当前位置:网站首页>The performance of the server's four channel memory is improved. How about the performance of the four channel memory
The performance of the server's four channel memory is improved. How about the performance of the four channel memory
2022-06-25 17:42:00 【Full stack programmer webmaster】
Hello everyone , I meet you again , I'm your friend, Quan Jun .
Four channel memory has been improved , But the cost performance is not very good , If you are an enthusiast, you don't need to think about anything , If it is a pragmatist, the performance improvement brought by the four channels may not be obvious in use .
Building a four channel memory is simple , There are enough memory card slots ,X79 Chipset of ,i7-3960K The processor of ,Apacer Yuzhan Cheetah DDR3-1600 2GB * 4 perhaps a pirate boat DDR3 1600 2GB*4 Memory module is OK , Of course, other brands of memory modules can also , This is just an example , Just four pieces of the same memory . But the processor must be better , Otherwise, it will become the bottleneck of memory performance .
Users upgraded from the old platform can continue to use the previous machine DDR3 Memory , Its performance can meet the needs of most users . If you want to install a brand new machine , So spend a little more money on 4 Pieces of memory can be put into place at one time , And single 2GB Memory can also meet the needs of game applications , And the cost of buying memory will not be very large .
As for the performance improvement , Just look at the data and comments below .
Chess test
First, let's start with chess , as everyone knows , Chess can be said to be a more objective test software to measure the multi-threaded parallel computing ability of the platform .
【 Analysis of test results 】:
From the actual test results, we can see , Multichannel + Large memory does not bring higher scores , In the four cases, there is no bottleneck for simple operation , Even a single channel can have very good performance . The results of the four specifications are more errors , At present, most memory specifications are 1333-1600 In this range , This small-scale operation has no bottleneck .
CineBench test
Next, let's test the sensitivity of graphics rendering to the number of memory channels and memory capacity , Use of test tools CineBench R11.5, This is also a professional testing software . The software can test the processor and graphics card respectively , And it can be compared with other performance samples . Compared with chess , This kind of test is more close to our daily work . So will it be more memory dependent in such professional applications .
【 Analysis of test results 】:
Same as the previous chess test , The increase in the number of channels and memory capacity still does not lead to a significant improvement in performance . The increase of quadruple channel number and memory capacity is not proportional to the increase of performance , It will inevitably disappoint people . although i7-3960X It's good enough , The demand can be satisfied if such a situation exists in a single channel , But such a result will still surprise many people .
AIDA64 Cache & Memory Benchmark
Next, let's see how the increase in the number of memory channels can affect the memory bandwidth , And the software that can be qualified for this test should be AIDA64 Cache & Memory Benchmark 了 . The software can objectively reflect the platform memory reading 、 Write and copy bandwidth .
【 Result analysis 】:
From the results we can see , Different from the previous test results , The increase in the number of channels can also lead to more objective bandwidth improvement . However, the improvement from single channel to dual channel is very obvious , However, the performance improvement from dual channel to three channel and then to four channel has once again encountered a small bottleneck . So we think that in terms of bandwidth , The bandwidth provided by dual channel memory can feed the memory controller of the processor .
PCMark Vantage test
PCMark It is also a regular customer in our test , The test tool can test the performance of each part of the platform , And give very impressive results . Here we use PCMark Vantage Rather than PCMark 7 The reason is that the former has a special test project for memory performance , Can better see the actual performance of memory . Next, let's take a look at the test results :
【 Analysis of test results 】:
PCMark Vantage The test results are different from the previous tests , From single channel to three channel, the performance increases linearly , And when it comes to the fourth channel, the result is basically the same as that of the third channel . stay PCMark Vantage Under such test software , The four channels still do not show their advantages .
3DMark 11 test
After testing the computing performance and memory bandwidth , Let's take a look at the impact of memory performance on graphics processing power .
【 Analysis of test results 】:
It is basically consistent with the previous results , Dual channel can be said to maintain a good balance between performance and price , The performance improvement from the three channels is very limited . Even on highly configured platforms , Memory does not cause bottlenecks .
Publisher : Full stack programmer stack length , Reprint please indicate the source :https://javaforall.cn/151150.html Link to the original text :https://javaforall.cn
边栏推荐
- Use diskgenius to expand the capacity of system disk C
- College Students' hot summer exchange, Rog star product phantom 16 flipped version / phantom 13 / phantom x appointment
- CVPR小目标检测:上下文和注意力机制提升小目标检测(附论文下载)
- C语言中%含义
- Why do we need ankeri's active power filter in frequency converter occasions?
- 杰理之定时器捕获(timer_cap.c)使用注意事项【篇】
- 求满足条件的最长子串长度
- 华为云GaussDB(for Redis)揭秘第19期:GaussDB(for Redis)全面对比Codis
- Interrupt operation: abortcontroller learning notes
- 【Matlab】数据统计分析
猜你喜欢

Mathematical modeling - linear programming
![[UVM practice== > episode_2] ~ VIP, VIP development, VIP release](/img/de/4befb321d5cf5b0aa630e9ad0fda2a.png)
[UVM practice== > episode_2] ~ VIP, VIP development, VIP release

Intelligent dialog 01 redis installation

Assembly language (5) register (memory access)

A development of student management system based on PHP

DDD concept is complex and difficult to understand. How to design code implementation model in practice?
![Jericho's method of obtaining reset source and wakeup IO port [chapter]](/img/93/30d5075a3da1ace7438ae06c1b1775.png)
Jericho's method of obtaining reset source and wakeup IO port [chapter]

杰理之系统时钟设置出现复位或无效问题【篇】

Treasure and niche Chinese painting 3D texture material website sharing
![How Jerry used to output a clock source to the outside world [chapter]](/img/ea/161be6416726bcd80bb2823a5f6389.png)
How Jerry used to output a clock source to the outside world [chapter]
随机推荐
[UVM practice== > episode_1] ~ MCDF design update, AMBA standard interface, UVM verification environment update
Vscode automatically generates ifndef define ENDIF of header file
bert之我的小总结
一些常用的知识点积累
匯編語言(5)寄存器(內存訪問)
js禁止浏览器默认事件
数学建模——非线性规划
Treasure and niche Chinese painting 3D texture material website sharing
How about qiniu's Zhangle TenPay? Is it safe
学习太极创客 — MQTT(三)连接MQTT服务端
Precautions for using Jerry's timer [chapter]
什么是公链开发?公链开发项目有哪些?
Next. JS hot update markdown file change
Introduction to the container of() function
Create a new ar fashion experience with cheese and sugar beans
有关QueryInterface函数
cgi通过odbc连接数据库
MVDR beam MATLAB, MVDR beam forming matlab[easy to understand]
卷积操作的本质特性+TextCNN文本分类
Precautions for using timer_cap.c of Jerry [chapter]