cheungmanhoi

引用:

原帖由 Puff 於 2013-11-6 17:09 發表
CineBench R15, 1 (for 1C)/2 (for 2C) rendering thread, same bucket of background applications
@ 3.67 Ghz
2M2C: ~153 cb
1M2C: ~131 cb
1M1C: ~76 cb (scaled down from 78 cb @ 3.71-3.84 Ghz)

yield ~70% o ...

睇完










都係唔明

Puff

引用:

原帖由 cheungmanhoi 於 2013-11-6 17:27 發表

睇完都係唔明

簡單 d 講可以睇淡依家又傳得好旺 gea 30% IPC boost 啦，然後同 intel 比差一截
要 1M2C 係 cinebench 先打到人地一粒 core... 無 SMT。


擺到明就係 design flaw


[ 本帖最後由 Puff 於 2013-11-6 17:37 編輯 ]

cheungmanhoi

引用:

原帖由 Puff 於 2013-11-6 17:30 發表

簡單 d 講可以睇淡依家又傳得好旺 gea 30% IPC boost 啦，然後同 intel 比差一截
要 1M2C 係 cinebench 先打到人地一粒 core... 無 SMT。


擺到明就係 design flaw
...

原地踏步


你叫d x6 1100t 用家情何以堪

45nm 去到28nm 咩進步都無咁

Puff

AMD claims Kaveri achieved 31 fps in BF4 at 1080p w/o Mantle, while GT 630 coupled to an Intel 4770K sucks.
4 SR cores, 8 CUs, TrueAudio confirmed, nothing really special at all
Desktop first, mobile next. The prototype machine seems nice anyway.

Just recycled stuff like C++ AMP and Java HSA enablement, and well, finally a real-time demo (this time N-body problem) running on Kaveri 35W chip, illustrating the speed-up of GPU acceleration versus single- and multi-core. Also the vaporware Media SDK is (eventually?) announced for release.








Last but not least, no roadmaps. Maybe we will see those things in the closing keynote.


[ 本帖最後由 Puff 於 2013-11-12 09:42 編輯 ]

dom

http://www.amd.com/us/press-rele ... ces-2013nov011.aspx
咦, 塞埋個 TrueAudio 入去咁好玩

dom

引用:

原帖由 Puff 於 2013/11/6 15:14 發表

average multi core performance
A10-5800K ~= i3-2100 < i3-3220  

That's why Faildozer is EPIC FAIL

dom

引用:

原帖由 Puff 於 2013/11/6 15:15 發表

依家先知 i5 mobile 無 QC

Mobile i5 = Desktop Core i3 Low Watt/ULV

dom

引用:

原帖由 Puff 於 2013/11/6 17:09 發表
CineBench R15, 1 (for 1C)/2 (for 2C) rendering thread, same bucket of background applications
@ 3.67 Ghz
2M2C: ~153 cb
1M2C: ~131 cb
1M1C: ~76 cb (scaled down from 78 cb @ 3.71-3.84 Ghz)

yield ~70% o ...

Haswell 1C VS 2M2C .....SR so suck ..

dom

引用:

原帖由 Puff 於 2013/11/12 09:30 發表
AMD claims Kaveri achieved 31 fps in BF4 at 1080p w/o Mantle, while GT 630 coupled to an Intel 4770K sucks.
4 SR cores, 8 CUs, TrueAudio confirmed, nothing really special at all
Desktop first, mobile  ...

Why not chase on Mobile first .......stupid AMD .....Desktop market no longer have large growth and DIY is dying breed......
Looks like AMD is signing their own dead warrant ...

dom

之前 rumor 傳情人節launch , 提早左一個月

cheungmanhoi

引用:

原帖由 dom 於 2013-11-12 10:47 發表


Why not chase on Mobile first .......stupid AMD .....Desktop market no longer have large growth and DIY is dying breed......
Looks like AMD is signing their own dead warrant ...

可能係 tune 個w數果度慢左
mobile 對watt 用電 熱量 性能比desktop 敏感好多...

dom

引用:

原帖由 cheungmanhoi 於 2013/11/12 11:38 發表

可能係 tune 個w數果度慢左
mobile 對watt 用電 熱量 性能比desktop 敏感好多...

咁 AMD 真係可以等死

Puff

引用:

原帖由 dom 於 2013-11-12 10:47 發表


Why not chase on Mobile first .......stupid AMD .....Desktop market no longer have large growth and DIY is dying breed......
Looks like AMD is signing their own dead warrant ...

Because you can launch PIBs whenever you want, while mobile release is bounded to OEM schedules. They are releasing machines for the Easter timeframe.

It is reasonable, though it sucks. Many thanks to GIrlFriend.



[ 本帖最後由 Puff 於 2013-11-12 15:10 編輯 ]

Puff

引用:

原帖由 dom 於 2013-11-12 10:45 發表


Haswell 1C VS 2M2C .....SR so suck ..

This is piledriver. But anyway, it doesn't change so much I guess.

Puff

OT: Does anybody have a clue of what speed the DDR4 majority will be at its launch?
2133? 2400? 2667?