Intel’s Haswell Core i5 4670K Gets Reviewed In China

/ 3 years ago


Intel’s Haswell launch is getting closer and closer, and in the past we have already seen an i7 4770K review from Tom’s Hardware, but today we get a look at the Core i5 K series Haswell processor – the i5 4670K. An extensive review has popped up in China at PC Online and the reviewers have really gone all out to run literally every benchmark. I have picked out those benchmarks which I think you will most want to see but if you want to see everything then be sure to check out the link to PC Online.

First up we have the classic test of CPU performance, Cinebench. Straight away we can see the increase over the i5 3570K is marginal at best. According to my calculations it grabs an extra 2.5% in performance.


Next up we have the 3DMark benchmark, running the CPU scores, and we can see approximately a 6% boost over the i5 3570K.


In the 3DMark 11 CPU test we see less performance gain over the previous generation i5 3570K to the tune of just 3.6%.


This graph details power consumption st Prime 95 load. We can see a marked improvement over the i5 3570K with a reduction of 25 watts, a power reduction of 17%.


Using all the benchmarks to summarise performance the CPU component of the i5 4670K is approximately 6% better than the i5 3570K.


Comparing the GPU component it can be seen that the i5 4670K is 23% better than the i5 3570K, this is mainly due to the integrated GPU redesign from HD 4000 to HD 4600.


Finally here is a power consumption summary graph using a wide variety of load scenarios. On average the i5 4670K consumes about 21% less power than the i5 3570K.


So there you have it some really extensive review testing on the Haswell i5 4670K. Don’t forget to check out PC Online if you want to see all the graphs, but make sure you have Chinese auto translate enabled!

What are your thoughts on the power consumption and performance of the i5 4670K?

Images courtesy of PC Online

Topics: , , , , ,

  • Wayne

    Doesn’t seem like it has any serious legs on the 3570K. Much lighter on gas though. I’ll test this chip myself then I’ll know for sure.

  • d6bmg

    Well, it don’t seem to be too good.
    Or might be because of a bad chip perhaps?

    • bad chip would only affect overclockability and power consumption, not stock performance. all the tests in that review are run at stock.

      • d6bmg

        Still, I was expecting improvement in performance, even at the stock.

        • there is. 6%.

          • d6bmg

            That’s next to nothing. Isn’t it?

          • d6bmg

            I think Intel is intentionally blocking their development due to lack of competition in the market..

          • Klimax

            No. It is getting much harder to extract more performance out without increasing power consumption. We’d need either new algos for basic operations or new machine type. (It’s still Turing machine)

            Free perf lunch was over for some time. (The only reason Arm has those increases is, that their starting position was so much worse. They’ll hit wall eventually.)