Verify Intel DDR5 wake-up functionality
Verify Intel DDR5 wake-up functionality
Linus has mentioned in his videos how boosting RAM speed enhances performance for AMD CPUs. Will the arrival of DDR5 guarantee a victory for AMD? I’ve been curious about this since Intel’s RAM speeds don’t impact frame rates. Whether Intel releases new chips or not, I think they’ll start designing CPUs based on RAM speed to keep up with AMD. Anyone have ideas on this? DDR5 is coming soon.
I discussed this with a tech expert who mentioned Intel has been leading for a long time due to their ability to pack more transistors onto a chip compared to AMD. He explained they invested heavily in R&D to maintain transistor density at 10nm and 7nm. I'm pleased we now have more competition, which should help keep Intel's prices fair. My thought is that AMD might shine during a new RAM era if Intel doesn't act. Using 3400mhz RAM versus 2133mhz can potentially double your frame rate at high resolutions with AMD. It's unclear when Intel will respond. We shouldn't overlook the M1 ARM chips, which are outperforming both AMD and Intel. The market dynamics are shifting in about 2 to 4 years. I'm not a fan of Apple or Intel; I enjoy seeing healthy competition!
Sure thing. The idea of a significant jump in FPS isn't realistic across the board. Performance gains are limited and vary by game. Higher RAM speeds help but aren't the main factor anymore. Intel has always had an edge with faster memory, and AMD is catching up. Things are evolving, but it's not a complete overhaul.
Intel's performance per instruction should rise significantly at the 7nm stage. Bypassing 10nm might be beneficial for chip development, with Alder Lake and 16-core designs showing potential. Since both AMD and Intel will utilize TSMC manufacturing facilities, we can expect closely matched IPC results. The real distinctions in speed will likely come from architectural and layout variations rather than raw numbers. Comparing 7nm and 5nm processes across AMD and Intel architectures could yield differences ranging from minimal to substantial—possibly as much as a 50% improvement. We should anticipate noticeable changes within the next five years. Looking back that far... computing has evolved dramatically. Today's systems are far more intricate than in 1999, and even more so than in 1979. When a $600 FX-55 is now worth just $40, it’s clear how fast technology has advanced!
In twenty years there will be a clear shift when physics restricts our capacity to miniaturize transistors. Although a university reached 2nm, an atom's average size is about 150 picometers or 0.150nm. Until we can control electrons effectively, further shrinking won't happen much soon. We'll transition to quantum computing and rely on cloud-based processing.