Intel Pentium G620T Pro Reviews

Despite boasting Intel's latest Sandy Bridge microarchitecture, the conclusions here are pretty much unchanged from the past. Intel maintains an advantage when it goes up against an AMD chip with the same number of cores, almost regardless of clock speed. In the case of the Pentium G620, even an Athlon II X2 265 with a 27% clock speed advantage can't outperform the Sandy Bridge based CPU. If you're choosing between chips with the same number of cores at the same price, Intel wins. The decision tilts in AMD's favor if you start comparing to the Athlon II X3. In heavily threaded workloads, the Athlon II X3's third core helps put it ahead of the entire SNB Pentium lineup. If you're building a machine to do offline 3D rendering, multithreaded compiling or video transcoding then AMD continues to deliver the best performance per dollar. It's in the lighter, less threaded workloads that the Pentium pulls ahead. If you're building more of a general use system (email, web browsing, typical office applications and even discrete GPU gaming), the Pentium will likely deliver better performance thanks to its ILP advantages. What AMD has offered these past couple of years is an affordable way to get great multithreaded performance for those applications that need it. Unfortunately the entire Sandy Bridge Pentium lineup is clock locked. Without turbo modes there's no support for overclocking at all. While these new Pentiums would have normally been great for enthusiasts looking to overclock, Intel has ensured that anyone looking to get more performance for free at the low end will have to shop AMD. Unfortunately Intel's advantage in single/lightly threaded performance is big enough that a clock speed advantage alone is generally not enough to make up for it (see G620 vs. Athlon II X2 265 comparison). It's sad that it has come to this. I was hoping we'd see more K-series SKUs at the low end but it seems like those will only be for the enthusiasts at the high end.

Sandy Bridge microarchitecture is amazingly diverse. We have already admired the high performance of processors based on it many time, and today we saw clearly that it is also a perfect fit for products to be used in quiet, compact and energy-efficient systems. However, the T-series processors discussed in our today’s article revealed a number of unique peculiarities, which aren’t mentioned anywhere in the official specs but can change your perception of these products. Take a look at their power consumption. Although I-series processors have half the TDP of the regular CPUs, it doesn’t at all mean that in reality they only consume half the power. Firstly, energy-efficient processors get very close to their TDP threshold, while regular CPUs can often consume less power and dissipate less heat than their TDP implies. Therefore, the actual difference in practical power consumption between T and non-T processors of the same class is never that significant. Secondly, we can really see a serious difference in practical power consumption between energy-efficient and regular processors only in a limited number of usage scenarios, while most of the time their power appetites are very close. In fact, T-series processors show their true energy-efficiency only under heavy multi-threaded load. In idle mode, under single-threaded load or during high GPU utilization, T-series processors do not offer any noticeable advantages in the power aspect. All this means that it doesn’t make sense to use energy-efficient Sandy Bridge modifications just for the sake of saving some power. Since in real life the processors are usually loaded sporadically, T-series won’t bring you power bill down a lot.