Nvidia's CPU would likely challenge both AMD and Intel, depending on advancements in performance and efficiency.
Nvidia's CPU would likely challenge both AMD and Intel, depending on advancements in performance and efficiency.
It's not unusual, under most workloads it's below 10%. You don't measure IPC during overclocking—it's about clock speed relative to time. My 2700x reached 4GHz, and my 6600k OC is a different story. Which one you're comparing depends on the specific issue: Green or AMD?
The biggest issue was the fact we're using 14nm+ (mostly). FX performed okay on cheaper builds. Honestly, it’s not too bad unless you push it too far. In fact, I’ve seen some cases where it outperforms the 98SE. RT on consumer cards is still early-stage tech. The main problem is limited adoption so far, and even lower-tier cards aren’t strong enough—just like with something new.
With a decent margin in the OC you could still surpass a Zen+ chip. The clock speed was mainly between Haswell and Broadwell. These points come from the part you quoted yourself, where you were puzzled about my understanding. I emphasized core-for-core because the key factor behind Zen/Zen+’s strong sales was having more cores at comparable prices. The architecture itself lagged behind Intel’s, which they compensated for by increasing core volume. That wasn’t ideal. My 2700X was quite underwhelming but it could still reach nearly that on PBO with a 4.1 manual; it couldn’t hit 4.2 easily, though many chips did. Only achieving four out of a Skylake chip felt odd, most would perform better. Even my 6950X (Broadwell-E, noticeably slower clock) could reach 4.2 on all cores, which is decent but not outstanding. Most Broadwells hit 4.2 or higher, and some even reached 4.6 or more, though not as reliably as Haswell.
FX Bulldozer projects often didn’t save money compared to budget i3 or i5 setups. Once you hit a solid FX processor like the 6 or 8 core and a reliable board that handles high power usage, spending the same or more than an i5 usually meant you could beat it in games. I upgraded from an FX 8320 at 4.4ghz to a 6600K at stock speeds, and FPS improved while frame drops vanished. The FX chips were terrible, and AMD faced lawsuits for it. I haven’t used ME myself, but the complaints are clear—if your system crashes under the OS, it’s not worth it. Still, RT gets a lot of attention, so some might think a GPU is justified, but Nvidia should focus on real performance gains rather than just marketing.
It seems like there were eight cores initially, but only half of them shared resources, making it feel like fewer than eight in practice.
Likely fair, I suppose. It might have been luck, but it looks like I ran into some FX deals during the Ivy/Haswell time. And honestly, I was mostly dealing with the rougher stuff—like using hardware with terrible drivers and trying to make real-mode DOS work. Cheers for not getting through this text! FX GPUs were really frustrating.
They utilize Nvidia tegra, their arm series (now owned by them), which controls the switch.
FX would have been a great offer when it was on sale. I recall the FX 9000 CPU came out really cheap compared to its original price, and one of those 9590 models would be enjoyable to use. Still, that makes sense. The drivers and DOS mode were problematic for me, which likely explains why most users stuck with the 98SE or 2000.