原文： Inside Of GM200 Full Maxwell Big Daddy
Whats this difference between this and earlier iterations?
In-Depth On The SMM
Maxwell Adds Performance Using Less Power. In short, the company says it increased utilization of the SMM by reconfiguring its resources.
Like the Ge
Force GTX 980 and 970 introduced to last September, GM200 is based on Nvidia’s efficient Maxwell architecture. Instead of GM204’s four Graphics Processing Clusters, you get six. And with four Streaming Multiprocessors per GPC, that adds up to 24 SMMs across the GPU. Multiply out the 128 CUDA cores per SMM and you get GeForce GTX Titan X’s total of 3072. Eight texture units per SMM add up to 192with a base core clock rate of 1000MHz, that’s 192 GTex/s (the original GeForce GTX Titan was rated at 188, despite its higher texture unit count).
Like the SMMs found in GM204, GM200 exposes 96KB of shared memory and 48KB of texture/L1 cache, doubling what it had in the GeForce GTX 750 Ti’s GM107. The other architectural elements are similar though; each SMM is broken up into four blocks, with their own instruction buffer, warp scheduler and pair of dispatch units. In fact, so much is carried over that double-precision math is still specified at 1/32 the rate of FP32, even though GM200 is the Maxwell family’s big daddy. Incidentally, an upcoming Quadro card based on the same GPU shares this fate. If FP64 performance is truly important to you, Nvidia would likely suggest one of its Tesla boards..
Some people feel that NVIDIA has been dishonest because it has omitted the FP64 performance, and is still charging a premium for this single card solution.
These functions are usually utilized for the scientific community, that being said if you can get this card at a reasonable price do so, as you will be doing yourself a favor and futureproofing your system for years to come.
12gb! everyone says that's overkill, guess what it's not. we are looking at a landscape that is requiring a larger and larger framebuffer and its just easier to program and code having the onboard Vram being used.
The card is a beast and has surpassed my expectations and I am an early adopter of bleeding technology. I purchased 2 of these cards. It was a big hit to the wallet accompanied with the thing I always tell myself ...I done ! im not upgrading anymore! I don't need anything more. I'm going to be homeless and hungry one day because of my addiction to the best tech has to offer.
Bottom line: there is no better card on the market today. Also its nice to purchase from Amazon because of the hassel free returns, I also purchased a protection plan, just in case. I think having the right insurance of bleeding edge computer parts is a no brainer, its like backing up. just do it.
If you have any specific questions, please feel free to ask.
显示更多>>
译文： 使用了GM200大核心
，那这个究竟和以前的迭代有什么不同？
更多的使用SMM Maxwell以使用更少的功率来提高性能。总之，公司表示会通过重新配置资源来提高SMM的利用率。
就像去年9月推出的GeForce GTX 980和970一样，GM200基于Nvidia高效的Maxwell架构。不像GM204的是四个图形处理集群，这里会有六个。而且每个GPC有四个流式多处理器，在整个GPU中共计24个SMM。将每个SMM的128个CUDA内核数乘以每个SMM的GeForce GTX Titan X总共3072个。每个SMM的8个纹理单元总计192个，基本内核时钟频率为1000MHz，即192 GTex / s（原始的GeForce GTX Titan评为188，尽管它的纹理单元数更高）。
就像GM204中的SMM一样，GM200显示96KB的共享内存和48KB的纹理/ L1缓存，比GeForce GTX 750 Ti的GM107增加了一倍。其他的结构元素比较相似，每个SMM分为四个块，分别有自己的指令缓冲区，warp调度程序和一对调度单元。事实上，即使GM200是麦克斯韦家族的优秀产品，双倍精度的数学仍然是FP32的1/32。顺便说一下，即将推出的基于同一GPU的Quadro卡也是差不多的命运。如果FP64的性能对你来说真的很重要的话，Nvidia可能会推荐其中一款Tesla主板。
有些人认为，NVIDIA并不够实惠，因为它已经略去了FP64性能，并且仍然在为这种单卡解决方案收费。
这些功能通常用于科学界，如果你能以合理的价格获得这张卡，那么你就可以这么做了，因为在未来的几年里，你可以做好准备了，因为这将使你拥有很好用的的系统。
12GB！大家都说这太夸张了，他们都猜测这个很可能达不到。但我们正在寻找一个需要更大和更大的帧缓冲的格局，而且更容易使用板载Vram的内存进行编程和编码。
这个显卡很好用而且远超过了我的期望，我是一个喜欢网络游戏的人。我购买了2个显卡。当然我也一直在告诉自己这是一个掏空的钱包的事情！我后来也没有再升级了！我什么都不需要做了，我觉得我会因为对高级科技产品的嗜好而有一天变得饥肠辘辘和无家可归。
最后一点：这个显卡已经可以算是如今市场上最好的显卡了。另外，因为亚马逊有着免费退货服务，所以我们是从这里购买的，我也购买了购物险，以防万一。我觉得购买这个保险是正确，为了计算机零部件买一份保险不需要考虑。去做就对了。
如果您有任何具体问题，请随时询问我。
