游戏脑力 wrote:
最新收到消息是alderlake qs版會在這個月流出
「這個月」的意思是今天？大核5.3小核3.9
嘩，真爽，他算的R20得分剛好和你一樣，還真的打了5950X。
以往的爆料多少有個照片或者截屏，這回一個沒頭沒尾的賬號的一段沒頭沒尾的文字，就有人要寫長文了（然後再憋回去）。每到此時我就不得不多説兩句了：
游戏脑力 wrote:
另外這些分數 真的沒有太多可討論的空間.
什麽都具體參數都沒有的跑分實在代表不了什麽
厲害了
進來一看.原來這麽多人為這種流出測試分數在高潮
我有點驚訝到
真的不知道在高潮什麽
上reddit 看一下, 看不懂英文?
我看得懂哦
it's not even a rumor, just numbers writen in a notepad. Pathetic and desperated.
（甚至都不算流言，只是記事本上寫的幾串數字。真是可悲又絕望）（大概是A粉，不作爲參考）
Intel will not stop until they reach 6ghz at 300W!
（不到6GHz 300W不算完！）
The fact that most people believe this is true based on zero proof whatsoever is very funny to me. If this was an AMD leak people would be incredible skeptical, but because it's Intel somehow I have to believe some random chinese dude has secret info about Intel with his source being trust me bro.
（許多人基於零證據而直接相信這件事讓我感到很好笑。如果這是AMD的泄露，人們會很懷疑，但是既然是Intel，不知怎地，我就直接隨便相信一個中國老兄有Intel的内部信源了。）
——It's not even a leak. Someone took an actual leaked score from an early ES and another leak of max. single core clocks of newer QS and then multiplied the old score with the new max. clocks to estimate the multi core performance level.
——（甚至都不算泄露。某人直接拿ES樣品泄露的數字和新QS的單核最大值乘起來就得到新的多核心表現了）（聽起來像是01上的有的人會做的事）
Gracemont cores are powerful（Gracemont 很强啊）
My job is in big part to optimize software and make it run on as many cores as possible. And from experience I can tell that performance doesn't increase linearly with number of cores or frequency due to various bottlenecks.
So until there are actual results instead of "computations", I am quite skeptical.
Ryzen 5800x scores 6k points but the 5950x scores 10400. The improvement is not linear.
（我的工作是對軟體最佳化使之能夠運行在盡可能多的核心上。就我的經驗來看跑分并不隨著核心數量或時脈的增加而綫性增長，我對這些結果表示很懷疑，Ryzen 5800X跑分是6000而5950x只有10400，增長不是綫性）
——The improvement is not linear here either. The more cores you have, the bigger the improvement 😀
Edit. Those estimates are totally wrong.
2 big cores: +2100 points
4 small cores: +1800 points
but the 6+4 core processor only scores 7400.
If you make up numbers at least make them realistic. Btw I think 9300 for the 12900K is not far off.
（在這裏增長也不是綫性的，核心越多，好像還增長越大了。
（更正：估算的數字全錯……
（就算你要編數字，起碼也編的像樣點。當然我覺得9300分對12900K是差不多的）
Beats a 5950X in multi threading. ADL-S looks to be pretty impressive.
（多核心上打敗5950X，ADL-S很强啊。）
——No way it beats it with less threads. （無可能打敗。畢竟執行緒都要少。）
——Yeah because it's not. Just read the articles and that score is just a speculation on the leaker's part and read how it needs windows 11, which is not officially released. （你説的沒錯，因爲沒有打敗。讀一下原文吧，那個分數是盧出來的）
11900KB boosts to 5.3 on older 10nm so why shouldn't the 12900K boost to 5.3 on newer 10nm?
（11900KB都能達到5.3了，爲啥覺得12900K用新的10nm達不到5.3？）
——5.3GHz actually sounds a dash high to me given all the IPC gains expected. I'm don't disbelieve it but... 2% more clock speed doesn't matter when you have something like 20% more IPC than RKL.
（5.3對我來説聽起來有點太高了，倒不是不信，而是有20%的IPC進步的話多那2%的時脈卻是無關緊要的）
——Because it would be more expensive to bin 5.3 for a more widescale product than a niche one and not that necessary if you have +25% IPC from Golden Cove. If it is actually the case then Intel are footing that bill in order to gain more market share.
（對於一般產品來説衝到5.3會更貴，而在25%的IPC提升下没有必要衝那麽高。如果這個事情屬實的話，Intel將爲此買單以換取市場份額。）
——i9s are not widescale products. i9s are enthusiast (read: niche) products By clocking higher, 12900K can beat 5900X. Intel can produce many more 12900Ks than AMD can produce 5900X (And producing at scale is cheaper, so some extra costs due to higher binning requirements can be offset by producing at scale) Result: 12900K is more powerful than 5900X and in larger volume than 5900X, while not being significantly more expensive, if intel allows higher frequencies.
—— （i9不是一般產品，i9是小衆發燒產品。更高的時脈使12900K可以擊敗5900X，而Intel能生產比5900X更多的12900K，所以分攤下來成本並不會很高。）
——Let's wait and see -- Sandy Bridge brought a significant IPC uplift over Westmere and also clocked higher at the same voltage. Both on 32nm process, and in this case 10nm ESF is a definite efficiency improvement over 10 SF.
——（等等看吧。Sandy Bridge就有大幅度的IPC和時脈雙提升。）
The final sample is posibility get 5,5GHz （成品會達到5.5 GHz）
——Please don't start. Rocket Lake was supposed to hit 5.7GHz. Does it? （求別説。Rocket Lake還説要到5.7呢。）
——Can it happen? Sure. Is it realistic? No. It is going to be a win if the top SKUs can hit 5.3GHz consistently. Keep in mind this is a leak by some random person on a forum. The whole story may be fake.
——（可能發生嗎？可能。現實嗎？不現實。如果頂級SKU能夠持續達到5.3GHz，那將是一場勝利，但是謹記，現在這個消息只是一個論壇上一個路人甲講出來的，整個故事可能是假的。）
——Frequency scaling doesn't work like that lmao。U see this is a qualification sample which is very close to the retail version so its almost impossible that this thing will boost to 5.5 ghz but there is still a quarter or half a quarter atleats until this thing launches so yea maybe a slight chance but I doubt it will change bec its a (QS) not (ES)
——（你看，這是一個非常接近最終販售版的QS，所以這個東西幾乎不可能提升到 5.5 GHz，但是正式推出之前還有四分之一或半個季度的時間，所以可能有一點機會，但我懷疑它會有什麽改變，因為它是QS而不是ES）
