin hali na wani kuskure za a iya sanya biyu Rokon da nufin babban memory: na farko da ake amfani da rikodin data cewa, an canja daga cache, da kuma na biyu - to karanta so abu.

Yana iya zama irin wannan cewa bayanai zama m zuwa matsakaici buffer. Wannan na faruwa a lokacin da data a cikin babban memory ba tare da yin wani gyara ga cache. Domin daidaito, duk da tace matakai amfani da jituwar ladabi.

zamani kalubale

Tare da karuwa a cikin mita na sarrafawa, kuma Ya ƙãra memory yi, wani sabon matsalar yankunan - da gazawa daga cikin dubawa canja wurin bayanai. Abin da zai iya lura da wani ilmi mutum? Kache na memory da amfani sosai idan mita ne kasa da RAM a processor. Da yawa daga cikinsu suna da nasu matsakaici buffer don rage samun lokaci zuwa ƙwaƙwalwar, wanda shi ne hankali fiye da rajista. A CPU cewa goyi bayan rumfa magance, sau da yawa sanya kananan amma sosai azumi adireshin translation buffer. Amma a wasu lokuta, da cache ba da amfani sosai, kuma wani lokacin kawai halitta matsaloli (amma shi ne yawanci a cikin kwakwalwa, wanda sun halartar da ba sana'a version). Magana daga abin da aka adana bayanai a cikin smartphone, NGO lura da cewa saboda da kananan girman da na'urar wajibi ne don haifar da sababbin dada cache aiwatar. Yanzu wasu wayoyin iya fariya sigogi kamar shekaru goma da suka wuce, m kwamfuta - amma abin da bambanci a cikin size!

Aiki tare da bayanai tsakanin daban-daban buffers

A cache ne amfani a lokacin da daya, da kuma yadda za a kula da tasiri na wannan fasahar, idan mai yawa daga gare su? Wannan matsala ne warware buffer coherency. Akwai uku zaɓuɓɓuka saboda data musayar:

1. M. A cache iya nuna hali kamar yadda ka so.
2. Exclusive. Tsara don kowane musamman hali.
3. Non-m. mai yadu yarda misali.

cache matakan

Adadin su ne yawanci daidai uku ko hudu. A mafi girma da matakin na memory, don haka shi ne bulkier da hankali:

1. L1 cache. The fi sauri matakin cache - na farko. A gaskiya, shi ne wani ɓangare na processor kamar yadda aka located a kan wani guda guntu, da kuma dangantaka da aiki tubalan. Yawancin lokaci zuwa kashi iri biyu: umurci cache da bayanai. Mafi zamani sarrafawa ba aiki ba tare da wannan matakin. Wannan cache aiki a processor mita, don haka saidai shi iya yin kowane Agogon sake zagayowar.
2. L2 cache. Yawancin lokaci located tare da baya daya. A raba memory yankunan. Don gano da darajar, ba za ka bukatar dukan girma, ba a karkashin caching bayanai, raba da adadin na tsakiya, wanda yake a cikin processor.
3. L3 cache. Slow, amma babbar dogara cache. Yawanci, fiye da 24 MB. Ana amfani da aiki tare da data samu daga makon jiya, inda daban-daban na biyu matakin.
4. L4 cache. Amfani da aka kubutar kawai ga high-yi multiprocessing sabobin da mainframes. Aiwatar da shi a matsayin mai raba guntu. Idan ka tambaya cewa an caching data a wani smartphone "Samsung" da kuma neman a shi, wannan matakin - Ina iya cewa ga 5 shekaru daidai sauri.

cache associativity

Wannan shi ne wani muhimman hakkokin halayyar. Associativity na Kama data ake bukata don nuna ma'ana rabawa. Ta, bi da bi, wajibi ne saboda gaskiyar cewa a bi da bi scan na dukan data kasance Lines daukan dubun hawan keke, kuma Yanã ba dukan amfanin. Saboda haka, amfani da RAM Kwayoyin m riko da data cache, domin rage search lokaci. Idan muka kwatanta da matsakaiciyar buffers ne na daya girma amma daban-daban associativity, daya tare da wanda ta yi wani babban nufin yin aiki kasa da sauri, amma tare da babban takamaiman yadda ya dace.

ƙarshe

Kamar yadda ka gani, da aka adana bayanai a karkashin wani yanayi, damar your kwamfuta da aiki da sauri sauri. Amma, alas, akwai har yanzu quite mai yawa al'amurran a kan abin da za ka iya yi aiki na dogon lokaci.