Compiling for Instruction-Level Parallelism: Advanced Topics, lecture by B. Ramakrishna Rau
1,009 views · Published 30 August 2016 · 1:21:25 · Indexed 20 September 2026
Channel: Computer History Museum · 2016 · People & Blogs
Compiling for Instruction-Level Parallelism Video Series: Advanced Topics. A lecture by B. Ramakrishna Rau. This video was recorded in March, 1994. From University Video Communications' catalog: "Instruction-level parallelism (ILP) is the principal architectural mechanism by which the microprocessor industry is currently maintaining the extraordinary rate of increase of microprocessor performance that we have come to expect. Both superscalar and VLIW architectures require compilers which go beyond conventional optimizing compilers by employing techniques to expose, enhance and exploit ILP. In this talk, Bob Rau discusses techniques for exposing and enhancing ILP including the single assignment form, memory disambiguation, load-store elimination, and critical path length reduction. In addition, he discusses the structure of ILP compilers and how it differs from that of conventional optimizing. " Lot number: X6636.2013 Catalog number: 102624739
More from this channel
-
29:38
Talking to the Future, 2013
-
1:10:29
KQED's John Boland & NPR's Jarl Mohn in Conversation with Museum CEO John Hollar
-
40:54
Principles of OLE 2.0, lecture by Tony Williams
-
35:17
The Digital Library and Future Technologies, lecture by Judith Klavans
-
3:27
DFJ’s Heidi Roizen & CloudFlare’s Michelle Zatlyn’s Advice for Entrepreneurs
-
1:13:13
Oral History of Richard Lyon
-
1:50:01
Oral History of George Gerpheide
-
1:03
CEO Mistaken for Her Assistant