Board logo

標題: [業界消息] Intel話請Raja翻黎整獨立GPU [打印本頁]

作者: Aware    時間: 2017-11-9 07:33     標題: Intel話請Raja翻黎整獨立GPU

心諗,你得咩?
啲垃圾Graphic狂堆transistor 都係唔夠人打
作者: Henry    時間: 2017-11-9 11:04

Intel有既係Video processing同bandwidth efficiency
無既係3D果邊Driver support或者GPU架構
其實如果Driver好返既話應該同NV有得砌
作者: BlackBird    時間: 2017-11-9 17:13

Intel 有的是製程優勢
VEGA係差頻率同太食電
兩樣野都可以用製程屈返黎, 效率起碼好過HD graphic
作者: qcmadness    時間: 2017-11-9 20:23

https://www.anandtech.com/show/1 ... -as-chief-architect
作者: qcmadness    時間: 2017-11-9 20:23

引用:
原帖由 BlackBird 於 2017-11-9 17:13 發表
Intel 有的是製程優勢
VEGA係差頻率同太食電
兩樣野都可以用製程屈返黎, 效率起碼好過HD graphic
我覺得Intel fab utilization太低
作者: dom    時間: 2017-11-9 22:05

Intel 加油
作者: Henry    時間: 2017-11-10 00:40

引用:
原帖由 qcmadness 於 2017-11-9 20:23 發表

我覺得Intel fab utilization太低
岩曬,終於有貨出。
作者: tat0801    時間: 2017-11-10 08:48

Intel fab 整曬nand 同optane 好過啦
作者: BlackBird    時間: 2017-11-10 11:26

引用:
原帖由 qcmadness 於 2017-11-9 20:23 發表

我覺得Intel fab utilization太低
搵Intel 做Vega64
作者: BlackBird    時間: 2017-11-10 11:28

引用:
原帖由 tat0801 於 2017-11-10 08:48 發表
Intel fab 整曬nand 同optane 好過啦
intel 唔會用自己edge node fab做flash哩d無成本效益的野
作者: Aware    時間: 2017-11-10 13:36

引用:
原帖由 qcmadness 於 2017-11-9 20:23 發表

我覺得Intel fab utilization太低
佢想搵客做OEM都冇人睬佢
作者: qcmadness    時間: 2017-11-10 20:45

引用:
原帖由 Aware 於 2017-11-10 13:36 發表

佢想搵客做OEM都冇人睬佢
因為high performance係有代價的
作者: Henry    時間: 2017-11-10 22:22

引用:
原帖由 BlackBird 於 2017-11-10 11:26 發表

搵Intel 做Vega64
用Intel製程再OC下可能同NV下代Volta有得比
作者: qcmadness    時間: 2017-11-10 22:25

引用:
原帖由 Henry 於 2017-11-10 22:22 發表

用Intel製程再OC下可能同NV下代Volta有得比
就算Intel 14nm++都無用

GCN太廢
作者: tat0801    時間: 2017-11-10 23:01

引用:
原帖由 BlackBird 於 2017-11-10 11:28 發表

intel 唔會用自己edge node fab做flash哩d無成本效益的野
Optane 唔係用最新製程
最新當然係要留比CPU GPU 呢啲
佢啲廠唔會通通都係14nm, 舊底果啲如果拎黎量產optane 搶生果單都肥啦
當然,做唔做又另一回事
作者: Aware    時間: 2017-11-10 23:06

引用:
原帖由 tat0801 於 2017-11-10 23:01 發表

Optane 唔係用最新製程
最新當然係要留比CPU GPU 呢啲
佢啲廠唔會通通都係14nm, 舊底果啲如果拎黎量產optane 搶生果單都肥啦
當然,做唔做又另一回事
生果單水位極低,睇死佢唔想Mac變AMD先繼續賣U俾Apple。
作者: tat0801    時間: 2017-11-11 08:45

