w*******g
发帖数: 140 | 1 【 以下文字转载自 Biology 讨论区 】
发信人: websurfgg (websurf), 信区: Biology
标 题: 得向学生物的致敬，看了wired 2007的十大发明有感
发信站: BBS 未名空间站 (Sun Jan 6 14:15:00 2008)
2007的10伟大发现有好几个都是生物的干细胞，牛！以前不觉得，但是看到intel 的
45nm的工艺才第十大发现(俺只知道这个很牛)，对比一下，生物还是很牛的。致敬 |
|
c*u
发帖数: 916 | 2 Intel 45nm 的工艺不能算“发现”吧，应当算发明。科学发现和工艺发明不一样的。
不过现在很多干细胞研究也是工艺问题。 |
|
i******g
发帖数: 3 | 3 急需可用的 BSIMSOI model card for 45nm or 65nm?
如果谁能提供一个，本人不胜感激！ |
|
k**f
发帖数: 11 | 4 They even made a 410GHz push-push oscillator this year using 45nm CMOS. |
|
a*******i
发帖数: 11664 | 5 我是ee烂校，排名30，40开外了，嘿嘿。
我说的那个人是git出来的，的确是名师名校。
cmos industry还能做今年，但是感觉越往后咱们这种device engineer或者process
engineer就越来越不重要
了。你从现在的趋势也能看出来，大家都要搞多核，横向扩展，当然有很多原因，一个
重要因素就是size往下
shrink不动了，从65nm到45nm还不难，但是到32nm就很难了，到22nm或者更小就非常难
了，而且这个东西
是有极限的。我感觉push到10几纳米就了不起了，以后的就是看怎么做connection，发
挥多核的优势。到那时候
就靠ic design, circuit level的了。真正还需要多少device/process engineering都
很难说。在si industry扎根越
深，越难往外跳啊。
不过这都我猜测，谁知道今后10年会是怎么样。
如果想读ph.d，还是安心继续吧。如果你都3，4年纪了，再换方向太不值了。
来的 |
|
R*******N
发帖数: 7494 | 6 45nm漏电流太大，不实用。65nm就很好了，现在CPU主要是拼结构了 |
|
l*******1
发帖数: 866 | 7
^^^^^^^^^^true, but not as severe as a show-stoper.
otherwise intel/AMD wouldn't push hard into
45nm lines. Also there are ways to reduce
leakage current. |
|
a*******i
发帖数: 11664 | 8 散热和scaling的问题。
现在65nm很成熟了，45nm的东西都陆陆续续出来的。正在开发的是32nm。22nm已经开始
研究了。再往下谁也不知道能做到什么程度。
从performance来讲，single core 2G的cpu > 1G Dual core的cpu > single core 1G
的cpu。类似的4 Core比Dual core要好，6 core会比 4 core好。但是继续往上就很难
说了，主要取决于芯片的设计和软件的支持。很有可能20 core会比16 core差。 |
|
c**********h
发帖数: 14 | 9 不知道Fab68的发展前景如何。 过几年45nm也要成为落后技术了，how would Fab68
change its infrastructure to keep up? 总感觉好像是Intel纯粹为了削减工资的开
支才把厂建到中国，也不知道能学到什么东西。 |
|
a******n
发帖数: 23 | 10 其实这个deep submicron NRE问题带来的fabless死期很早就有人在dac和eet上预测了
。现在一个45nm的mask就要烧掉几百万美金。不知道为什么还有人前仆后继进去。
以后不死的fabless，多是broadcom之类的寡头，或者是拿了老旧成熟工艺做超低成本
产品的，比如玩具芯片之类。 |
|
h*******a
发帖数: 184 | 11 seriously?有了20nm25那么就不用45nm，90nm等以上的吗？ |
|
|
|
ET
发帖数: 10701 | 14 啥technology node?
现在的65nm 45nm flicker noise corner 上MHz也完全现实的。
input referred noise的唯一物理意义是可比性，无法可测。
都是通过测量output noise density, 折算回去。 在simulation中，
也不要设定noise souce, 因为transistor本身就是个noise source, 要measure的就是
ouput noise density..
complain
source |
|
ET
发帖数: 10701 | 15 昨晚simulted a 45nm nmosfet using predictive spice model.
attached here. the spectral density also included a 10K resistor's noise
performance tho.
the flicker noise corner frequency is in the range of ~Mhz.. |
|
ET
发帖数: 10701 | 16 90nm: 1/90n
45nm : 1/45n
工具比我的强大呀.. 我只能眼睛看了.. |
|
e***7
发帖数: 118 | 17 1. What will determine the speed of a transistor?
2. When we talk about 45nm, 32nm node, what physical parameter of a CMOS are
we talking about? The width?
