It all started with node size, that you seem to hold as the only valid performance per watt indicator,
You keep reguritating things I never said or asserted. I don't understand how you can keep making that oversight ... again and again.
LeDoyen wrote:which downright made me know you know nothing of the manufacturing side of things.
Which has what to do with any of this? I'm taking design, layout and other things. Putting things together has nothing to do with how Intel didn't design its paging correctly in hardware.
LeDoyen wrote:When i mention "front end fabs".. apparently you do not know that term either? well..
Being at fabless firms, I know exact what they are.
LeDoyen wrote:If you worked in design, well, i work in manufacturing.. so you know things i dont know, and i know things you don't know.
I've worked in design, clearly more than yourself. That's why you don't understand what SoC is. You think it means whatever goes on-die, but it's not.
LeDoyen wrote:Also, nobody asked for your curriculum vitae, but since you ask about mine, i have worked for semiconductor tools OEM for quite some time now, namely Novellus (before the acquisition by LAM) and Applied, on PVD, dry etch and CVD. I didn't care to mention because it brings nothing to the arguments you bring out which are totally irrelevant to the original question from this dude about his next gaming laptop (and since you like that t erm too, i have NDA up the ass both with my company and the various customers i work at).
You just come to flatter your ego, taking the post nowhere, and not helping at all in his choice.
You aren't familiar with semiconductor design and architecture. That is clear. That's why you keep saying what SoC is, in stark contrast to even how Intel defines it, just like everyone else.
I only brought up my NDA because it's with Intel over its exact issues -- that have existed since 2007 -- in its paging, TLB and other units.
They haven't fixed it in 12 years. It's why Intel keep shaving to ship firmware and kernel fixes to stop the hemorrhaging. They finally admitted it publicly in 2017, and there is no end in sight.
That's why many people are looking to AMD ... for everything, including gaming on Windows. The fixes are heavily impacting, while AMD doesn't have the same problem in x86-64. And even ARMv8, and some in ARMv7, has these protections. They are a no brainer. Intel's failure to design a real x86-64 chip with these protections is going to be its downfall.
Just like I warned them back in late 2007 when Nehelem was in our engineered trading systems on Wall Street. The Linux kernel patches, under embargo, have the same. It's a real, big, continuing, 12+ year mess.
I have nothing against Intel personally, just their designs ... as well as their FPUs (which is another story). There's a reason why, for precision, engineers and scientists have long preferred AMD FPUs. And now we're seeing a serious issue with Intel not doing things in hardware. It's why graphics are so unstable on x86-64 platforms, especially with PCIe.
It didn't have to be this way. But Intel didn't care. And now ... they have to, especially as AMD is going to have a fabrication lead over Intel thanx to TSMC.NOTE: My current gaming system is actually Intel, but the next one won't be.