引用:
原帖由 Aware 於 2017-11-10 23:06 發表

生果單水位極低,睇死佢唔想Mac變AMD先繼續賣U俾Apple。
想拖mac 唔轉arm 住咋 sosad
作者: Aware    時間: 2017-11-11 10:10

引用:
原帖由 tat0801 於 2017-11-11 08:45 發表

想拖mac 唔轉arm 住咋 sosad
佢仲玩IPC增長=0既話生果一定飛
作者: tat0801    時間: 2017-11-11 10:27

引用:
原帖由 Aware 於 2017-11-11 10:10 發表

佢仲玩IPC增長=0既話生果一定飛
其實都忍夠啦
呢單Intel AMD 配
一來要應對NVidia
二來要應對Apple
格離台仲係度話生果只佔pc shipment 唔夠一成
無咁大影響力夾到Intel 做野
之但係生果呢一成,差唔多食曬啲premium market 呢又
A11X 配強化版 in house GPU
我其實好奇啲graphic patent 係邊面黎
AMD 定Nvidia 還是PowerVR
POLARIS embeds Intel Core Gen8
只可以頂到年零
A chip 都仲講緊3成複合增長(先唔好理堆左幾多core, transistors, die size)
作者: Henry    時間: 2017-11-11 11:14

引用:
原帖由 Aware 於 2017-11-11 10:10 發表

佢仲玩IPC增長=0既話生果一定飛
難怪iOS 11係Pad既介面同OSX開始有D似。
Intel新架構要2019-2020先出到來,趕唔趕得切?
作者: Aware    時間: 2017-11-11 13:16

引用:
原帖由 Henry 於 2017-11-11 11:14 發表

難怪iOS 11係Pad既介面同OSX開始有D似。
Intel新架構要2019-2020先出到來,趕唔趕得切?
佢咁玩法Apple轉會真係唔奇
作者: qcmadness    時間: 2017-11-11 14:01

引用:
原帖由 tat0801 於 2017-11-11 10:27 發表

其實都忍夠啦
呢單Intel AMD 配
一來要應對NVidia
二來要應對Apple
格離台仲係度話生果只佔pc shipment 唔夠一成
無咁大影響力夾到Intel 做野
之但係生果呢一成,差唔多食曬啲premium market 呢又
A11X 配強化版 in  ...
關Intel咩事, Apple賣US$10000, Intel都唔會收多d

Apple A11再上就會用勁多transistor, 而家已經大粒過Zen / Skylake
作者: tat0801    時間: 2017-11-11 15:48

引用:
原帖由 qcmadness 於 2017-11-11 14:01 發表

關Intel咩事, Apple賣US$10000, Intel都唔會收多d

Apple A11再上就會用勁多transistor, 而家已經大粒過Zen / Skylake
生果堆transistors 唔係第一日
個CPU 唔會再有好多增長
但GPU 就另一回事
作者: qcmadness    時間: 2017-11-11 15:49

引用:
原帖由 tat0801 於 2017-11-11 15:48 發表

生果堆transistors 唔係第一日
個CPU 唔會再有好多增長
但GPU 就另一回事
GPU勁食transistor
作者: tat0801    時間: 2017-11-11 16:19

引用:
原帖由 qcmadness 於 2017-11-11 15:49 發表

GPU勁食transistor
Tim Cook 講明黎緊GPU 會保持到增長
而家轉左in house
佢點堆都得
當然成本功耗平衡
等佢搞
呢粒A11現時未發揮曬出黎
我估到iOS 12 又會有啲野放出黎
作者: qcmadness    時間: 2017-11-11 21:31

http://www.xtremesystems.org/for ... wfull=1#post5262338
引用:
The last 'successful' card he worked on at (still) ATi was the R700 (Radeon HD 4XXX series). But back then his involvement was much smaller - he wasn't in charge of the whole design, only the software part of it. Note that this is the time when ATi's graphics drivers were particularly garbage. Then, he left to become chief graphics architect at Apple, and worked on the GPUs in iPhones 1, 3G, and 4. These weren't built from scratch, but rather, modified PowerVR designs. Note that at the time, iPhones were notorious for terrible, underperforming GPUs that were a royal pain in the ass to program for. He's also partially responsible (through his connections at Intel) for Apple's PowerPC -> x86 switch, although you can't really attribute the success to him; Intel did all the hard engineering work.