3. Data rate of an ossiliscope is given, how do you calculate the resolution?
4. N type and P type, which one is stronger?
可能有的题描述的不是很准确，但还是希望大家给点意见，感谢感谢。 |
|
p********e
发帖数: 6030 | 18 1. What will determine the speed of a transistor?
electron mobility, mean free path, channel/gate length
2. When we talk about 45nm, 32nm node, what physical parameter of a CMOS are
we talking about? The width?
the gate length
3. Data rate of an ossiliscope is given, how do you calculate the resolution?
4. N type and P type, which one is stronger?
what do you mean by stronger? |
|
j******e
发帖数: 526 | 19 1. What will determine the speed of a transistor?
temperature,doping,thickness
2. When we talk about 45nm, 32nm node, what physical parameter of a CMOS are
we talking about? The width?
minimum length
3. Data rate of an ossiliscope is given, how do you calculate the resolution?
resolution = bit rate / clock freq /#channel
4. N type and P type, which one is stronger?
stronger? i don't understand clearly,
may be N type |
|
y********g
发帖数: 81 | 20 1.Current and load cap. Its current is a
function of length, width, bias, etc.
2. 45nm, 32nm is half pitch dimension. This
is not transistor physical gate length nor
effect channel length. Refer to device
physics for detail.
3. 1/rate? how do you define resolution?
4. NMOS is stronger than PMOS in term of
larger current under the same condition.
NMOS conducts electron current whereas PMOS
conducts hole current. Whichever carrier it
is, current is proportional to mobility.
mobility is inverse pro... 阅读全帖 |
|
l********g
发帖数: 68 | 21 帮朋友转到版上,感兴趣的请直接和他联系。
Email: s*******[email protected]
Position Title: Junior Circuit Design Engineer (0-4 years’ experience)
Description/
Qualifications: RESPONSIBILITIES:
- High performance, low power custom circuit design in Digital-IP.
- Design, optimization, simulation and characterization of custom digital
circuits.
- Logic equivalence checking and transistor level function verification.
- 6T/8T/10T sram design.
- Library cells and datapath cells design.
- Knowledge of multi-port RF, ROM,... 阅读全帖 |
|
l********g
发帖数: 68 | 22 Posted job opennings several weeks ago. Three opennings have been filled and
there are still several left. Please send your resume to sramcache@yahoo.
com ASAP if you are interested.
The employer sponsors H1B, and starts green card process immediately after
you join us in full time based on current policy.
Position Title: Junior Circuit Design Engineer (0-4 years experience)
Description/
Qualifications: RESPONSIBILITIES:
- High performance, low power custom circuit design in Digital-IP.
- Design... 阅读全帖 |
|
l******1
发帖数: 85 | 23 拿到一个ARM memory design group 的onsite. 求复习建议。。看job description上
大多关于transistor level，memory,RTL coding的知识。。准备了半天。。电面的时
候问了好多 verification的东西。。觉得答得不太好。。。反到拿到onsite了。。真
心求建议。。贴个job description..
Accountabilities
Contribute in all parts of the memory development flow, starting at
design spec
Understand memory design and development in advanced technology (45nm
and beyond)
Perform physical verification, memory characterization, FE verification,
release procedure and QA flow of the memory compilers
Exp... 阅读全帖 |
|
d****y
发帖数: 76 | 24 【 以下文字转载自 JobMarket 讨论区 】
发信人: dooffy (dooffy), 信区: JobMarket
标 题: Opening of Circuit Design Engineer position at a Silicon Valley company
发信站: BBS 未名空间站 (Thu Nov 10 15:49:32 2011, 美东)
A reputable IC design company is now hiring Circuit Design Engineer with the
focus on transistor level circuit design. Please find the detailed job
description at the end of this message. If you are well qualified, we would
like you to join us as soon as you can. For those who are still in the
middle of degree pr... 阅读全帖 |
|
m*****t
发帖数: 3477 | 25 在摩尔精英上看到这篇评论，主要是讲国内IC特别是Analaog的现状。绝得靠谱。
这里的人对美国这边业界的情况可能更了解一些。板上也不少师出Wooley, Gray,
Allen, Abidi, Razavi, Sánchez, Song, Moon等名门的高徒。欢迎大家讨论和发表看
法。
PS: 像“半导体怎么样也不及CS”这种谁都知道且没营养的话，就不用再在这里累述了。
————————————————————————————————————
近来闲来无事看了一下招聘网站，发现电子行业工程师与互联网工程师的差距越拉越大
，颇感凄凉。而在电子行业内，不同岗位的待遇也千差万别。就好像模拟这个岗位，从
2007年走到高峰以后就一直往下走，待遇感觉也逐渐抛离。在与朋友讨论这个问题的同
时，朋友甩过来水木社区上的一个讨论，我们来看一下大家是怎么看待模拟待遇式微的
这个现象：
回答者：icore
现在做模拟的兄弟很多是给当年的老炮给忽悠来的。
什么模拟是艺术，什么格雷，拉扎维，艾伦，三森吹的像圣经。还有什么bob pease，
widlar江湖人物一般的个人传记，本版也是帮凶，什么模拟ic的九... 阅读全帖 |
|
dn
发帖数: 43 | 26 45nm with DUV 193nm source
Otherwise use EBL, or nano-imprinting |
|
S*****n
发帖数: 6055 | 27 intel CPU不是大量使用直径45nm的Cu nanowire么 |
|
C******g
发帖数: 2930 | 28 Sure, it was to be expected. News that the PS3 Slim would not only be
smaller (33% according to Sony; it's slim!), but use less power (34% less),
weigh less (36% lighter), and cost less (a solid Benjamin!) all pointed to
one thing: black magic.
