RSS订阅信息安全技术跟踪与研究:技术、平台、会议、论文、产业
你现在的位置:首页 / 技术积累 / 正文

52知识点之三:各种计算设备的计算和存储能力比较

0 技术积累 | 2015年3月24日
转载申明:本站原创,欢迎转载。但转载时请保留原文地址。
原文地址:http://www.vonwei.com/post/3outof52.html

52 Things: Number 3: Computational and storage power of different form factors

Posted by Luke Mather

This is the third in a series of blog posts to address the list of '52 Things Every PhD Student Should Know' to do Cryptography. The set of questions has been compiled to give PhD candidates a sense of what they should know by the end of their first year. We will be presenting answers to each of the questions over the next year, one per week, and I am the student assigned to the third question:

Q3: Estimate the relative computational and storage capabilities of 

  • a smart-card

  • a micro-controller (i.e. a sensor node)

  • an embedded or mobile computer (e.g., a mobile phone or PDA)

  • a laptop- or desktop-class computer.

To measure the computational capability of a device we could assess the clock speed of its processors. This may be misleading if the processor enables some form of parallelism---two cores running at 2 GHz obviously possess more computational power than a single core running at 2 GHz, and so finding a direct quantitative measure is not a realistic expectation. For specific devices like general purpose graphics cards, often the total FLOPS (floating point operations per second) the device is capable of sustaining is reported (for either single or double precision arithmetic) but even this measure is not a particularly reliable choice when applied to any given problem---indeed, some services facilitate a comparison by benchmarking the performance of different devices on a variety of problem instances---see, for example, CompuBench. Fortunately the range of capabilities of the devices included in the question makes a sufficient answer less dependent on quantitative metrics.

A measure for the storage capabilities of each device is much simpler to find: we can simply compare the approximate number of bytes of information the device is capable of holding on permanent storage.

A smart-card is the least computationally powerful device: obviously clock speeds vary for different implementations, but one might expect to see around a 20 MHz core speed. In terms of storage, a typical smart-card might have around 2 kilobytes (KiB) available.

A microcontroller is "a small computer on a single integrated circuit containing a processor core, memory, and programmable input/output peripherals" [1]. The range of storage and compute capability available will vary significantly according to the exact definition of microcontroller, but taking the suggested sensor node as an example, a typical microcontroller is likely to have similar computational capabilities as a smart-card and slightly more storage available, perhaps in the order of a few KiB to a few megabytes (MiB).

A mobile computer such as a mobile phone has significantly more storage and computing power, and the amount of power available is rapidly increasing over time. Taking the 2008 iPhone and the 2013 Nexus 5 phone as an example, the iPhone used a 412 MHz 32-bit RISC ARM core, and the Nexus 5 CPU used is a 2.3 GHz quad-core processor. In terms of storage, if we ignore the ability of some phones to use removable storage, then a high-end phone in 2013 might expect to provide in the order of 16 to 32 gigabytes (GiB) of storage.

Finally, most laptop or desktop class computers are likely to have more processing power than a mobile phone: the high-end Intel "Haswell" i7 4960K processor contains 4 cores each clocked at 4 GHz, and the AMD "Piledriver" FX-9590 CPU contains 8 cores at 4.7 GHz---note that a direct comparison between these two processors requires more than just assessing core counts and clock speeds! There are other factors that can affect the computing capabilities of a desktop or laptop computer---in particular, the addition of a graphics processing unit can, for certain problems, provide a large increase in performance. The storage capacity of a laptop or desktop can vary tremendously, but a typical amount of storage in a consumer machine might be between hundreds of gigabytes and several terabytes (TiB)---the largest single hard drive capacities are now around 8 TiB

 

翻译和补充

         这周的问题是评估各种计算设备(智能卡、微控制器、嵌入式或者移动计算机、便携或者桌面类计算机)的相对计算能力和存储能力。这个问题还是比较具有实际意义的,在实际研究中,避免不了使用各种计算设备和环境来实现你的原型系统,这时选择怎样的计算设备,充分考虑它们的计算能力和存储能力是非常必要的。例如,一些密码安全协议可能是针对传统PC的,而在嵌入式或者移动设备上使用,考虑到计算能力和存储能力的不一致,可能需要设计更加轻量级的协议。

