qcmadness

引用:

原帖由 dom 於 2013-10-29 10:43 發表
http://www.expreview.com/29101.html
快左少少 咁多多

真係先算啦

p.s. 如果係真, i3 / i5有難

qcmadness

引用:

原帖由 cheungmanhoi 於 2013-10-30 10:06 發表

如果佢唔2個platform 合一無咩興趣

SteamRoller得FM2+

qcmadness

引用:

原帖由 cheungmanhoi 於 2013-10-30 10:15 發表

咁 pure cpu果d fx都係落fm2+?
定係唔會有 fx

no FX

because I suspect AMD will go 20nm in 2015

qcmadness

引用:

原帖由 Puff 於 2013-11-12 15:13 發表
OT: Does anybody have a clue of what speed the DDR4 majority will be at its launch?
2133? 2400? 2667?

expect 2133

qcmadness

引用:

原帖由 cheungmanhoi 於 2013-11-13 10:50 發表
amd.....

睇來佢真係無plan b
真係見步行步咁
好彩d engine人爆到個勁idea就有得番身者
唔好彩就

Even Intel does not have a plan B

qcmadness

Cinebench R11.5 Single-Thread
AMD FX-8350: 1.10
AMD FX-6300: 1.07
AMD Phenom II X6 1100T BE: 1.11

WinRAR 3.8 compression (dual-thread)
AMD FX-8350: 74.9s
AMD FX-6300: 78.3s
AMD Phenom II X6 1100T BE: 89s

World of Warcraft FRAPS (dual-thread)
AMD FX-8350: 91.5
AMD FX-6300: 79.4
AMD Phenom II X6 1100T BE: 77.8

引用:

1yrs ago AMD finally pick up their sorry a$$ and fix some of their Faildozer and make Faildozer 1.5 ( Piledriver )
it  is getting closer to "Phenon II" (Star) class single thread performance , but not there yet

This is certainly not true

你唔好同我講同頻呀, 你講既係single-thread performance而唔係IPC呀

qcmadness

引用:

原帖由 dom 於 2013-12-7 16:14 發表

Pls show Sandy Bridge to Haswell result

Cinebench 11.5 - Single Thread

i7 2600K: 1.52
i7 3770K: 1.66
i7 4770K: 1.78

WinRAR 3.8 compression (dual-thread)
i7 4770K: 53s
i7 3770K: 57.7s
i7 2600K: 59.6s

qcmadness

引用:

原帖由 dom 於 2013-12-7 22:11 發表
see how epic fail faildozer is

But more core wor

AMD prioritize more core over higher single-thread performance.

qcmadness

引用:

原帖由 dom 於 2014-1-1 23:55 發表


買得起 i5 價落粒U 既 都買得起獨顯啦

獨顯一樣要錢

qcmadness

Slightly larger than my estimation (220-240mm^2)

http://www.xbitlabs.com/news/cpu ... ess_Technology.html

引用:

With Kaveri, AMD chose to use GlobalFoundries’ 28SHP (28nm super high performance) process technology that was designed with various types of chips in mind. Improved transistor density and thinner elements of 28nm process tech allowed the company to integrate whopping 2.41 billion of transistors into a 245mm2 Kaveri die, which is 85% higher transistor count compared to Trinity/Richland design (246mm2, 1.303 billion transistors, 32nm SOI). Now the company has four Steamroller cores, hUMA [heterogeneous unified memory architecture] memory controller, Radeon graphics processing unit with 512 stream processors (AMD calls them eight GPU compute units) and a lot of various special-purpose hardware inside its flagship APU.

Still SOI-based.

qcmadness

引用:

原帖由 Puff 於 2014-1-11 00:01 發表

Bulk. Page 17

Steamroller: An x86-64 Core Implemented in 28nm Bulk CMOS 3:15 PM
K. Gillespie1, H. R. Fair III1, C. Henrion2, R. Jotwani3, S. Kosonocky2,
R. S. Orefice1, D. A. Priore1, J. White1, K.  ...

但係SHP @ GF真係指SOI

唔通係hybrid?