0
\$\begingroup\$

I am quite sure that this is a question which will be marked as "primarily opinion-based" but I would still like to ask it here only as you do not get so many intellectual, experienced and skilled people at same place. I have acquired a keen interest in the field of Computer Architecture (CA) and when I say keen I mean I would like to pursue it no matter what (even though I am not from computer background).I read a few books on this and almost every one of them said that it is a study of Hardware/Software Interface.

Now my first question is "when we talk about this interface is it something like we can study both hardware and software in this field and their interdependencies"?

My primary interest is towards the hardware part of CA. I love playing with muxes, decoders and all. We are still using the old x86 architecture (primarily!) in the modern world.

So second question " Are the researchers in this field not able to come up with anything new in terms of the architecture"?

All the simulations which researchers have to carry out today are based on software which by their own admission gives inaccurate results but since we do not have other alternatives we have to stick with it.

Third question "Where are we in terms of adaptive/reconfigurable architectures which researchers can modify according to their own design to get the results instead of playing around with software scripts and altering them"?

My concerns or a sense of jealousy lies in the fact that in CA software and hardware both plays an equal role. So why can't we limit the software research to compiler optimizations etc. and try and revitalised the research in terms of Hardware/FPGA based simulations.Software people have thousands of things to do but for the hardware guys the opportunities are limited. So why can't we try and be competitive enough to go for this sort of research. I went through the profile of professors in many of the top universities, was not able to find many who may be working on something like this.

P.S: I am not suggesting that software in hardware are bad or something. I am informed enough to realise the fact that software makes our task easier but is it the only way of doing things?

I am writing this after feeling disappointed so have not cared much about the flow. Everyone is welcomed to make suggestion whether it be in Hardware part or the software one.

\$\endgroup\$

closed as too broad by PeterJ, Daniel Grillo, tcrosley, Olin Lathrop, Asmyldof Feb 4 '16 at 22:32

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • \$\begingroup\$ "We are still using the old x86 architecture (primarily!) in the modern world." I would not at all be surprised if ARM has caught up already; it's used in almost every cellphone in existence. \$\endgroup\$ – Ignacio Vazquez-Abrams Jan 16 '16 at 20:33
  • \$\begingroup\$ Intel recently bought Altera, one of the two largest manufacturers of FPGAs, so we are likely to se some major changes in Intel processors in the future. \$\endgroup\$ – Rasmus B. Sorensen Jan 16 '16 at 21:27
5
\$\begingroup\$
  1. Software is all about the manipulation of abstract binary values, whether they are individual bits, or collections of bits that represent characters or numbers. The hardware-software interface is the definition of how those bit values relate to real-world physical phenomena, such as measuring a voltage, reading a switch and turning on/off an LED, sending or receiving a packet of information on a network, or creating a display on a screen.

  2. Researchers are coming up with new architectures all the time. You don't see them in mainstream computers, mainly because the mainstream computers are expected to be able to run all of the software that was primarily written for previous-generation computers. Nowadays, by far the larger investment is in software development, not hardware development, and the value of that investment needs to be preserved.

  3. Today, we have an unprecedented level of flexibility to experiment with new architectures by means of FPGAs. These provide the ultimate ability to program a "sea of gates" to do literally anything you want. There are several types that include one or more "hard-core" conventional processors (e.g., ARM) along with a large configurable "fabric" of gates, allowing you to experiment by creating customized co-processors that are optimized for specific applications.

\$\endgroup\$
  • \$\begingroup\$ I assume that the hardware-software interface is the instruction set architecture (ISA). The ISA is not a mapping between bits and real-world phenomena. It is a mapping between software instructions and abstract binary operations. \$\endgroup\$ – Rasmus B. Sorensen Jan 16 '16 at 21:21
  • \$\begingroup\$ @RasmusB.Sorensen: No, actually they're quite separate concepts. The instruction set architecture is completely abstract, and simply defines what software operations (binary data to binary data) can be performed by the CPU on a per-instruction basis. What that binary data means in the real world is the domain of the hardware-software interface. \$\endgroup\$ – Dave Tweed Jan 16 '16 at 21:26
  • \$\begingroup\$ Ok, to clarify my understanding, how would you define an add operation? Is this not implemented in hardware and defined in the ISA? \$\endgroup\$ – Rasmus B. Sorensen Jan 16 '16 at 21:31
  • \$\begingroup\$ @RasmusB.Sorensen: Of course it's implemented in hardware -- everything is. However, executing the add operation does not normally affect anything other than the internal state of the CPU. There are no external effects unless the ISA permits the result to be written to an I/O register of some sort that is part of the hardware-software interface. Still two separate issues, though. \$\endgroup\$ – Dave Tweed Jan 16 '16 at 22:47
  • \$\begingroup\$ Ok you answered most of the question. I saw this last night BEE. This is something i am talking about. Are there any other universities working on something like this? \$\endgroup\$ – abhishek tyagi Jan 17 '16 at 3:15

Not the answer you're looking for? Browse other questions tagged or ask your own question.