联想拯救者Legion 5 Pro(国内型号R9000P 2021)体验

之前用的笔记本是dell precision 7520丐版(i7-7700HQ+M1200 升级到32GB 内存加上M1200显卡)。但是最近感觉他的软件越来越力不从心了。所以购买了新的机子。

之前调研了很多产品,主要考虑有限元计算,所以想着cpu要强。之前的是4核,准备升级个八核。市场上的8核cpu最强的就是AMD 5900HX,但是欧洲目前只有华硕有,但是华硕的在欧洲的价格比国内贵太多。丐版的魔霸(5900HX+GTX3070)都要2300欧左右(18000左右)。所以还是支持下联想。

联想拯救者legion 5 pro,欧洲这边报价1799欧元(14000)。我的型号是82JQ001PGE,主要配置(AMD 5800H+RTX3070+32GBRAM+1TSSD)具体配置可以看下面的链接https://psref.lenovo.com/Detail/Legion/Lenovo_Legion_5_Pro_16ACH6H?M=82JQ001PGE

这里采用另一个类似型号的配置截图(主要键盘背光不同,我的是4分区RGB)

我将原先dell的ssd (三星970EVO plus 1TB, NVME 2280)硬盘移到新的机子中(原装西数SN730 1TB),机子内硬盘插口处预留了螺丝。拆机有点麻烦,机子卡扣比较多,特别注意风扇出风口处的卡扣。之前没有拆过拯救者的话,最好看下视频比如b站这个。安装之后新的机子中有两个NVME 2280 SSD。采用了AS SSD Benchmark以及三星自带的软件对于ssd进行了测速,两者性能相近。

原装的系统是windows 10 home。安装硬盘开机之后,原系统直接可以使用且为windows 10 pro 版本,之后去联想官网下载Lenovo vantage。软件挺好用的,特别是Lenovo vantage toolbar应该是从thinkpad那里移植过来的。更新了驱动之后,电脑在系统上和原先的笔记本没有什么区别。所有软件该激活的都激活了,不过matlab要求更新license。

硬件上,联想的笔记本的屏幕是165Hz的2K高刷屏幕,肉眼可见的泛黄,提高亮度则情况有所好转。喇叭的低音一般般,高音还是可以的,音质上感觉不如我的sony小音箱(SRS-XB12立体声)。

键盘背光有几种典型模式,呼吸,变色(两类颜色可选),键盘的手感一般,比不上thinkpad,相对于dell这个也差了点(键程比较短),另外键盘的装配也一般,导致了,每个键的透光程度不一致。

整个机子的做功还是不错的,特别是A面和D面,全金属。B面的显示占了很大的面积,观感很好。重视有点重,机身快2.6KG了,加上300W适配器1.05KG(真的重)比老的dell工作站(机身2.79 kg,180W适配器0.58 kg)还重。

测试了下cpu的性能,Cinebech R23, 单线程 1312,多线程 12242。分别比我之前的笔记本(单线程851,多线程3602))提高54%和240%。还是比较满意的。暂时还没有测试有限元的计算速度。通过了abaqus 自带的verification,正常使用没有问题。跑了下鲁大师,结果如下:

测试了下MATLAB(2020a)的速度,跑了下知乎中的一段代码,原代码附在文末,共用时14.22s, 一般水平吧。相对的7700HQ为22s,4800H为18.3s。具体可以参考知乎回答https://www.zhihu.com/question/435741427/answer/1687421295

机子打开大型软件的速度还是比老机子快很多的。

试了两把Call of duty cold war,外接4K(HDMI接口),采用软件默认设置70fps左右,画面有撕裂,画质很好(比M1200好多了)。同时测试了戴森球,4K 4倍抗锯齿,fps在35帧左右。

另外显存8GB感觉以后可能不够用。可能是4K显示器的缘故,call of duty用了70%,戴森球用了73%。CPU在两个游戏中的占用都很低,戴森球占用小于10%。

暂时就测试这些,后面看看新电脑能否提高工作效率。

2021年4月14日更新

在外接显示器时(机子显示器开启状态),机子在玩大型游戏时,切出游戏到桌面时,电脑会比较卡。同时切回游戏时,回闪回桌面,再切游戏会恢复正常。

代码如下:

