Steamをインストール
ログイン
|
言語
简体中文(簡体字中国語)
繁體中文(繁体字中国語)
한국어 (韓国語)
ไทย (タイ語)
български (ブルガリア語)
Čeština(チェコ語)
Dansk (デンマーク語)
Deutsch (ドイツ語)
English (英語)
Español - España (スペイン語 - スペイン)
Español - Latinoamérica (スペイン語 - ラテンアメリカ)
Ελληνικά (ギリシャ語)
Français (フランス語)
Italiano (イタリア語)
Bahasa Indonesia(インドネシア語)
Magyar(ハンガリー語)
Nederlands (オランダ語)
Norsk (ノルウェー語)
Polski (ポーランド語)
Português(ポルトガル語-ポルトガル)
Português - Brasil (ポルトガル語 - ブラジル)
Română(ルーマニア語)
Русский (ロシア語)
Suomi (フィンランド語)
Svenska (スウェーデン語)
Türkçe (トルコ語)
Tiếng Việt (ベトナム語)
Українська (ウクライナ語)
翻訳の問題を報告
(Intel side / AMD side)
8P + 4E cores < 12 Cores
20 Threads < 24 Threads
Base clock 3.6GHz < 3.7GHz
Boost 5.0GHz < 4.8GHz
Cache 25 MB < Cache 64 MB
10 NM < 7 NM
5.0 PCIe > 4.0 PCIe
TDP 125 (190) < TDP 105 (142)
By raw data the 5900X is superior work wise, but I believe after optimisation the 12700k will be the choice for gamers again.
Beside that, you require a new Mainboard (LGA 1700) and DDR5... which will cost a lot for the average person for this small upgrade (When you sit at the current generation).
I also don't trust Intels "Single Core" performance cheating at any point anymore, the boost will drop after a while which cheats the entire result.
Edit: If I had to choose, I would buy the 5900X for many reasons when I go for a small workstation, the 12700K for gaming.
Talking about higher upgrade possibilities, as we know already the 12900K can push up to whopping 400w, which never can't be hold up (Not even a custom loop) and it downclocks itself instantly. It's obvious that Intel tries to break records, by doing non sense features and cheating their results. For the casual gamer how ever something like 400w is something you will NEVER reach. But the fact that Intel also tried to cover their 10nm by calling it "7nm" is also a poverty certificate. Don't get me wrong, Intel is fine, but most people just don't see how dodgy intel push out their results.
5900X/5950X have proved to not be so great in a long list of Games.
edit
i expect good performance due to the fact the core for system thread will be boosted . But this change of thread from one core to anodther will be a hindrance to the perfomance . Power draw will be not perfect though .
Also what is the difference between Performance and High Efficiency cores?
Arent they the same cores with different clock speed and voltage usage.
Alder Lake may have high core count, but the E-Cores are just garbage, so anything that needs more than 6/12 to 8/16 is gonna be better off on a 10900K or 5900X.
AMD doesn't use a monolithic design so you can't compare it to Intel, and while I think Intel won't be either anymore starting with Alder Lake, it's still going to be different as Intel is also implementing a hybrid design. In traditional terms, the 12700K would be like an octo core CPU with Hyper-threading, but it's also going to have an additional lesser powered Arm cores.
Intel has a far different architecture. As of late, we can summarize it as "a bit slower than AMD's but Intel's chips clock higher to somewhat negate it" and how something performs will depend on how the software is written, but generally the latest generation chips from both are very close.
The upcoming Alder Lake has no official metrics but the leaks suggest it will take the top place. So what this means is an 12700K will likely turn in better performance "per core" (going to be harder to refer to this given there's going to be hybrid designs becoming the norm), but your CPU will still probably edge it out, if slightly, in highly parallelized tasks. Keep in mind that a lot of the results seen recently by Intel compared them against AMD on Windows 11 which was suffering from two different performance issues, and Intel has stated they will run the comparisons again, and while Intel likely does have a faster chip, keep this in mind that they released these unfavorable initial results knowing people would see them and take them at face value, because as they say, first impressions are everything (you see this with GPUs too where AMD sometimes has more performance increases through drivers and may shrink or even a gap, but recommendations will still go to nVidia because it's impossible for everyone to keep updated on every last change).
But anyway, this new advancements is... pretty much the norm (and hopefully it stays as such because CPUs were becoming concerning there for a while). A CPU newer is faster. I'm confused by your comparison if you got a CPU recently, looking for something for around four years, and not even a year or official launch later you're looking at comparisons to what you have? What are you going to do when AMD's refresh with 3D cache, and then later, Zen 4 comes? What about Intel's generation after Alder Lake then?
Developments that happen after you buy something are actually great, because it means when you upgrade, you get a nice jump. It was a sad feeling getting a 2500K and watching the feet dragging for years and generations afterwards. There's a reason I stayed with it so long.
Keep in mind the Alder Lake CPUs will command a price jump over the existing generation prices (similar to how Zen 3 raised them over Zen 2, further suggesting Intel won't be releasing a chip slower than Zen 3, which is highly unlikely anyway given the small gap between them with released stuff), and DDR5 will be expensive and likely in its slower iterations at first. While Alder Lake may have a bigger than usual (for Intel) jump (finally!), it makes more sense to pay all of these premiums if you're on something more older, not anything recent, let alone high end recent.
Even then though, Alder Lake's architecture is different enough, coupled with DDR5 and PCIe 5.0 support, that I don't think you can really compare it to what AMD is offering or even Intel's previous generation. I think the single threaded performance of Alder Lake will be good enough compared to AMD Zen 3 that it will beat it in game performance and overall core for core speed. But the fact the 5950X has 16 'performance' cores and 32 threads will make it better for traditional workloads and multi threaded applications that prefer more cores and take advantage of multi threaded CPUs, compared to the 12900K and the 12700K.
Hence, since the 5900X has 24 threads and the 12700K has 20 threads, I believe the 5900X will be better in those scenarios as well. But for gaming, that prefer better single threaded performance, over how many cores a CPU has, Alder Lake will most likely perform exceptional. Thus, why the 12900K and 12700K, while priced less than the 5950X and the 5900X, respectively, will most likely perform better than those CPUs overall in gaming.
Just my opinion.
Also just food for thought. All these benchmarks are done at 1080p with an RTX 3090. If they were to do the benchmarks in 4K lets say, the results would all be the same. At 1440p and 4K the difference in performance of CPUs really begins to equal out and the GPU is the botlleneck. Thus if you are just buying a CPU for gaming and you play at 4k, you could buy a cheaper CPU or even a CPU that is a generation or two old and still get the same exact performance as the most expensive current gen CPU.
What I think is funny is that when AMD shows gaming benchmarks for their CPUs, they do it at 1080p coupled with an Nvidia GPU. Really? I mean AMD GPUs are good enough that they could use one of their own GPUs to show the benchmarks. I understand doing it at 1080 to show the differences in the CPUs but come on AMD, have some confidence in your own hardware. LMAO.
Just buy a 10900k, it'll be far cheaper and as fast or faster than both in gaming and as close in production that it won't make a huge impact.
Beyond that, wait for reliable benchmarks, reading that otof guys response tells me he doesn't really understand stuff, especially the cooling side of things.
I cant see a reason to hate this that much also my phone (Redmi T8) which have 4cores @ 2.0GHz and 4cores @ 1.8GHz, and for around 10 sec can get to deep 24 but in the first sec can get to deep 18, yes this is not the best but this is enough for normal use and helpfull to some of the chess players, but in competitive you will use threadripper or 4 X (Ryzen 9 5900X)