In an unexpected turn of events, a Hong Kong-based media outlet (XFastest) has leaked benchmarks of Intel’s upcoming Core i9-10900, before the chip has even been announced. The leaked tests include Cinebench R15, R20 and the CPU-Z scores. In this post, we compare them to the existing Intel 9th Gen and AMD’s Ryzen 3000 parts and see how they fare. Before we start, keep in mind that this is an Engineering Sample (ES), so the final scores will be slightly higher due to improved core clocks.
The engineering sample has all the same features of the final chip, except lower core clocks. The base core clock is a paltry 2.5GHz while the boost varies on the workload, usually topping out below 4.5GHz The final sample will have a boost frequency just south of 5GHz, so we can expect 15-20% higher performance than these tests.
10th Gen Intel Core i9-10900 ES Benchmarks
The Cinebench R15 score is rather dismal. Despite having the highest thread count of the lot, it manages to beat only the Core i7-9700K. While the single-core score can be explained on the basis of lower clock speeds, the multi-core score is a bit troubling.
In Cinebench R20, the 10th Gen Core i9-10900 once again fails to impress. Other than the octa-core i7-9700K, it fails to beat any of the other CPUs despite having the highest core (and thread) count.
Having a look at CPU-Z stats, the low scores can be explained on the basis of abysmally low operating clocks. In the Cinebench multi-core tests, the Core i9-10900 is limited to just 3GHz. That’s quite low. Even the 12 core Ryzen 9 3900X manages to run well over 4GHz under load in CB R20.
The CPU-Z scores are slightly better. Although the Core i9-10900 still fails to beat the Ryzen 7 or the Core i9-9900K, the deltas are much smaller here. You can tell that at launch, thanks to higher boost clocks, it should be able to edge past all these chips. Furthermore, in CPU-Z SC the 10900 manages to boost as much as 4.4GHz (vs 4.3GHz in Cinebench SC and 3GHz in MC). This means that in Cinebench, the 10th Gen CPU is running into a power bottleneck of sorts. As the temps are already quite sound, there’s no other plausible explanation.
There are a few conclusions that can be drawn from this little activity. Firstly, don’t expect any significant gains in multi-threaded workloads from Intel’s 10th Gen desktop CPUs. The base design is the same and considering that other than the top-end i9 part, the rest of the chips just get hyperthreading as an add-on, the performance won’t be much higher than the 9th Gen lineup. Secondly, as far as the single-threaded performance is concerned, yes, the higher boost clocks will allow for some decent gains, but what will that cost? How much will the 10 core i9-10900K draw under load and what kind of cooling will you need to keep it from exploding? For the answers to these questions, we’ll have to wait a few more months.