The upcoming Core i9-10900K desktop processor – the flagship of the Comet Lake-S family – will outperform the current flagship Core i9-9900K by up to 30%. In any case, Intel cites such information in its internal documents, one of the pages of which became public by the efforts of one of the users of the Chinese Weibo.
Intel compared the future flagship of the mass segment with the current one in a series of synthetic toasts using benchmarks such as SYSmark, SPECrate, XPRT and Cinebench. Each of them, except the last, Intel constantly uses in its internal tests. Unfortunately, specific test results are not presented, and only an increase relative to the Core i9-9900K is indicated, so it is difficult to compare the Core i9-10900K with any other processors.
The increase of 30% mentioned at the beginning is observed in one of the tests of the SPECrate 2006 benchmark. Also, a very high increase of 26% is observed in the popular Cinebench R15 test. Actually, there is nothing surprising in such changes in the results in these tests: the Core i9-10900K has 10 cores and 20 threads, while the Core i9-9900K has only 8 cores and 16 threads. And the frequencies of the first of the new products should be higher.
And, in fact, due to the increase in frequencies, an increase is observed in other benchmarks, because from an architectural point of view, the novelty completely repeats its predecessor. In the XPRT test suite, using a single core, the new Core i9-10900K was only 2–4% faster than the Core i9-9900K. Also, a slight increase of 4-8% is observed in single-core tests of SPECrate 2006, and only 10% increased performance in SYSmark 2014 SE.
It is important to note that the Core i9-10900K processor was tested with standard parameters, including TDP, equal to 125 watts. Also in the document for this chip, the power of 250 W is mentioned, which is necessary for him to achieve turbofrequency for all cores. Note that for the Core i9-9900K this figure is 210 watts.
If you notice an error, select it with the mouse and press CTRL + ENTER.