clear, clc
A = rand(3e3);
B = rand(3e3);
num = 3;
T = zeros(8,num);
for i = 1:num
    % Test1 乘法
    clc
    disp('^-------')
    disp(['第' sprintf('%4i',i) ' 轮乘法测试中...'])
    tic, X1 = A*B; T(1,i) = toc;
    % Test2 稀疏矩阵
    clc
    disp('-^------')
    disp(['第' sprintf('%4i',i) ' 轮稀疏矩阵测试中...'])
    tic, X2 = sparse(A); T(2,i) = toc;
    % Test3 逆矩阵
    clc
    disp('--^-----')
    disp(['第' sprintf('%4i',i) ' 轮逆矩阵测试中...'])
    tic, X3 = inv(A); T(3,i) = toc;
    % Test4 快速傅里叶
    clc
    disp('---^----')
    disp(['第' sprintf('%4i',i) ' 轮快速傅里叶测试中...'])
    tic, X4 = fft(A); T(4,i) = toc;
    % Test5 LU分解
    clc
    disp('----^---')
    disp(['第' sprintf('%4i',i) ' 轮LU分解测试中...'])
    tic, [L5,U5,P5] = lu(A); T(5,i) = toc;
    % Test6 QR分解
    clc
    disp('-----^--')
    disp(['第' sprintf('%4i',i) ' 轮QR分解测试中...'])
    tic, X6 = qr(A); T(6,i) = toc;
    % Test7 奇异值分解
    clc
    disp('------^-')
    disp(['第' sprintf('%4i',i) ' 轮奇异值分解测试中...'])
    tic, [U7,S7,V7] = svd(A); T(7,i) = toc;
    % Test8 特征值与特征向量
    clc
    disp('-------^')
    disp(['第' sprintf('%4i',i) ' 轮特征值与特征向量测试中...'])
    tic, [V8,D8] = eig(A); T(8,i) = toc;
end
clc
% 各项测试平均时间
t = sum(T,2)./num;
disp(['Multiplication :   ' sprintf('%6f',t(1))])
disp(['Sparse         :   ' sprintf('%6f',t(2))])
disp(['Inverse        :   ' sprintf('%6f',t(3))])
disp(['FFT            :   ' sprintf('%6f',t(4))])
disp(['LU             :   ' sprintf('%6f',t(5))])
disp(['QR             :   ' sprintf('%6f',t(6))])
disp(['SVD            :   ' sprintf('%6f',t(7))])
disp(['Eigen          :   ' sprintf('%6f',t(8))])
% Total
total = sum(t);
disp('----------------------------')
disp(['Total          :   ' sprintf('%6f',total)])

达芬奇剪辑软件对于内存的巨大占用

今天利用达芬奇软件剪辑了两个时间长度为1分40秒的FHD@50Hz的视频,cpu占用30%, RAM占用直接上了40G,显存占用9.0G,其中6.6G为显卡显存,2.4G为共享内存。真的是伤不起。

以下为资源管理截图:

平台配置

Device name DESKTOP-KO7EL1P
Processor AMD Ryzen 9 3950X 16-Core Processor 3.50 GHz
Installed RAM 64.0 GB
System type 64-bit operating system, x64-based processor
Edition Windows 10 Pro
Version 20H2
Nvidia GeForce GTX 1080 8G
Installed on ‎2020/‎7/‎15
OS build 19042.804

a set of interesting papers written by George M. Whitesides

Here I collect some interesting paper written by George M. Whitesides. Most of them are not serious scientific papers but interesting and useful.

Whitesides, George M. “Curiosity and science.” Angewandte Chemie International Edition 57.16 (2018): 4126-4129.

Whitesides, George M. “Soft robotics.” Angewandte Chemie International Edition 57.16 (2018): 4258-4273.

Whitesides, George M. “Whitesides’ group: writing a paper.” Advanced Materials 16.15 (2004): 1375-1377.

Ilievski, Filip, et al. “Soft robotics for chemists.” Angewandte Chemie 123.8 (2011): 1930-1935.

Nature energy 投稿记录 record of submission

First round

Review received 23. September 20
Manuscript under consideration 23rd September 20
Review assigned 17th July 20 Editor Decision Started 17th July 20
Manuscript under consideration 2nd July 20
Editor assigned 2nd July 20
Manuscript received 1st July 20
Manuscript under submission 1st July 20

Second round

Manuscript under consideration 1st December 20
Editor Decision Started 13th November 20
Manuscript under consideration 13th November 20
Editor assigned 12th November 20
Manuscript received 12th November 20
Manuscript under submission 12th November 20

Third round

Accepted 18th December 20
Manuscript under consideration 18th December 20
Editor assigned 7th December 20
Manuscript received 7th December 20
Manuscript under submission 6th December 20

Published 28th January 2021

海外使用FTP访问阿里云服务器奇慢无比

RT

使用阿里云独立虚拟主机国内节点,在欧洲通过FTP访问服务器的文件,速度特别慢,而且时常没有响应,需要重新链接。所以需要使用这个服务的话,尽量不要悬着阿里云的国内主机节点。

除此以外,在利用wordpress自动升级的时候,发现网站本身无法连接到wordpress 的更新服务器,所以只能采用手动升级(5.5->5.6).升级到5.6之后,原先不能自动升级的主题也可以自动升级了。正在使用5.6 暂时没有什么问题。

FTP客户端,Windows 10 操作系统,WinSCP 客户端,客户端设置参照阿里云官网设置。

Ecoflex, PDMS comparasion

materialViscosity (base & Mixed)Tensile strengthdesity
g/cm^3
ModulusPot life (Room temp)
PDMS (10: 1)5100 cp/5.5 Pa-sec, 3500 cp 3.5 Pa-sec6.7 Mpa0.9651.35 Mpa1.5 h
ECOFLEX 00-30 (1:1)3, 000 cps/3 pa-sec1.379Ma1.070.068947 MPa45 min
Viscosity of water, 0.001

A method to fitting the magnetic field of a plate magnet.