为了度量一个设备的计算能力,可以通过评估其处理器的时钟速度来达到。如果仅仅依赖这一点,会存在一定的误导,有些处理器可能支持并行操作,例如,一个2GHz的双核处理器的计算能力显然要强于一个2GHz的单核处理器。因此,找一个直接的量化度量方法并不一定符合实际期望。对于像通用显卡等特殊设备,通常报告其能支持的每秒浮点运算次数(FLOPS),无论其执行单精度运算还是双精度运算。但是,浮点运算这种度量方式也并不能满足所有问题。不过,确实也存在有些网络服务(如CompuBenchhttp://compubench.com)可以帮助比较各种设备、各种处理器的性能,这些比较结果可以提供一定的参考。比较幸运的是,本问题考虑的计算或者存储能力范围,并不需要依赖量化指标就可以给出比较满意的回答。

度量一个设备的存储能力就更容易解决:可以简单的比较设备长久存储支持的比特数。其实关于这一点,我并不是很同意原作者。一个设备的存储能力并不能只由其长久存储能力决定。通常计算设备的存储也分很多类型,如内存、磁盘、缓存、外设存储、安全存储等。要决定一个设备的真实存储能力也需要依赖具体需求。有些应用场景可能需要大容量的磁盘,而有些场景(不希望断电后拥有数据)可能需要更多的内存。

一个智能卡可以称得上拥有最少计算能力的设备,智能卡应用很广泛,如加密解密卡、IC卡、门禁卡、支付卡、校园卡等。不同的实现,其时钟速度肯定不一样,不过一般可以期望的是20MHz左右(一个核)。至于存储,一个典型的智能卡大概是2KB左右。

一个微控制器(也称单片机)的定义为:一个微型计算机,把处理器核、存储和可编程输入输出外设等集成在一块集成电路芯片上[1]。微控制器的计算和存储能力与其在实际中的使用和确切定义有关。不过以问题建议的传感器节点为例,一个典型的微控制器的计算能力与一个智能卡差不多,而存储能力可能会增加一点,大约几KB或者几MB的增加。

相比智能卡和微控制器,一个嵌入式或者移动设备的计算和存储能力增长就比较明显了,而且随着时间推移,移动设备的能力也在不断增长。以2008年的iPhone手机和2013年的Nexus 5手机为例,iPhone使用412MHz 32ARM核,Nexus使用2.3GHz四核处理器。至于存储,如果忽略一些手机拥有可移动外设(如SD卡)的能力,2013年大部分手机的存储大概以16GB32GB为主。现在情况更加严重,可以说移动设备的计算和存储能力增长非常明显,如笔者用的小米note配置为2.5GHz四核,64GB机身存储,确实够奢侈了。

最后,桌面类计算机就不用多说了,处理能力远超前面几个。如高端Intel Haswell i7 4960K处理器包含四个核,每个核的时钟速度为4GHzAMD Piledriver Fx-9590 CPU包含8个核,每个核时钟速度为4.7GHz。注意,想要直接比较这两个处理器,仅仅依靠核数和时钟速度还是远远不够的,还需要考虑其它能够影响桌面类计算机计算能力的因子,如增加一个GPU图形处理单元,在处理一些问题时计算能力会提升很多。桌面类计算机的存储能力是千差万别的,对于普通用户,通常是几百GB到几TB之间,目前最大的独立硬盘空间大约8TB左右。而至于服务器那计算能力和存储能力就更加强大了,可以分割中多个虚拟机,每个虚拟机可以相当于一个桌面机器的能力。

 

[1] https://en.wikipedia.org/wiki/Microcontroller


  • ------------------分隔线----------------

  • 如果感兴趣,欢迎关注本站微信号,跟踪最新博文信息,手机微信扫一扫下面的二维码,即可关注!
  • 微月信公众号
  • 推荐您阅读更多有关于“ 智能卡  Smart card  微控制器  嵌入式设备  移动设备  桌面计算机  52知识点   ”的文章

    请填写你的在线分享代码
    上一篇:ThinkPad T440p入手体验下一篇:如果没有360,还有这些堪称神器的软件可供选择

    猜你喜欢

    评论列表:

    发表评论

    必填

    选填

    选填

    必填,不填不让过哦,嘻嘻。

    记住我,下次回复时不用重新输入个人信息

    本站介绍
    最近发表
    本年最热文章
    本月最热文章
    网站分类
    文章归档