4G, 64bit, PAE

The memory was added to 4 GB last weekend. The reason is that in normal working environments, outlook, communicator, ie, Firefox, VMWare, or an IDE is enabled under Windows 7, the background is running MSE, youdao dictionary, wakoopa, Evernote, and Gae proxy. The total number of processes is 70 +, and the memory usage is 80% +. This is really high, and aero, these features are all turned off. To alleviate this problem, I bought two Kingston 2G 800, but since the Intel gm965 White Paper says it only supports DDR2 667 and 533, for example, so the two slices run at a frequency of 667:

The CPU-Z is detected at 332.5 MHz, because it is a DDR, multiplied by 2 is 667.

How can I make the memory work at MHz? At first, I wanted to use setfsb and SPD tool overlook. When talking about setfsb, there was another interesting thing. Several years ago, it seemed that there was a Chinese user fool who wrote setfsb, you still need an activation code to use it, and a warning message is displayed ,:

Later, I upgraded from t5500 to t8100, and the performance improvement has been very obvious. Even if it reaches t9300, it will certainly not change much. More importantly, setfsb does not seem to support gm965. However, the author has been updating and may support it.

"Of course, whether using clockgen or setfsb, you must know the model of the PLL to change the external frequency. However, there are thousands of different types of things on the market. The author of clockgen has been suspended for a long time because of this headache. At present, the pm965 and gm965 series on the market are using a total of more than 10 kinds of PLL crystal oscillator chips, such as mainstream manufacturers: HP, Acer, Fujitsu, Toshiba 965 series notebook boards are basically usedIcs9lpr363dglfThis model. Some domestic manufacturers, such as Shenzhou and Qixi, use slg8sp510t and other cheap chips on the pm965 motherboard to reduce costs, such as the 825d series l23xt/l5x5t/l7xxt series in Shenzhou. These crystal oscillator chips are rare in the Japanese market and are not yet supported by setfsb. Their authors are also adding and improving these chips"

"Can I use it? It has nothing to do with the CPU. The key is whether the software contains the model of the frequency generator on your motherboard?
ICS model, near the CPU and beiqiao, similar"9lprs365bglf"

Image of the Shenzhou Forum theft:

If the memory is used, even if the SPD information is modified, it is unstable when the motherboard is not supported, and My bios seems a little old. The blue screen is displayed during the last standby time, windows reminds me to consider upgrading the BIOS.

The above seems to be a little far-fetched. It comes with a normal working environment. The effect is obvious, and the response speed is obviously faster. Even in peak hours, it is only 60% +

The information on the system properties page is as follows. It was updated to Windows 7 SP1 last week.

 

I installed a 32-bit Windows 7. By default, the system displays "4.00 GB (2.99 GB usable)". Of course, the 32-bit system certainly supports memory larger than 4 GB, otherwise, how can the 32-bit OS servers be explained? (the server technology is at least five years ahead of the normal PC technology ..).

Speaking of 64-bit, in order to catch up with fashion, when I first got out of RC in win7, I said goodbye to XP, but installed Windows Server 2008x64, and to support cpuvt technology, I changed the CPU, at that time, I was quite interested in the server and had a strange idea. I wanted to use the server hardware and system. At that time, I had to go through VMware esxi (hardware is not supported), xen server, hyper-V (installed on the home desktop ).

I was running a dual-system, and the other was centos 64-bit. Later I changed to Gentoo's x64 version. Although it is based on 64-bit hardware, it does not feel that 64-bit OS runs faster. Even now, there are few software that supports 64-bit native, if you run a 32-bit application through the 32-seat system (syswow64 directory is impressive), the speed will be somewhat compromised, as well as the driver. If it is not a hardware driver, many database connection drivers for ODBC are not 64-bit, but you can run the 32-bit ODBC Data Source Management Program.

Similarly, in Linux, the situation is slightly better, because many packages can be compiled and installed from the source, but some programs also need 32-bit libraries, in this way, a lot of hard disk space (Lib, lib32, and lib64 Directories) will be wasted. In this case, my laptop is replaced by the motherboard, such as CPU, hard disk, memory, fan, and Bluetooth, I haven't thought about the optical drive yet. I can find a screw change interface and replace it with the optical drive. The previous optical drive burned several hundred images and then sacrificed them. Maybe I can also buy a hard drive box with a drive space, you can consider adding an SSD hard disk to the original hard disk location, so that it is faster to start and load the OS, and now it is much easier to use a USB system.

Usually speaking of the 64-bit, most of the time refers to Intel 64 (em64t/amd64/x86-64/x64), there is also the IA64 andt processing, in addition to the server, basically not touch; the advantage of 64-bit is that,

In theory, the 1.64-bit address space can support 64-power 2 and 16 EB memory;

• 2. extended register set (Extended register set). In my understanding, this can make the call conventions of the function consistent, and the parameters and return value pointers can be passed through registers, this is more efficient and can avoid some traditional stack overflow vulnerabilities.
• 3. for developers, the CPU Instruction Set changes are not large, which is equivalent to a union;
• 4. Forward compatibility with applications and operating systems using similar wow64 (Windows-on-Windows 64) Technologies
• Maximum memory actually supported by Windows

How does a 32-bit OS support memory larger than 4 GB? You can use PAE (physical address exetension) to implement it. For detailed principles, refer to this article, which is very well described.

I am using the readyfor4 GB tool. After the patch is restarted, the "test mode" watermark is displayed in the lower right corner. This way: however, you can remove it using another starforce tool, it's cool.

Select Remove watermark

As for Linux, this is simple. For details, refer to this article. In ubuntu, run sudo apt-Get install linux-server directly and restart it.

The default 32-bit Ubuntu only recognizes 3 GB

Install linux-image-server, Linux-headers-generic-PAE, and the package name indicates its main purpose.

Recognized