Disabling CPU cores might not greatly boost your maximum overclock.
Disabling CPU cores might not greatly boost your maximum overclock.
Hey there, overclocking folks, I was just curious if turning off CPU cores really boosts my maximum overclock for top single-core performance. I’m familiar with how to push the limits and manage temperatures, but I’m wondering why disabling cores might help. Since I’m using an FX 6300 and single-core stats matter a lot for games, I figured switching to a 43xx lineup could improve that. Please share your thoughts, thanks! Also, my target is around 500MHz, right?
[UPDATE 24.12.] It looks like it’s not worth the effort now. After checking a similar discussion, it didn’t seem worthwhile. I also attempted it myself—even with two cores enabled, I only managed a 400MHz increase. So, no real gain. I think I’ll stick with boosting the NB bus speed and raising the FSB again.
The single-core performance at 4.7GHz was similar, maybe even lower than the 4.3GHz version with NB OC single-core results. Everyone knows multi-core performance drops when only two cores are active...
Sadly, turning off a CPU core won't actually boost performance for single-core tasks. This should be done with Fx-based AMD CPUs. Improving single-core CPU performance mainly depends on the number of IPCs the core can handle. IPC stands for Instructions Per Clock Cycle, and more instructions executed in each cycle mean better single-core processing. Overclocking slightly raises the frequency, allowing the CPU to complete more instructions per second in less time. However, this is influenced by the CPU's design and how it processes data, which also affects its single-core IPC rating. Intel CPUs are generally superior at handling data due to their architecture.
Haha, it really depends on how you're using it. If there are no background programs running, it might improve performance a bit, but once you start multitasking, you'll miss out on the extra cores. I'm not even sure it's worth it for gaming CPUs with HT capability, let alone AMD systems. Just focus on getting a better cooler—an Evo 212+ is solid and costs around $20-30, fitting most sockets. If memory matters, you can easily boost the FX6300 to 4.7 on air cooling.
Sadly, turning off the CPU core does not actually boost performance for single-core tasks. This approach works better with Fx-based AMD CPUs. Improving single-core CPU performance mainly depends on the number of IPCs the core can handle. IPC stands for Instructions Per Clock Cycle, and it reflects how many instructions are processed within one clock cycle. The higher the IPC, the more efficient the CPU is at handling single-core operations.
Increasing the frequency to a certain point allows the CPU to execute more instructions per second in less time. However, this depends on the CPU's internal construction and its efficiency in processing data. This also influences its single-core IPC rating.
Intel CPUs excel at processing data due to their design and efficient execution of instructions. This is why Intel-based CPUs generally outperform AMD FX CPUs in single-core applications and tasks.
With the new Rizen CPU, you might have heard that its single-core IPC is expected to be 40% higher or faster compared to AMD's new CPUs. This improvement comes from the new layout and architecture used in the Rizen design, which aims to achieve a greater overall IPC than current FX-based CPUs on the market.
Since CPU architecture cannot be altered, redesigning the CPU to enhance data handling and execution paths is necessary for achieving higher IPC. This includes optimizing data paths, cache size, and processing stages.
Using some unusual water cooling and refrigeration, I achieved my old FX8320 at 5.2ghz across all cores at 1.66V. The improvement was minimal, less than 15% compared to its stock 3.5ghz when benchmarks favored all eight cores, even with a 48.5% boost in clock speeds. The single-core gains were essentially within the error range. Performance and clock speeds didn’t improve together. Ultimately, everything is limited by the cache, so pushing speeds only makes the bottleneck worse.