Intel MMX

Views:
 
Category: Education
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

Seminar on Intel MMX Technology Presented by G.EETHANJALI MCA third ROLL NO:1011120 :

Seminar on I ntel MMX Technology Presented by G.EETHANJALI MCA third ROLL NO:1011120

CONTENTS:

CONTENTS Introduction Definition process Basic concepts Packed data format Conditional execution Saturating arithmetic Fixed-Point Arithmetic Data alignment Features New data types 64-bit mmx registers Performance advantage Conclusion References

INTRODUCTION:

INTRODUCTION Intel’s MMX technology is an extension to the basic Intel Architecture (IA) designed to improve performance of multimedia and communication algorithms . The definition of MMX technology evolved from earlier work in the i860 architecture . The i860 processor provided instructions that operated on multiple adjacent data operands in parallel.

Definition process:

Definition process MMX technology’s definition process was an outstanding adventure for its participants, a path with many twists and turns . It was a bottom-up process. Engineering input and managerial drive made MMX technology happen . The definition of MMX technology was guided by a clear set of priorities and goals set forth by the definition team . It was also imperative that processors with MMX technology retain backward compatibility with existing software, both operating systems and applications

Basic Concepts :

Basic Concepts Saturating arithmetic vs. wrap-around arithmetic Fixed-point arithmetic. Repositioning data elements within packed data format. Data alignment.

Packed Data Format :

Packed Data Format MMX technology defines new register formats for data representation. The key feature of multimedia applications is that the typical data size of operands is small Most of the data operands’ sizes are either a byte or a word (16 bits). . These two properties lend themselves to the use of SIMD computation.

Conditional Execution :

Conditional Execution What happens when a computation is only done if the operand value passes some conditional check? Operating on multiple data operands using a single instruction presents an interesting issue. Using a branch approach does not work well for two reasons: first, a branch-based solution is slower because of the inherent branch misprediction penalty, and second, because of the need to convert packed data types to scalars.

Fixed-Point Arithmetic :

Fixed-Point Arithmetic Media applications involve working on fraction values . for example, the use of a weighting coefficient in filtering averaging. Looking at application requirements and the trade-off of performance and design complexity leads to the use of a fixed-point arithmetic paradigm for several media applications.

Data Alignment :

Data Alignment Use of packed data also presents data alignment issues. In some cases, the data may be aligned on its natural boundary and not on size of the packed data operand. Therefore, in some cases, there is a need to support efficient access of unaligned data for media applications. shift-left and shift-right operations on 64 bits. These instructions enable using a sequence of Shift left, Shift right, and Or operations to assemble the desired byte from the aligned data that encompasses the desired bytes.

features:

features New data types built by packing independent small data elements together into one register. An enhanced instruction set that operates on all independent data elements in a register, using a parallel SIMD fashion. New 64-bit MMX registers that are mapped on the IA floating-point registers. Full IA compatibility.

New Data Types :

New Data Types MMX technology introduces four new data types: three packed data types and a new 64-bit entity. Each element within the packed data types is an independent fixed-point integer. This adds a burden on the user, but it also leaves a large amount of flexibility to choose and change the precision of fixed-point numbers during the course of the application in order to fully control the dynamic range of values.

64 bit MMX registers:

64 bit MMX registers MMX technology provides eight new 64-bit general purpose registers that are mapped on the floating-point registers. MMX registers are random access registers, that is, they are not accessed via a stack model like the floating-point registers. MMX registers are used for holding MMX data only.

MMX registers:

MMX registers The exponent field of the corresponding floating-point register (bits 64-78) and the sign bit (bit 79) are set to ones (1’s), making the value in the register a NaN (Not a Number) or infinity when viewed as a floating-point value. This helps to reduce confusion by ensuring that an MMX data value will not look like a valid floating-point value. The dual usage of the floating-point registers does not preclude applications from using both MMX code and floating-point code.

Use of FP DLL Model for MMX Code:

Use of FP DLL Model for MMX Code To enable common multimedia applications for processors with and without MMX technology, we chose to promote the Dynamic Linked Library (DLL) model as the primary model to support MMX instructions. In the DLL model, depending upon whether the processor provides MMX technology support in hardware (the processor CPUID provides this information), the appropriate version of the media library function is linked dynamically. MMX technology DLLs suggest the same guidelines as that of FP DLLs.

Performance Advantage :

Performance Advantage We will analyze the performance enhancement due to MMX technology. The multiply-accumulate (MAC) operation is one of the most Fequent operations in multimedia and communications applications used in basic mathematical primitives like Matrix multiply and filters. A multiply-accumulate operation (MAC) is defined as the product of two operands added to a third operand (the accumulator).

:

This operation requires two loads (operands of the multiplication operation), a multiply, and an add (to the accumulator). MMX technology does not support three operand instructions; therefore, it does not have a full MAC capability.

Intel has introduced two processor families with technology: Pentium processor with mmx technology and pentium2 processor.:

Intel has introduced two processor families with technology: Pentium processor with mmx technology and pentium2 processor . Intel Pentium processor Pentium 2 processor

The performance of both processors was compared on the Intel Media Benchmark (IMB) [5,6], which measures the performance of processors running algorithms found in multimedia applications. The IMB incorporates audio and video playback, image processing, wave sample rate conversion, and 3D geometry. Figure 6 and Table 2 compare the Pentium processor with MMX technology and the Pentium II processor against the Pentium processor and the Pentium® Pro processor. :

The performance of both processors was compared on the Intel Media Benchmark (IMB) [5,6], which measures the performance of processors running algorithms found in multimedia applications. The IMB incorporates audio and video playback, image processing, wave sample rate conversion, and 3D geometry. Figure 6 and Table 2 compare the Pentium processor with MMX technology and the Pentium II processor against the Pentium processor and the Pentium® Pro processor.

The performance of both processors was compared on the Intel Media Benchmark (IMB) which measures the performance of processors running algorithms found in multimedia applications.:

The performance of both processors was compared on the Intel Media Benchmark (IMB) which measures the performance of processors running algorithms found in multimedia applications.

Intel media bench mark performance comparison :

Intel media bench mark performance comparison

conclusion:

conclusion MMX technology implements a high-performance technique that enhances the performance of Intel Architecture microprocessors for media applications. The core algorithms in these applications are computer intensive . These algorithms perform operations on a large amount of data, use small data types, and provide many opportunities for parallelism.

References :

References http://pentium.intel.com/] Intel Media Benchmark URL:procs/per/ icp /imb.htm . ] A. Peleg, U. Weiser, MMX™ Technology Extension to the Intel Architecture, IEEE Micro, Vol. 16, No. 4, August 1996, pp. 42-50. [2] A. Peleg, S. Wilkie, U. Weiser, Intel MMX for Multimedia PCs, Communications of the ACM, Vol. 40, No. 1, January 1997, pp. 25-38. [3] Intel Corporate Literature, i860™ Microprocessor Family Programmers Reference Manual, Order number 240875.Intel Corporate Literature Sales, 1991. [4] Pentium®Family User’s Manual, Volume 3: Architecture and Programming Manual, Order number 241430, Intel Corporate Literature Sales, Mt. Prospect, IL,1994. [5] M. Slater, The Land Beyond Benchmarks, Computer and Communications OEM Magazine, Vol. 4, No. 31,

authorStream Live Help