Intel Core i7-4770K Haswell Processor Performance Review Posted

Have you been wondering how the upcoming Intel Fourth Generation Core i ‘Haswell’ processors will perform on the LGA1150 platform? If you said yes, your wait is over. Tom’s Hardware has posted up performance numbers in a ‘preview’ article on the upcoming flagship Intel Core i7-4770K processor with Intel HD 4600 graphics.  We don’t even have our hands on a processor yet, so we went and read the review ourselves. THe Intel Core i7-4770K has a base clock of 3.5GHz and boosts up to 3.9GHz thanks to Turbo Boost 2.0. The processor has 8MB of L3 cache, uses Intel’s GT2 HD 4600 GPU at 1250MHz and has a TDP of just 84 Watts. The review on the pre-production Intel Core i7-4770K showed performance gains of 7 to 13% over the current Intel Core i7-3770K ‘Ivy Bridge’ processor when it comes to current threaded workloads. It also showed 12 to 26% when gaming at a resolution of 1920×1080 with the integrated Intel HD Graphics 4600 iGPU. No power numbers were shown in the review for some reason, but we expect that the Intel Core i7-4770K will use less power as well. The AVX2 instruction set on Intel’s Core i7-4770K, and all Haswell processors, looks pretty damn good. Take a look at the Sandra Processor Multimedia benchmark below!

Intel Core i7-4770K Processor Memory Bandwidth

Intel Core i7-4770K Processor Gaming Performance

The integer test employs the AVX2 instruction set on Intel’s Haswell-based Core i7-4770K, while the Ivy and Sandy Bridge-based processors are limited to AVX support. As you see in the red bar, the task is finished much faster on Haswell. It’s close, but not quite 2x. Floating-point performance also enjoys a significant speed-up from Intel’s first implementation of FMA3 (AMD’s Bulldozer design supports FMA4, while Piledriver supports both the three- and four-operand versions). The Ivy and Sandy Bridge-based processors utilize AVX-optimized code paths, falling quite a bit behind at the same clock rate. Why do doubles seem to speed up so much more than floats on Haswell? The code path for FMA3 is actually latency-bound. If we were to turn off FMA3 support altogether in Sandra’s options and used AVX, the scaling proves similar.


Comments are closed.