GTX 970带宽计算
GTX 970 bandwidth calculation
我正在计算gtx970的理论带宽。根据给出的规格:-
http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-970/specifications
内存时钟为 7Gb/s
内存总线宽度 = 256
带宽=7*256*2/8(*2因为是DDR)
= 448 GB/s
但是,在规格中它被指定为 224GB/s
为什么有 2 倍的差异?我是不是说错了,如果是的话请指正。
谢谢
7 Gbps 似乎是有效时钟,即包括数据速率。另请注意,此 Wikipedia list 的字段解释说 "All DDR/GDDR memories operate at half this frequency, except for GDDR5, which operates at one quarter of this frequency",这表明尽管 DDR 缩写,所有 GDDR5 芯片实际上都是四倍数据速率。
最后,让我指出来自 Wikipedia 的注释,它取消了琐碎的 effective clock * bus width
公式:
For accessing its memory, the GTX 970 stripes data across 7 of its 8 32-bit physical memory lanes, at 196 GB/s. The last 1/8 of its memory (0.5 GiB on a 4 GiB card) is accessed on a non-interleaved solitary 32-bit connection at 28 GB/s, one seventh the speed of the rest of the memory space. Because this smaller memory pool uses the same connection as the 7th lane to the larger main pool, it contends with accesses to the larger block reducing the effective memory bandwidth not adding to it as an independent connection could.
报告的时钟速率是 "effective" 时钟速率,并且已经考虑了上升沿和下降沿的传输。麻烦的是DDR的2倍。
这里有一些关于 devtalk 的讨论:https://devtalk.nvidia.com/default/topic/995384/theoretical-bandwidth-vs-effective-bandwidth/
其实你的格式是对的,只是内存时钟不对。 GeForce GTX 970的显存频率为1753MHz(参考https://www.techpowerup.com/gpu-specs/geforce-gtx-970.c2620).
我正在计算gtx970的理论带宽。根据给出的规格:-
http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-970/specifications
内存时钟为 7Gb/s 内存总线宽度 = 256
带宽=7*256*2/8(*2因为是DDR)
= 448 GB/s
但是,在规格中它被指定为 224GB/s
为什么有 2 倍的差异?我是不是说错了,如果是的话请指正。
谢谢
7 Gbps 似乎是有效时钟,即包括数据速率。另请注意,此 Wikipedia list 的字段解释说 "All DDR/GDDR memories operate at half this frequency, except for GDDR5, which operates at one quarter of this frequency",这表明尽管 DDR 缩写,所有 GDDR5 芯片实际上都是四倍数据速率。
最后,让我指出来自 Wikipedia 的注释,它取消了琐碎的 effective clock * bus width
公式:
For accessing its memory, the GTX 970 stripes data across 7 of its 8 32-bit physical memory lanes, at 196 GB/s. The last 1/8 of its memory (0.5 GiB on a 4 GiB card) is accessed on a non-interleaved solitary 32-bit connection at 28 GB/s, one seventh the speed of the rest of the memory space. Because this smaller memory pool uses the same connection as the 7th lane to the larger main pool, it contends with accesses to the larger block reducing the effective memory bandwidth not adding to it as an independent connection could.
报告的时钟速率是 "effective" 时钟速率,并且已经考虑了上升沿和下降沿的传输。麻烦的是DDR的2倍。
这里有一些关于 devtalk 的讨论:https://devtalk.nvidia.com/default/topic/995384/theoretical-bandwidth-vs-effective-bandwidth/
其实你的格式是对的,只是内存时钟不对。 GeForce GTX 970的显存频率为1753MHz(参考https://www.techpowerup.com/gpu-specs/geforce-gtx-970.c2620).