Turns out, our hunch was wrong (at least in part). IBM has revealed that the
Cell processor under the newly matte-ified hood of the PS3 Slim was "
manufactured using an advanced 45-nanometer manufacturing process" that will
deliver "many performance impro |
|
w*******y
发帖数: 60932 | 29 $33.35 shipped after coupon code FIRE1205
Coupon code good through July 4th.
Link:
https://www.ewiz.com/detail.php?name=AD5000OGJ&title=AMD-Athlon-X2-Dual-Core
-Processor-5000-2-2GHz-AM2-OEM
Product Info:
Specifications
Mfr Part Number: AD5000OGJ22GI
Process Type: AMD Athlon X2 Dual-Core 5000+
Frequency: 2.2 GHz
FSB: 3200MHz
Cache: 2x 512KB
Process: 45nm SOI
Socket: Socket AM2+
Power Consumption: 65W
Package: OEM
|
|
w*******y
发帖数: 60932 | 30 Here is a great price on a decent processor, the deal is IN STORE ONLY. The
link is just for reference.
______ Price $32.33______ ____ Lowest is $55 on google products (even
microcenter is $61)
Link:
http://www.frys.com/product/5942414?site=sr:SEARCH:MAIN_RSLT_PG
Processor: AMD Athlon II X2
Model: 250 - ADX250OCGQBO X
Core Speed (MHz): 3000
Wattage: 65 W
L2 Cache: 1000KB
CMOS: 45nm SOI
Socket: AM3
Warranty: 3 YR
Tech Specs
L1: 256KB total
L2: 2MB total
L3: none
Multiplier: 15.0x
Package: 938- |
|
w*******y
发帖数: 60932 | 31 Specifications
IntelTechnology Intel64; Idle States; Enhanced IntelSpeedstep
Technology; Thermal Monitoring Technologies; Execute Disable Bit
Processor Core 2 Quad Q8300
Operating Frequency 2.5GHz
Number of Cores Quad-Core
Number of Threads 4 Processing Threads
Level 2 Cache 4MB
CPU Voltage 0.85V - 1.3625V
Thermal Power 95W
Processor Architecture 45nm SOI
Processor Data Width 64-Bit
Socket Type Socket 775
Front Side Bus 1333MHz
What's in the Box ... 阅读全帖 |
|
|
w*******y
发帖数: 60932 | 33 Athlon II x2 250 + free motherboard $58 + tax YMMV Microcenter B&M
I know it has been cheaper but this is not currently adverftised in
Microcenter's motherboard deals (free motherboard only with more expensive
cpus). Also, with the apparent death of Frys motherboard deals I think this
is worth mentioning.
In Houston, TX, I was able to place an order for:
AMD Athlon II X2 250 Dual Core Processor = 57.99
I could add for .01:
Biostar A780L Socket AM2 /AM3 760G mATX Motherboard (Reg. $49.99 DDR2 me... 阅读全帖 |
|
w*******y
发帖数: 60932 | 34 This is kind of local, not sure if it's at other stores, but I was just at
the Mayfield Heights, OH Micro Center and they had Core i7 950 open-box CPUs
for $159.96+tax. This is B&M only, not online. There were several when I
was there, not sure if other stores have them.
New is $199.99+tax, which is still considerably cheaper than anywhere else
online. Newegg is still $294.99 free shipping. The open box item can be
returned within 30 days if it doesn't work, and the balance of Intel's
warranty s... 阅读全帖 |
|
u****d
发帖数: 23938 | 35 All Clarkdale CPUs come with an IGP (Integrated Graphics Product) on die
which means it has a small embedded GPU inside the processor. Though the
processors will be manufactured at 32nm the totally weird thing is that the
graphics core is produced at 45nm. That can only mean one thing, the
Clarkdale processors will feature a multi-chip package, the CPU and GPU won'
t be merged into a single die but will have two chips merged in one package.
跟我想象的差距太大； |
|
a*******i
发帖数: 11664 | 36 08年的时候ibm把45nm技术卖给中芯了。
IBM
leverage
a |
|
