Why is the AMD FX series so hated?
Why is the AMD FX series so hated?
I really can't grasp why these chips were so disliked. I created a lot of FX setups, like my gaming rig using the FX-9590. I really liked it—it had air cooling, was pretty quiet, and performed well even when things got tough.
Its PERCORE speed was disappointing. Back then there were many games available, and forward remained single-threaded with a single dual-core thread, making it slower than an old overclocked Core2Duo PerCore. Multithreading was acceptable for its era, but single-threading was problematic. CPUs weren't utilizing full independent FPUs; they relied on shared FPU capabilities...and the rest of the details aside.
Beyond the poor single-core results, questions arose about whether the CPUs actually matched their claimed core numbers. In several parallel workloads, the FX 8150 underperformed against the Phenom X6 1090T. This situation becomes confusing when comparing a newer 8c/8t processor to an older 6c/6t chip. Ultimately, whether a Bulldozer module is considered two cores or one with simultaneous multithreading hinges on the specific tasks. If calculations demand floating-point math, the CPU appears to function with only half its intended cores. Although a 4c/8t model remained cutting-edge, FX was still ahead, which isn't entirely surprising but wasn't clearly communicated.
I was okay with them since they were very affordable and I even assembled a few systems myself. But they came with many downsides. At the moment, their power usage was poor, and the temperatures were extremely high (though for today's standards those temps were acceptable, but the maximum operating temperature on those chips was only around 70°C). In terms of single-threaded speed, they lagged behind the previous generation Phenom CPUs. If a newer CPU is faster than an older one, it usually means something went wrong. They also lasted a long time longer than expected. FX CPUs were available for about five years. They performed sluggishly when launched, and five years later they only got a refresh and a few other models like the 9590—essentially just pre-optimized versions of the 8350 to claim they weren't obsolete. Even AMD acknowledged FX was lacking, especially when they tried benchmarking a GPU (probably the Fury) on an Intel platform because they knew their CPUs would be the bottleneck. If it runs well for you, feel free to keep using it. It's essentially a platform only worth it if you can buy it almost for free. That was one of the main reasons back then you could get an 8320 for $110—at that price it was hard to compete unless you added something like a GTX 980 SLI, and it let you use a much better GPU than an i5 4690k would allow. You'd rather have the i5 all week if you needed real CPU performance.
FX was initially implemented on socket 939 with a focus on speed. Later, during the TLB errata for C2 Phenom arena chips, they relied on Phenom, which often succeeded in reaching 3ghz to 3.2ghz. These instances were rare, but they helped solidify the FX branding. The FX name became associated with slower performance, and eventually the FX identity was established.
Significant power and warmth, yet noticeably lagging behind competitors—around 40% slower.
They are certainly flawed yet possess some merits as well. FX tends to be odd in this regard, but that’s part of the charm. It can struggle with single-threaded tasks, though it performs adequately in multi-threaded scenarios. I’ll break it down simply: There are essentially two approaches to handling this situation. Picture a dragster speeding through a track—this FX chip is like the engine. It delivers maximum power, but if you can’t channel it efficiently, it wastes potential. The other option is to let it “hook” and succeed, achieving results without overtaxing resources. This is the core of FX: it has capability, yet its inefficiency often leads to frustration until it finally meets expectations.
For single-core setups, it feels sluggish and frustrating, whereas multi-threaded environments see better performance. The trade-off lies in power versus practicality. If you can push it to exceed 200 FPS in a game or test, it’s impressive—but only because the numbers matter more than real-world smoothness. Most people care about frame rates that look consistent, not just the theoretical max.
Higher MHz ratings are appealing, but human perception of quality stops at around 150 FPS before noticeable differences appear. In games, unless you’re tracking precise FPS values, the difference between 160 and 210 is negligible. What matters is achieving higher numbers, even if they don’t translate into tangible benefits.
Unlike Intel chips, FX lacks advanced features like CB/CBB, making it more accessible for enthusiasts who prioritize performance over optimization. This design choice helped it gain traction, especially when paired with affordable pricing and reliable builds. It’s a case where raw capability, combined with cost-effectiveness, allowed it to thrive before AMD stepped in.
For everyday gaming or daily use, I wouldn’t recommend it unless you’re into optimization. However, for enthusiasts looking to push hardware limits, it’s worth exploring. If you’re considering using it, ensure your system can handle the demands—otherwise, it might just sit idle.
A helpful tip: always inspect the file before extraction. Even if it looks clean, double-checking ensures a smooth installation. Once ready, create a shortcut and launch it directly to start testing.
I value all the feedback and details you shared. It intrigued me because during my gaming session on a gaming PC, these chips consistently outperformed the top I7 models we had built with similar parts. While they generated more heat than the Intel options, they were built to handle that. From a performance standpoint, what others observed was mostly modest gains in everyday use and noticeable improvements in gaming. I’ve come to understand benchmarks as limited in their value since they can be manipulated to show any result. Personally, these chips consistently exceeded benchmark scores. By the way, I’m an AMD enthusiast. Over many years using their products, I’ve always thought they offered good value for the price. Of course, my confidence has sometimes wavered, but overall it’s clear. I also want to note that overclocking isn’t necessary for most gamers today—CPUs are powerful enough that the difference is rarely perceptible. Still, for specific tasks, it can be useful. I don’t want others to think I’m the only one who sees this way, and I’m open to hearing different perspectives. As for my setup right now, I run a Ryzen 7 with 32GB RAM and a 1TB SSD, paired with an RTX 5500 graphics card. I was considering upgrading my old 9590 into a case with a slightly older GPU and selling it, mainly because I’m trying to cover rent after losing my job.