User Profile
ChinmoyDip2007
Copper Contributor
Joined 5 years ago
User Widgets
Recent Discussions
Should Microsoft use custom AMD processors on surface devices?
Microsoft is using off the shelf processors for their Surface devices from Intel and AMD. But the processors are not reaching the performance per watt leadership and are too expensive in compared to their performance efficiency. Many gaming consoles are using custom AMD processors which are cheaper and better at performance per watt.442Views0likes0CommentsMy idea of Unified Binaries
Unified Binaries is like one binary that with proper runtime or framework integrated into Windows, can work in any architecture (ARM/x86). It's quite like Java or Python or .NET core, where a runtime for a specific architecture can run a same universal code for windows. I gave two different types of Unified Binaries: 1.for older binaries, There will be a runtime created that can run that can run non native (x86 on ARM or ARM on x86) windows binaries without any modification into the actual software. 2.Newer binaries will be compiled from scratch in a way that proper frameworks on windows can easily run it on any architecture. For .NET binaries you already have what to do. Just try to create a framework for C++ binaries It's like Minecraft java, the underlying java part is the same but different launchers and C++ code for system communication. This method I assume can be designed in a way that the apps can think that they are running natively. Please also consider graphics2.2KViews0likes8CommentsShould arm based surface devices use nvidia processors instead?
This might sound wired. But when I saw the Nintendo switch using a custom tegra based processor that can run mid resource intensive games while giving about 5 hours of battery life out of just a 16 whr battery, it surprised me. I kept searching nvidia’s website and found out that the tegra x1 on shield to is very powerful yet efficient. Although 10 watts may seem a lot, but it’s 20nm compared to today’s 5nm. Nvidia markets them as mobile processors, and scaling it down to 5nm means a lot of efficiency for the same performance, which should even give some headroom for to increase the power by a little margin to get some more performance. I know that the processor has older cpu and gpu but it can be simply made up by using today’s new RTX and arm counterparts.460Views0likes0Comments
Groups
Recent Blog Articles
No content to show