I find I make wrong table in my paper (Soft electromagnetic actuators), here I try to correct it.

In the paper, we use a plate magnet (Fig. S3A, with a diameter=10 cm and height=3 cm, N45 coated with Nickel) which can be purchased via the link(www.magnet-shop.com).

At first, we use a Gauss meter (Gaussmeter HGM09s, Goudsmit Magnetic Systems) to measure the magnetic field. Then we fitting the experimental data with Comsol (Fig. S3B& Fig. S3C). Only one parameter needs to fit, the remnant magnetization.

From the manufacture, we obtain the remnant magnetization is 1.32-1.37 T. We obtain the remanent magnetization of the plate magnet, 1.31 T from fitting (Fig. S3D). It can use a simple equation to fit the magnetic field at the center of the magnet where R=0 (Fig. S3E).

Also we can know the distribution of the magnetic field from simulation(Crosection: Fig. S3 F&G, Field: Fig. S3H&I). With Matlab (lsqcurvefit, or fit) we are able to fitting the magnetic field with the following equations:

The parameter P and Q are not correct in the paper (Table S1) which comes from a unknow version even thou it can generate a similar magnetic field, here I provide the orginal data (Matlab code and data) and the program to plot the Fig. S3 F,G,H,I:

Fig. S3. Characterization of a plate magnet. (A) Picture of the plate magnet. (B) Simulated spatial distribution of the magnetic field around a plate magnet. The legend shows the strength of the magnetic potential. (C) Measurements of the magnetic field were taken at the labeled sites 6 mm above the surface of the magnet. (D) Comparison of magnetic flux density between the experimental measurement and the simulation corresponding to the labeled sites in (C). (E) Simulated and fitted magnetic flux density along the symmetry axis of the magnet at R=0. (F&G) Distribution of the two components of the magnetic field at different heights Z above the surface of the magnet. (H&I) Contour maps of a 3D numerical fitting of the magnetic field on the top of the plate magnet compared with the simulated results. (Photo Credit: Michael Drack and Guoyong Mao/Johannes Kepler University Linz)

Made in China in a view of a researcher

While reading a paper, I got a intresting finding. The paper is about a magnetic refrigerant bought from a chinese company Baotou Steel and Rare Earth Company of China. In a section, the researcher anaylsed the elements of the material. Even tough the company claimed it is 99.9% pure, but actually it is 93% (99.7% wt.%) which is not so pure. Then its performance is slightly inferior to the data from Ames Laboratory. However, then the researcher compared the price of the Chinese material with the material used in the Ames Laborotory, which are $110/kg and $23,000/kg, respectively. Very happy to use chinses product.

original short-cut.

paper: Zimm, C., A. Jastrab, A. Sternberg, V. Pecharsky, K. Gschneidner, M. Osborne, and I. Anderson. “Description and performance of a near-room temperature magnetic refrigerator.” In Advances in cryogenic engineering, pp. 1759-1766. Springer, Boston, MA, 1998.

奥地利科研人员居留延期申请

materials for renewing residence permit 拘留延期申请需要办理的材料。摄制于2019年9月
  1. 护照# 需要护照首页,理论上所有含有信息的页面都需要复印,但是签证官也没有看。
  2. 照片。 3.5×4.5 cm,#最好带有日期的照片,6个月内。
  3. 租赁合同 #租房,如果住房地址没有更换的话,理论上不需要
  4. 健康保险说明(一般在合同中), 带上E-card(for insurance) #研究人员的保险合同
  5. 工作合同(保证自己能在奥地利生活下去)#工作合同中有工资说明
  6. 最近三个月的工资单,要求单位打印出的。银行的收入history并没有用。
  7. 带上之前的ID card
  8. 被要求 Hosting Agreement,主要由单位出具,接收证明(如果是project post-doc申请researcher type的residence permit那么这个是必须的。如果是University assistant, 那么就不需要这个, 学校不提供,而且签证类型改为 Settlement permit special cases of gainful employment)。
  9. KSV-AUszug (非必要的认证,从来没有过。)# 类似蚂蚁信用的东西,没什么用。

所有材料需要带上原件与复印件。

其他的参考资料包裹OEAD上的信息

https://oead.at/en/to-austria/entry-and-residence/settlement-permit-researcher/#checklists

关于 初次签证居留相关信息参考