O meni:

Sem še relativno mlad (v '30 letih) inženir informatike iz Ljubljane. Teme, ki me najbolj zanimajo: dogodki po svetu (novice, politika), zgodovina, znanost (sploh vesolje)

Nesmiselnost uporabe registry cleaner-jev

Za pisanje te današnje objave me je navdahnila debata v diskusiji z naslovom “Registry Cleaners“: http://episteme.arstechnica.com/…/853006300931 na “Ars OpenForum” forumu (moj nick tam je shirker), katera se je razvila kot posledica vprašanja OP-ja v stilu: “What do you do about registry bloat?” (jah in posledično vprašanja smiselnosti t.i. “registry cleaner” programov) in v kateri sem seveda tudi sam sodeloval. In pa, tisto tretjo objavo v seriji objav v zvezi s PC-jevimi trdimi diski pa prestavljam nekam v prihodnost, ker sem se pač odločil dodati še četro objavo in bom s tem to serijo smiselno razdelil na 2 dela po dve posamezni objavi, kateri sta bajdvej obe že skoraj v celoti napisani/sformatirani.

Torej, v zgoraj omenjenem thread-u v temle postu uporabnik z nickom PeterB omenja (bajdvej, on je eden najbolj spoštovanih uporabnikov na Ars forumu), da je napisal programček, s katerim je naredil en zelo zanimiv test: testiral je namreč razliko v času, ki jo OS porabi za odpiranje/iskanje keys-ov in values-ov (tudi npr. neobstoječih) itd. po registry-ju, šlo pa je za razlike med ogromnim test registry-hive fajlom, registry-hive fajlom srednje velikosti in čisto majhnim registry-hive fajlom.

OK, so. Registry performance. I wrote a crude little program that created a bunch of registry entries in various configurations, and then timed some operations on those entries. I did kinds of operation; opening keys, opening non-existent keys, querying values, and querying non-existent values. It’s not scientific, but it does give a feel for how the performance of the registry is influenced by registry size.

The largest test I did created 512 keys. Each key contained a further 512 keys, and each of those keys contained 8 values. This resulted in a registry hive containing 262144 keys and 2097152 values, which took up approximately 140 MB of space. The next smallest test was 256x256x8, for a total of about 35 MB space. Then 128x128x8, for a total of about 9 MB.

For 100 key opens, the largest registry file took 0.0018 seconds. The middle one took 0.0014 seconds. The smallest took 0.0010 seconds. For 100 failed key opens, the largest took 0.0018 seconds; the middle one took 0.0012; the smallest 0.0008.

Clearly, the registry does not suffer any significant performance impact from being large. The value in “registry cleaners” is, from a performance standpoint, highly questionable. Even if the registry were shrunk to 10% of its size, it is unlikely to provide any discernible difference in performance.

Naknadno pa dodajam še kopijo tagale posta od nekega drugega uporabnika z nickom Hat Monster (ki je tudi zelo spoštovan med Arsian-i), ki uporabi preprosto analogijo, s katero prikaže, na kakšen način (npr. napram branju .txt fajlov) deluje registry.

Ugh. This again?

The registry is a demand paged heirarchial database file. It’s not a text file. It’s not a bloody Word document.

Performing a random read on the registry should not depend on the registry’s size on disk or size in RAM (which should be as near as damn it identical).

Let me use a quick analogy.

I ask you to open c:my worktext filesreportsq108.txt.

Are you slowed down if c: also contains Windows, Program Files, System Volume Information, Documents and Settings or indeed any other number of folders? Are you slowed down if My Work contains, as well as text files, other folders such as images, presentations, webpages, etc?

No, you’re not. You press winkey+R and type the path out. Neither is the registry.

Medtem ko se pa moj post v tej diskusiji (za katerega vam tudi priporočam, da si ga preberete) nahaja na strani 1, direktno nanj pa lahko pridete s klikom na tale link; mimogrede, v njem omenjam tudi program Registry First Aid, katerega jaz uporabljam za urejanje vrednosti več keys-ov naenkrat itd. Aja no, je pa tu sicer še moj drugi post (na strani 2, kot tudi oba zgornja sem skopirana posta), ki pa sam po sebi niti ni tako zelo relevanten.

P.S. – Veliko bolj obširno pa o vsem tem pišem tudi sam na myths.html strani v že dolgo časa nazaj napisanem (recimo temu) članku pod sekcijo naslovljeno z “THE REGISTRY-CLEANING SOFTWARE MYTH” na mojem glavnem izključno z računalništvom povezanem domačem spletnem mestu, ki je sicer resda v celoti napisano v angleškem jeziku!!

Tadej

Advertisements

One Comment on “Nesmiselnost uporabe registry cleaner-jev”

  1. programov pravi:

    […] – but not always on opensubtitles:) Czech dance radio Santamusic.cz will promote opensubtitles.orgNesmiselnost uporabe registry cleaner-jev Tadej: spletne objaveZa pisanje te danaÅ¡nje objave me je navdahnila debata v diskusiji z naslovom Registry […]


Oddajte komentar

Fill in your details below or click an icon to log in:

WordPress.com Logo

Komentirate prijavljeni s svojim WordPress.com računom. Odjava / Spremeni )

Twitter picture

Komentirate prijavljeni s svojim Twitter računom. Odjava / Spremeni )

Facebook photo

Komentirate prijavljeni s svojim Facebook računom. Odjava / Spremeni )

Google+ photo

Komentirate prijavljeni s svojim Google+ računom. Odjava / Spremeni )

Connecting to %s