X-Nico

unusual facts about Floating point



Am386

Floating point performance of the Am386 could be boosted with the addition of a 80387DX or 80387SX coprocessor, although performance would still not approach that of the on-chip FPU of the 486DX.

Microsoft Binary Format

In computing, Microsoft Binary Format (MBF) was a format for floating point numbers used in Microsoft's BASIC language products including MBASIC, QuickBasic and GW-BASIC.

OpenEXR

It is notable for supporting 16-bit-per-channel floating point values (half precision), with a sign bit, five bits of exponent, and a ten-bit mantissa.


see also

Adjusted Peak Performance

APP was itself replaced in November 2007 when the BIS amended 15 CFR to include the December 2006 Wassenaar Arrangement Plenary Agreement Implementation's new metric - Gigaflops (GFLOPS), one billion floating point operations per second, or TeraFLOPS, one trillion floating point operations per second.

High-level emulation

Emulating such an architecture programmatically on a PC, characterised by the emphasis put on floating-point operations, and specialised graphics hardware with memory separate from the system memory would be extremely difficult, especially taking into account the scarcity of documentation typical for specialised, proprietary hardware.

J Strother Moore

He and others used ACL2 to prove the correctness of the floating point division operations of the AMD K5 microprocessor in the wake of the Pentium FDIV bug.

RGBE image format

By using a shared exponent, the RGBE format gains some of the advantages of floating point values without the 12 bytes per pixel needed for single precision IEEE floating-point values, or 6 bytes in half precision (and which would cover smaller range).

SGI Indigo² and Challenge M

There was a special version of the "Teal" Indigo², called Power Indigo²; this system had increased FPU (floating point unit) capabilities and used specially designed R8000 CPUs.

Transmeta Efficeon

Internally, the Efficeon has two arithmetic logic units, two load/store/add units, two execute units, two floating-point/MMX/SSE/SSE2 units, one branch prediction unit, one alias unit, and one control unit.

Vancouver Stock Exchange

The history of the exchange's index provides a standard case example of large errors arising from seemingly innocuous floating point calculations.