The Multimedia Acceleration eXtensions or MAX are instruction set extensions to the Hewlett-Packard PA-RISC instruction set architecture (ISA). MAX was developed to improve the performance of multimedia applications that were becoming more prevalent during the 1990s.
MAX instructions operate on 32- or 64-bit SIMD data types consisting of multiple 16-bit integers packed in general purpose registers. The available functionality includes additions, subtractions and shifts.
The first version, MAX-1, was for the 32-bit PA-RISC 1.1 ISA. The second version, MAX-2, was for the 64-bit PA-RISC 2.0 ISA.
Notability
editThe approach is notable because the set of instructions is much smaller than in other multimedia CPUs, and also more general-purpose. The small set and simplicity of the instructions reduce the recurring costs of the electronics, as well as the costs and difficulty of the design. The general-purpose nature of the instructions increases their overall value. These instructions require only small changes to a CPU's arithmetic-logic unit. A similar design approach promises to be a successful model for the multimedia instructions of other CPU designs.[1][2][3] The set is also small because the CPU already included powerful shift and bit-manipulation instructions: "Shift pair" which shifts a pair of registers, "extract" and "deposit" of bit fields, and all the common bit-wise logical operations (and, or, exclusive-or, etc.).[2]
This set of multimedia instructions has proven its performance, as well. In 1996 the 64-bit "MAX-2" instructions enabled real-time performance of MPEG-1 and MPEG-2 video while increasing the area of a RISC CPU by only 0.2%.[1]
Implementations
editMAX-1 was first implemented with the PA-7100LC in 1994. It is usually attributed as being the first SIMD extensions to an ISA. The second version, MAX-2, was for the 64-bit PA-RISC 2.0 ISA. It was first implemented in the PA-8000 microprocessor released in 1996.[1]
The basic approach to the arithmetic in MAX-2 is to "interrupt the carries" between the 16-bit subwords, and choose between modular arithmetic, signed and unsigned saturation. This requires only small changes to the arithmetic logic unit.[2]
MAX-1
editInstruction | Description |
---|---|
HADD | Parallel add with modulo arithmetic |
HADD,ss | Parallel add with signed saturation |
HADD,us | Parallel add with unsigned saturation |
HSUB | Parallel subtract with modulo arithmetic |
HSUB,ss | Parallel subtract with signed saturation |
HSUB,us | Parallel subtract with unsigned saturation |
HAVE | Parallel average |
HSHLADD | Parallel shift left and add with signed saturation |
HSHRADD | Parallel shift right and add with signed saturation |
MAX-2
editMAX-2 instructions are register-to-register instructions that operate on multiple integers in 64-bit quantities. All have a one cycle latency in the PA-8000 microprocessor and its derivatives. Memory accesses are via the standard 64-bit loads and stores.
The "MIX" and "PERMH" instructions are a notable innovation because they permute words in the register set without accessing memory. This can substantially speed many operations.[2]
Instruction | Description |
---|---|
HADD | Parallel add with modulo arithmetic |
HADD,ss | Parallel add with signed saturation |
HADD,us | Parallel add with unsigned saturation |
HSUB | Parallel subtract with modulo arithmetic |
HSUB,ss | Parallel subtract with signed saturation |
HSUB,us | Parallel subtract with unsigned saturation |
HSHLADD | Parallel shift left and add with signed saturation |
HSHRADD | Parallel shift right and add with signed saturation |
HAVG | Parallel average |
HSHR | Parallel shift right signed |
HSHR,u | Parallel shift right unsigned |
HSHL | Parallel shift left |
MIX | Mix 16-bit sub-words in a 64-bit word; MIX Left, Ra,Rb,Rc, Rc:=a1,b1,a3,b3; MIX Right, Rc:=a2,b2,a4,b4[2] |
MIXW | Mix 32-bit sub-words in a 64-bit word; e.g. MIXW Left, Ra,Rb,Rc, Rc:=a1,a2,b1,b2; MIXW Right, Rc:=a3,a4,b3,b4[2] |
PERMH | Permute 16-bit sub-words of the source in any possible permutation in the destination register, including repetitions.[2] |
References
edit- ^ a b c Lee, Ruby B. (August 1996). "Subword Parallelism with MAX-2" (PDF). IEEE Micro. 16 (4): 51–59. doi:10.1109/40.526925. Retrieved 21 September 2014.
- ^ a b c d e f g Lee, Ruby; Huck, Jerry (February 25, 1996). "64-bit and multimedia extensions in the PA-RISC 2.0 architecture". COMPCON '96. Technologies for the Information Superhighway Digest of Papers. pp. 152–160. doi:10.1109/CMPCON.1996.501762. ISBN 0-8186-7414-8. S2CID 13081443.
- ^ Lee, Ruby B. (April 1995). "Accelerating Multimedia with Enhanced Microprocessors" (PDF). IEEE Micro. 15 (2): 22–32. doi:10.1109/40.372347. Retrieved 21 September 2014.