[ad_1]
As expected, Intel officially announced this week its new 9th generation Core processors, as well as the Z390 chipset. We have written everything here with a few brief comments on what to expect given the embargo on the benchmarks and the full reviews are not up until October 19th.
Although, technically, we are not bound by this because we have not signed a confidentiality agreement to obtain our sample, we will wait for a number of reasons, by professionalism and by respect for other readers who are going run days after days. in their review of one day.
However, when PCGamesN released the first Core i9-9900K results today, we were a little surprised. The title said "The Core i9 9900K Intel is up to 50% faster than the Ryzen 7 2700X AMD in the games", seemed to me wrong, but I read …
"Intel has officially announced its new Core i9 9900K processor, calling it the" best gaming processor in the world. "It's not just a marketing crisis either … well, not quite. Principled Technologies to conduct a comparative analysis of its latest 9th generation chips and their competitors participate in 19 of the most popular PC games. "
Thus, Intel may publish its own "tests" made suspiciously by a third party ten days before revisions, while it is forbidden for reviewers to rebut the claims because of the NDA. First bad sign.
By scrolling through PCGamesN, this is what is said when we look at Intel's performance tests …
"But the real interest of all this is that Intel can present the 9900K as the best gaming processor compared to the AMD competition, and it seems to have excelled. On some games, such as Civ 6 and PUBG, the delta performance is not necessarily great, but most of the time, you get a 30-50 percent higher rate than the 9900K compared to the 2700X. "
At first, many results seemed very suspicious, after spending countless hours analyzing the performance of the 2700X and 8700K, I have a good idea of how they compare in a wide range of titles and these results seemed very disappointing. After spotting some wacky results, my next thought was why PCGamesN publishes this misleading data and why do not they destroy the paid benchmark report? Do not they just know better?
On the Principled Technologies website, you can find the full report that describes how they were tested and the material used. The official memory speeds have been used, which is not a big deal, although they have managed to cripple Ryzen, or at least expose the weaknesses.
Ryzen does not work very well with full memory DIMMs. Two modules are optimal. However, timings are also important and they used Corsair Vengeance memory without loading the extreme memory profile or the XMP parameter. They simply set the memory frequency to 2933 and left the ridiculously loose default memory timings. These low times ensure compatibility and allow the systems to boot, but then you must enable the memory profile. It is misleading to make landmarks without performing this crucial step.
Nevertheless, it would be almost right if they had done the same thing for Intel, but they did not do it. For all Intel platforms, they first set the memory on XMP, then adjusted the frequency manually, giving Intel a significant performance advantage, especially for gaming.
The next step in their handling of the results was to test only in 1080p with a GTX 1080 Ti using quality presets located at one or two levels of the maximum level. In many cases, this simulates the kind of performance we see when we test in 720p with higher quality presets. Of course, we also test 1080p and 1440p to give readers a complete picture.
One of the worst results chosen by PCGamesN was that of Ashes of the Singularity. Ignore the 9900K at the moment because I can not show you these results, but let's focus on 8700K and 2700X instead. Here, the 8700K was 29% faster than the 2700X, it's a margin far superior to what I thought I would find. Principled Technologies uses the built-in reference, processor-centric reference with the game running in DirectX 12 mode with high quality presetting.
So I installed two Corsair Vengeance Pro DDR4-3200 modules, loaded the XMP profile on the AMD and Intel platforms and ran the tests with exactly the same parameters. I've also re-run the tests with XMP memory timings but at the official memory speeds for each CPU. I will only show three of the results of the game. I did not feel the need to spend more time on this test after a few hours, it was enough to make a point.
Here are the results of Ashes of the Singularity, let's talk about it. Compared to the results generated by Principled Technologies and using the stock memory, my 2700X was 18% faster, which is already a shocking result.
The 8700K, on the other hand, was 4% slower, which means that, compared to the 2700X, it was 4% faster and up to 9% faster with a memory with a higher number of screen points. This has nothing to do with the 29% performance advantage that Intel enjoyed in the paid test.
Then I looked at Far Cry 5 and there the change was not so extreme for the 2700X, it was 3% faster in my test and 10% faster with a specification memory of 3,200. However, the 8700K was even slower, 7% slower with the 2666 memory used by Principled Technologies. Even with 3200 memories, I could not match their result.
This means that even though they claim that the 8700K processor is 26% faster in Far Cry 5, it actually corresponds to 14% or 12% with 3200 memories, which represents a clear victory for Intel, but not as much. extreme that the landmarks of Principled Technologies believe.
The latest results I reviewed were about Assassin's Creed: Origins. They again used the built-in reference with the third highest quality preselection, 1080p, with a GTX 1080 Ti. According to Principled Technologies, the 8700K model was 36% faster than in reality, it is closer to 8% with memory speeds in stock and 10% with overclocked memory.
It is therefore quite obvious that the results of Principled Technologies are a waste load and that no one should reproduce them. Of course, Intel's main goal was to emphasize how great the Core i9-9900K is, and for the moment, I can not show you the real results for this processor. Needless to say though, it will not be 50% faster than the 2700X in games.
The 9900K will be faster than the 2700X for games, this is not doubtful, but it will cost twice as much once you have taken into account the price of the motherboard. On average, we found that when we used memory optimized for the 2700X and 8700K, the Intel processor was about 9% faster in 1080p with a GTX 1080 Ti. Realistically, we expect the 9900K to be a little faster than the 8700K, at least in the games, as most titles are not about to come out of the 6-core / 12-thread processor.
The fact that Intel controls the report itself is not too much of a problem, and the Principled Technologies report is very transparent because it clearly indicates how it tested the games and set up the hardware. The results and test methods are heavily biased, but they have not tried to hide their questionable methods. You can consult the specifications and find all the details. It's still doubtful, but it's a paid report, so it's a bit expected.
Nvidia's deceptive benchmarks for launching the GeForce 20 series were rather bad, but it's on another level and I sincerely hope that these results will not be published anywhere else with the intention of promoting the 9900K. We understand, Intel is under pressure, but in the end, this kind of approach makes society a bad service.
We expect more information on this in the coming days and that goes without saying, please wait for our independent review of i9-9900K and the reviews of other trusted media that will arrive next weekend. . Tomorrow, we also hope to follow up on our thoughts on the new 28-core Xeon CPUs and Intel Core X processors in the unlocked series: see where they oppose Threadripper.
Source link
