问 都是选择哪种编译器？ 用realview，keil的cARM 还是 gnu ? 本来很想用CARM，结果不支持 double，哎！ 答 1: 还好有long long 答 2: float 呢？ 答 3: 我觉得奇怪，帮助文档里面明明写了支持double，但我试了试，好象double跟float一样？？难道哪里有参数可以设置么？？？

Data Storage Formats

The CARM C Compiler offers a number of basic data types to use in your C programs. The following table lists these data types along with their size requirements and value ranges.
Data TypeBitsBytesAlignmentValue Range

signed char1	8	1	1	-128 to +127
unsigned char	8	1	1	0 to 255
signed short1	16	2	2	-32768 to +32767
unsigned short	16	2	2	0 to 65535
enum2	32	4	4	-2147483648 to +2147483647
signed int1	32	4	4	-2147483648 to +2147483647
unsigned int	32	4	4	0 to 4294967295
signed long1	32	4	4	-2147483648 to +2147483647
unsigned long	32	4	4	0 to 4294967295
signed long long1	64	8	4	-263 to +263-1
unsigned long long	64	8	4	0 to 264-1
float	32	4	4	±1.175494E-38 to ±3.402823E+38
double	64	8	4	±1.7-308 to ±1.7E+308
pointer	32	4	4	32-bit Address

Note

1. Signed types are stored in two's complement.
2. The enum type is stored using the shortest signed or unsigned integer type required to contain its value.

Other data types, like structures and unions, may contain scalars from this table.
Copyright (c) Keil Software, Inc. and Keil Elektronik GmbH. All rights reserved. 答 4: 选realview比较好! 答 5: float 和 double 差很多啊float   24bit + 8bit = 32bit

double  48bit + 16bit = 64bit 答 6: 圈圈帮我找找，怎么才能让CARM支持double好像要一个caldp.LIB 现在的是calfp.LIB 答 7: 在keil web上找到这个，呜呜！CARM: CALDP.LIB NOT FOUND

--------------------------------------------------------------------------------

Information in this support solution applies to:

CARM All Versions

--------------------------------------------------------------------------------

QUESTION
My project is using double-precision floating arithmetic. When linking my program, the linker does not find CALDP.LIB. This library CANnot be found in \LIB folder. CAN you please provide it?

ANSWER
The CARM compiler toolchain does not support the double-precision (64-bit) floating point data type.

SOLUTION
In MDK-ARM, double-precision floating point support is provided by using the RealView Compiler.

MORE INFORMATION
Refer to Selecting the Toolset in the µVision User's Guide.



答 8: 晕菜了……既然CARM中没有double，怎么跟帮助里不一样呢。看来只好用realView了。 答 9: 晕菜~~~早该用realView了~~~看你再美~~~ 答 10: 晕~~~~~~hotpower大叔这会在旁边幸灾乐祸了…………RealView好象没搞__at这一套？ 答 11: RV就是没at,其他都不晕~~~ 答 12: 比较喜欢 __at 哈哈 答 13: Keil 的C就是不支持double我以前实验过，只支持float而已。要想可靠点还是选RV算了。