dom

其實我想講...

小鏡先係最萌...
untvb 發表於 10-10-2011 07:01 PM

呢個我不嬲都知

cise4832

都唔係唔 make sense...
証明左兩條 threads share resource 係好難, intel 真係技高一籌... ...
untvb 發表於 2011-10-10 19:27

又唔係咁講既
bulldozer畢竟係新architecture
再講...睇benchmark single thread似乎大鑊D

dom

挑娜星.....推土機8核同頻表現仲差過6核 1090T,點算????
車道士 發表於 10-10-2011 07:19 PM

其實正式計
Bulldozer 係 4C8T  (唔理佢話咩 Eight Core 啦 , 跟據同樣道理 Intel 都可以話佢粒 Core i7 係 8Core )
應該用 Phenom II X4 / Core i5 黎比

dom

回復  dom

oic....但聽講會賣2k港紙以上,咪....
車道士 發表於 10-10-2011 07:54 PM

Bulldozer = Epic Fail

cise4832

覺得好唔make sense, 有乜可能仲慢過上一代?
AltezzaXP 發表於 2011-10-10 19:23

I too, think it doesn't make sense.
They should have known the performance before taping out.

If I had a product that failed to deliver and just want to cover part of the loss,
I will just let it slip out to market quietly, for example dynamic switchable graphic for AMD, there isn't much marketing around it.

Consider that in the recent months:
1) We have comic FX.
2) Give-away contest hosted by AMD.
3) World record overclocking event.
4) 7Ghz stable SC2 demonstration.
5) Cray XT6 interlagos upgrade.
6) ScaleMP supports interlagos.
7) A lot of active AMD blogs covering details.

That's quite a lot of marketing and good news,
yet the chip failed to deliver?

真係似自殺多D

~蜜蜂~

piledriver如果追到/超過sandy bridge就得架啦,ivy bridge效能提升都唔大
2012 AMD都會出Piledriver ...
vcfirstben 發表於 2011-10-10 17:52

假設 ivy bridge IPC 一樣
基於製造工藝的進步，Intel 趾高時脈
AMD Piledriver 的 10% 一樣唔夠打

~蜜蜂~

覺得好唔make sense, 有乜可能仲慢過上一代?
AltezzaXP 發表於 2011-10-10 19:23

AMD 可能自己都高估左推土機