▶ 纹理内存访问补充（见纹理内存博客 http://www.cnblogs.com/cuancuancuanhao/p/7809713.html）
▶ 计算能力
● 不同计算能力的硬件对计算特性的支持。
● 不同计算能力的硬件技术特性（重要）。
● 浮点运算技术标准描述（原文）
■ All compute devices follow the IEEE 754-2008 standard for binary floating-point arithmetic with the following deviations:
There is no dynamically configurable rounding mode; however, most of the operations support multiple IEEE rounding modes, exposed via device intrinsics;
There is no mechanism for detecting that a floating-point exception has occurred and all operations behave as if the IEEE-754 exceptions are always masked, and deliver the masked response as defined by IEEE-754 if there is an exceptional event; for the same reason, while SNaN encodings are supported, they are not signaling and are handled as quiet;
The result of a single-precision floating-point operation involving one or more input NaNs is the quiet NaN of bit pattern 0x7fffffff;
Double-precision floating-point absolute value and negation are not compliant with IEEE-754 with respect to NaNs; these are passed through unchanged;
Code must be compiled with -ftz=false, -prec-div=true, and -prec-sqrt=true to ensure IEEE compliance (this is the default setting; see the nvcc user manual for description of these compilation flags).
Regardless of the setting of the compiler flag -ftz,
Atomic single-precision floating-point adds on global memory always operate in flush-to-zero mode, i.e., behave equivalent to FADD.F32.FTZ.RN,
Atomic single-precision floating-point adds on shared memory always operate with denormal support, i.e., behave equivalent to FADD.F32.RN.
■ IEEE-754R 标准，CUDA 设备上函数 fminf()，fmin()，fmaxf()，fmax() 在有且仅有一个参数为NaN时，返回值将是另一个参数。
■ 浮点数转化为整数，且超出整形格式边界时，IEEE-754 称为未定义，而 CUDA 设备将该数映射到整形格式边界（改格式能表示的最大或最小的数）。
■ 整数除以零或整数溢出时，IEEE-754 称为未定义，而 CUDA 设备将抛出一个非特定的值。
● 各代 CC 设备的新特性介绍
■ 调整共享内存和 L1 缓存的平衡
1 // driver_types.h 2 enum __device_builtin__ cudaFuncCache 3 { 4 cudaFuncCachePreferNone = 0, // 默认缓存模式 5 cudaFuncCachePreferShared = 1, // 扩大共享内存缩小 L1 缓存 6 cudaFuncCachePreferL1 = 2, // 扩大 L1 缓存缩小共享内存 7 cudaFuncCachePreferEqual = 3 // 等量大小的共享内存和 L1 缓存 8 }; 9 10 // cuda_runtime_api.h 11 extern __host__ cudaError_t CUDARTAPI cudaDeviceSetCacheConfig(enum cudaFuncCache cacheConfig);