TL/DR : Will current dual-core CPUs (say, the i3-6100) be able to run mainstream desktop applications (not games) for another 10 years? When will the programmers who design this kind of applications (browsers, etc.) really start to make use of more than two physical cores?
My story and more specific questions:
I'm currently building a new computer, which will replace the Athlon 64 with 2 GB of RAM which I built in early 2005. I don't play any games beyond chess, and the heaviest task my computer ever sees is running a virtual machine to use some Windows Xp-only applications (I use Linux). The main reason why I'm building a new box now is because there is a new piece of highly specialized software that I need which will only run on Windows 7 and up, and I can't run that in a virtual machine on my current PC (at least not with an acceptable responsiveness). Since this software was written for the special purpose of analysing a +10 GB database of medieval Arabic texts, no alternative for it exists. It moreover runs smoothly only when it has access to 5 GBs of RAM, which probably has something to do with the fact that it was not written by a professional programmer (?).
I want my new build to be good for another 10 years or longer, assuming the heaviest task it will be used for is running Windows 10 based data analysis software through a virtual machine. I like to imagine that the utter disappearance of non-3D versions of mainstream websites will be what will force me to move to a new platform in the unforeseeable future, but since I'm not exactly an expert in the futurist department, I would like to ask your opinions on this. I already made up my mind on buying a Z170 chip-set (MSI Z170A PC MATE), 16 GBs of DDR4 RAM (2400MHZ/CL14), and a Samsung SSD 850 EVO 250GB. Although I'd prefer not to have to upgrade anything before the whole thing becomes obsolete, I chose this configuration because upgrades (NVMe SSD, up to 64GB DDR4 RAM with higher clock-speeds and lower latencies) at least remain possible.
The only performance related part I find hard to decide on is the CPU. Surely today's desktop applications rely mostly on single-thread performance and do not need more than two physical cores, so I naturally gravitate towards the I3-6100. The question that is important to me, however, is if this will also hold true for the next 10 years? I find quite a lot of discussion of I3s vs I5s or the (relative) importance of multiple cores on the web, but they are all limited to questions of gaming performance, in the next 4-5 years. Of course it is exceedingly hard to try to see further than that into the future, but some people might at least have an idea of how the programming of desktop applications will evolve, and more specifically if and when desktop software designers (browsers, word processing, etc.) will be starting to program in a way that capitalizes on the availability of four physical cores?
I already overspent a little on the RAM and Z170 chip-set, but given the fact that I want to run (relatively light) data analysis software in a virtualised Windows 10 environment, the 16 GBs and having the option of higher clock-speed DDR4 in the future (the H170 is limited to 2144MHZ) seemed to make this necessary. The question remains now if I should also overspend on the CPU and get a I5-6500 instead of the I3-6100? Their single-thread performance appears to be exactly the same, although it remains unclear if the benchmarks [urlExt=http://www.cpu-monkey.com/en/compare_cpu-intel_core_i5_6500-524-vs-intel_core_i3_6100-556]cpu monkey[/urlExt] cites are making use of the I5's Turbo-boost option or not. In multi-core benchmarks however the I3-6100 is only at 70-80% of the I5-6500. What I need to know, then, is how important multi-core performance will be in the next 10 years?
This is my particular situation, but I think all those who are planning to do a long-term build might benefit from your informed opinions on this question.
My story and more specific questions:
I'm currently building a new computer, which will replace the Athlon 64 with 2 GB of RAM which I built in early 2005. I don't play any games beyond chess, and the heaviest task my computer ever sees is running a virtual machine to use some Windows Xp-only applications (I use Linux). The main reason why I'm building a new box now is because there is a new piece of highly specialized software that I need which will only run on Windows 7 and up, and I can't run that in a virtual machine on my current PC (at least not with an acceptable responsiveness). Since this software was written for the special purpose of analysing a +10 GB database of medieval Arabic texts, no alternative for it exists. It moreover runs smoothly only when it has access to 5 GBs of RAM, which probably has something to do with the fact that it was not written by a professional programmer (?).
I want my new build to be good for another 10 years or longer, assuming the heaviest task it will be used for is running Windows 10 based data analysis software through a virtual machine. I like to imagine that the utter disappearance of non-3D versions of mainstream websites will be what will force me to move to a new platform in the unforeseeable future, but since I'm not exactly an expert in the futurist department, I would like to ask your opinions on this. I already made up my mind on buying a Z170 chip-set (MSI Z170A PC MATE), 16 GBs of DDR4 RAM (2400MHZ/CL14), and a Samsung SSD 850 EVO 250GB. Although I'd prefer not to have to upgrade anything before the whole thing becomes obsolete, I chose this configuration because upgrades (NVMe SSD, up to 64GB DDR4 RAM with higher clock-speeds and lower latencies) at least remain possible.
The only performance related part I find hard to decide on is the CPU. Surely today's desktop applications rely mostly on single-thread performance and do not need more than two physical cores, so I naturally gravitate towards the I3-6100. The question that is important to me, however, is if this will also hold true for the next 10 years? I find quite a lot of discussion of I3s vs I5s or the (relative) importance of multiple cores on the web, but they are all limited to questions of gaming performance, in the next 4-5 years. Of course it is exceedingly hard to try to see further than that into the future, but some people might at least have an idea of how the programming of desktop applications will evolve, and more specifically if and when desktop software designers (browsers, word processing, etc.) will be starting to program in a way that capitalizes on the availability of four physical cores?
I already overspent a little on the RAM and Z170 chip-set, but given the fact that I want to run (relatively light) data analysis software in a virtualised Windows 10 environment, the 16 GBs and having the option of higher clock-speed DDR4 in the future (the H170 is limited to 2144MHZ) seemed to make this necessary. The question remains now if I should also overspend on the CPU and get a I5-6500 instead of the I3-6100? Their single-thread performance appears to be exactly the same, although it remains unclear if the benchmarks [urlExt=http://www.cpu-monkey.com/en/compare_cpu-intel_core_i5_6500-524-vs-intel_core_i3_6100-556]cpu monkey[/urlExt] cites are making use of the I5's Turbo-boost option or not. In multi-core benchmarks however the I3-6100 is only at 70-80% of the I5-6500. What I need to know, then, is how important multi-core performance will be in the next 10 years?
This is my particular situation, but I think all those who are planning to do a long-term build might benefit from your informed opinions on this question.