Taksonomia Flynna
(c) I, Cburnett, CC-BY-SA-3.0
SISD
(c) I, Cburnett, CC-BY-SA-3.0
MISD
(c) I, Cburnett, CC-BY-SA-3.0
SIMD
(c) I, Cburnett, CC-BY-SA-3.0
MIMD
Taksonomia Flynna – klasyfikacja architektur komputerowych, zaproponowana w latach sześćdziesiątych XX wieku przez Michaela Flynna, opierająca się na liczbie przetwarzanych strumieni danych i strumieni rozkazów.
W taksonomii tej wyróżnia się cztery grupy:
- SISD (ang. single instruction, single data) – przetwarzany jest jeden strumień danych przez jeden wykonywany program - komputery skalarne (sekwencyjne).
- SIMD (ang. single instruction, multiple data) – przetwarzanych jest wiele strumieni danych przez jeden wykonywany program - tzw. komputery wektorowe.
- MISD (ang. multiple instruction, single data) – wiele równolegle wykonywanych programów przetwarza jednocześnie jeden wspólny strumień danych. W zasadzie jedynym zastosowaniem są systemy wykorzystujące redundancję (wielokrotne wykonywanie tych samych obliczeń) do minimalizacji błędów.
- MIMD (ang. multiple instruction, multiple data) – równolegle wykonywanych jest wiele programów, z których każdy przetwarza własne strumienie danych – przykładem mogą być komputery wieloprocesorowe, a także klastry i gridy.
Linki zewnętrzne
- Strona Michaela Flynna. arith.stanford.edu. [zarchiwizowane z tego adresu (2007-10-11)].
Media użyte na tej stronie
MIMD.svg
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a MIMD design: multiple instruction, multiple data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a MIMD design: multiple instruction, multiple data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
SISD.svg
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a SISD design: single instruction, single data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a SISD design: single instruction, single data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
SIMD.svg
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a SImD design: single instruction, multiple data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a SImD design: single instruction, multiple data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
MISD.svg
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a MISD design: multiple instruction, single data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.
(c) I, Cburnett, CC-BY-SA-3.0
Flynn's Taxonomy of a MISD design: multiple instruction, single data. Each "PU" (processing unit) does not necessarily correspond to a processor, just some functional unit that can perform processing. The PU's are indicated as such to show relationship between instructions, data, and the processing of the data.