Anyway, he left before the iPhone 5 came out, with a much-improved GPU. Since his absence, iPhone GPUs became better and better (and are actually somewhat painless to program for now), and actually outperform Qualcomm's offerings. Of course, he went right back to AMD, where he (unsuccessfully) tried to spin RTG off, threatening to quit and go to Intel. He kept throwing tantrums until Lisa Su gave up and just gave him free reign to do whatever he wanted at RTG. He immediately put his newfound freedom to work by producing several hopelessly uncompetitive designs: starting with Hawaii (r9 29X), then following it up with the pointless Tahiti & Tonga (r9 280, 285X), and then the even less competitive Fiji (Fury). The 290 was actually better positioned in the market than Fury. This trend continues with his latest failure, Vega, which is in an even more awkward position than the Fury was. The card literally has no reason to exist.

作者: tat0801    時間: 2017-11-12 11:30

引用:
原帖由 qcmadness 於 2017-11-11 21:31 發表
http://www.xtremesystems.org/forums/showthread.php?295056-News-Intel-Hires-Raja-Koduri-to-Develop-Discrete-GPUs-This-Time-for-Real&p=5262338&viewfull=1#post5262338



Vega 叫可以同1070 打平手囉
慢對家成兩年先打平手
老黃fatfat 啦
但生果無選擇下都要用vega for iMac Pro
但另一方面用eGPU 去將個最易壞,最可以成日更新既既硬件外判出去
但啲driver 唔知點處理,涉及好多硬件加速既問題
作者: BlackBird    時間: 2017-11-14 09:06

引用:
原帖由 tat0801 於 2017-11-10 23:01 發表

Optane 唔係用最新製程
最新當然係要留比CPU GPU 呢啲
佢啲廠唔會通通都係14nm, 舊底果啲如果拎黎量產optane 搶生果單都肥啦
當然,做唔做又另一回事
Optane=20nm
作者: tat0801    時間: 2017-11-14 10:45

引用:
原帖由 BlackBird 於 2017-11-14 09:06 發表

Optane=20nm
Intel 家下啲閒置產能都唔知做咩
做chipset 又用唔曬
Optane 似乎良率終於得,開始有實物賣街
啲fab 盡其用又未必搵最多既
Intel 而家都樣樣沾沾手
作者: qcmadness    時間: 2017-11-14 21:09

引用:
原帖由 tat0801 於 2017-11-14 10:45 發表

Intel 家下啲閒置產能都唔知做咩
做chipset 又用唔曬
Optane 似乎良率終於得,開始有實物賣街
啲fab 盡其用又未必搵最多既
Intel 而家都樣樣沾沾手 ...
唔同類型既chips唔一定可以共用fab
作者: Aware    時間: 2017-11-15 09:30

引用:
原帖由 qcmadness 於 2017-11-14 21:09 發表

唔同類型既chips唔一定可以共用fab
啲製程又HP又LP又+/++/+++....,共用到先奇
作者: qcmadness    時間: 2017-11-15 21:26

引用:
原帖由 Aware 於 2017-11-15 09:30 發表

啲製程又HP又LP又+/++/+++....,共用到先奇
問題係application都影響好大
作者: qcmadness    時間: 2017-11-15 21:27

http://www.guru3d.com/news-story ... ll-development.html

大家明去左邊啦




歡迎光臨 HKSpot (https://bbs.hk-spot.com/) Powered by Discuz! 6.0 Lite