DerBauer's content covering CPU silicon lottery and reviews is essential for anyone starting to understand CPUs.
DerBauer's content covering CPU silicon lottery and reviews is essential for anyone starting to understand CPUs.
He eliminated the power cap, allowing the CPU to operate at maximum capacity, which is crucial for clarity. Yet, claiming every reviewer follows the same approach overlooks important differences (a bit exaggerated). I’d find it useful if derBauer explained his decision to disable the limit and why, and I’d like to hear his reasoning. Your feedback was well received!
You might tolerate it, but when two items from the same model differ by more than 9%, they likely aren't identical. This raises doubts about the consistency of your production methods if you overlook that variation. Tolerances are inherent everywhere. You won’t escape them entirely. Still, there should be a sensible idea of how the product will behave.
I can observe a slight performance gap, though it remains within my awareness range. My graphics card isn't very powerful for overclocking, and I typically see only about 8-10% improvement. Still, I can confirm whether the system is running with the overclock active or disabled in Control, a title I've played extensively and understand its behavior on my machine. The extra frames per second and smoother operation from the overclock are just perceptible, even if modest. Yes, it seems a player could detect differences between the highest and lowest tested 7600 der8auer models. But only in games and tasks that matter to that person. This isn't a difference that stands out instantly. I think the reason der8auer chose this approach was to highlight the variation among CPUs, which is logical, though I wish he had been more explicit about it. Also, could you recommend a review site that focuses on PBO-enabled units? It's unusual to see such detailed testing, and every review I've read with PBO always includes the CPU at its standard setting. I'm sure someone else has done this, but I usually check around five or six reviewers for each product. None of them test in this manner. Remember, the observed variance here is only noticeable when using PBO and power limits are unlocked—it's not a 9% difference at normal settings. At regular prices, these chips should perform very similarly under all-load conditions.
Over the past two decades I've observed similar trends among fellow enthusiasts. When multiple people push identical components—some achieving higher speeds or lower voltages than others—it's a familiar pattern.
So if there's no overclocking, there probably wouldn't be any noticable difference, right? I think hard-core OCers and pro gamers would be much more likely to notice things than average people. I mean, I can tell the difference between my iGPU and my GC, lol, but I'm not sure I would notice the variance he showed. LOL! Agreed. As a reviewer myself, I find it hard to remember to say everything. I don't script stuff out for food reviews, so I sometimes forget to mention, say, the texture of a food. For some reviews, I'll take notes to reference, especially if there's a lot to talk about, but I'd have to sit down and create basic templates for all the different types of things I review...Not sure how many that would be, but it's not an insignificant number. lol! Not justifying his omission by any means, though. He's a professional. I'd imagine that he sometimes forgets that he has viewers who don't know things that are common sense for him. Again, I'm not apologizing for him. I think his video has produced some interesting revelations. I don't have a wide enough range of reviewers that I watch regularly to say and, of late, I've been very focused on coolers and fans because that's my main focus for IT products for my channel. At least for now. I doubt I'll ever get to the level of expertise necessary to do CPU, GPU and mobo reviews but sometimes I surprise myself. Sometimes. Agreed. That makes sense.
Problem is its not just PBO given the chips by default clock as high as they can for the thermals and power available, the latter often the motherboard manufacturers differ in their defaults rather than using the CPU manufacturers settings. Were seeing more and more than your motherboard choice impacts RAM timings causing the same problem with CPUs performing worse on certain boards with certain RAM. There is no real solution to this though other than educating people to realise this is a thing and there's not much you can do about it. Building a PC was never the exact science people tend to think it is.
It's a science, yet its complexity requires exceptional talent and tools to simplify it effectively.
It would be great if CPU, RAM, and motherboard makers adopted common XMP/EXPO sub-timing settings that work automatically without manual setup. I know these settings would need to be more relaxed compared to today’s top boards, but the solution would be for motherboards to offer an automatic optimization feature at their current levels. Think of a built-in RAM "Auto OC" for sub-timings. Right now, Gigabyte is focusing heavily on AM5 systems with the most precise timings, which seems to be drawing more complaints over time due to instability and boot problems as they approach limits. I imagine two possible outcomes: either Gigabyte leverages its lead to capture market share among hobbyists, prompting competitors to push harder, or they might retreat under pressure from increased support demands. Neither path is ideal for users—choosing between stable stock settings would reduce inconsistency, and reviewers would have a clearer benchmark to compare results.
These are just some thoughts I’ve had about this. In practice, running your system close to stock levels isn’t a big deal unless your chip has any defects. It’s similar with RAM and ICs—different sets can vary in timing precision, speed capabilities, or stability. This explains why you get different “kit models” depending on the PC specs and stock timings. The same applies to GPUs. I’m referring to silicon, which all chips are built from, and it has consistent properties regardless of its use. When you increase clock speeds and voltage, the differences between parts become noticeable, affecting performance. Remember, Roman does what I do, but his focus isn’t always for everyday users. Still, gamers care about this a lot. If you’re aiming for better performance, you need a chip that balances speed with stability. Other aspects like cooling solutions and case airflow also play a big role—don’t just buy the cheapest cooler; make sure it delivers adequate airflow and volume (CFM). This is especially true for air-cooled systems, as they depend heavily on existing airflow inside the case.