Jump to content
Search In
  • More options...
Find results that contain...
Find results in...
Spirit`yL

So,what`s difference between 512 and normally?

Recommended Posts

Basically it lets the processor work on 512 bits worth of data per clock cycle. Not relevant to most workloads, but it's incredible for massive data processing work.

 

https://www.intel.com/content/www/us/en/architecture-and-technology/avx-512-overview.html

image.png.c3a4e612b3fbe691e47ee8666a290c0f.png


¯\_(ツ)_/¯

 

 

Desktop:

Intel Core i7-3820 | Corsair H100i | ASUS P9X79-LE | 16GB Patriot Viper 3 1866MHz DDR3 | MSI GTX 970 Gaming 4G | 2TB WD Blue M.2 SATA SSD | 2TB Hitachi Deskstar HDD | 1TB WD Black HDD | Corsair CX750M Fractal Design Define R5 Windows 10 Pro / Linux Mint 19 Cinnamon

 

Laptop:

Dell XPS 15 9560 4K Touch | Intel Core i5-7300HQ | 12GB Generic (Crucial?) 2133MHz DDR4 | Nvidia GTX 1050 | 256GB Toshiba M.2 NVMe SSD | Windows 10

Link to post
Share on other sites
1 hour ago, Spirit`yL said:

Do it improve gaming performance?😂

 

AVX 512 and GPUs address similar problems in different ways. AVX 512 does more data computation with the same number of cores, GPUs just shove in a huge amount more cores. For gaming you don't need the ability to access hundreds of GB worth of RAM, so a GPU is much more useful there, whereas in something like a financial simulation you might need that much memory, which is where AVX 512 comes in since CPUs do have the ability to have that much RAM. Nvidia's NVLink is something trying to address the problem of limited memory amount on GPUs by making super high speed links so they can share RAM.

 

TL;DR:

AVX 512 no make game fast.


¯\_(ツ)_/¯

 

 

Desktop:

Intel Core i7-3820 | Corsair H100i | ASUS P9X79-LE | 16GB Patriot Viper 3 1866MHz DDR3 | MSI GTX 970 Gaming 4G | 2TB WD Blue M.2 SATA SSD | 2TB Hitachi Deskstar HDD | 1TB WD Black HDD | Corsair CX750M Fractal Design Define R5 Windows 10 Pro / Linux Mint 19 Cinnamon

 

Laptop:

Dell XPS 15 9560 4K Touch | Intel Core i5-7300HQ | 12GB Generic (Crucial?) 2133MHz DDR4 | Nvidia GTX 1050 | 256GB Toshiba M.2 NVMe SSD | Windows 10

Link to post
Share on other sites

Some people seem to think that the purpose of SSE/2 and AVX/2/E is solely to allow operations on longer integers. That is actually the least of it's purposes, and is only a by product of their implementations.

These instruction set extensions provide SIMD, or Single Instruction Multiple Data, models to programmers. They are meant to process the same operation on more than one unit of data at once.

For example, each 256 bit register can be treated as 8 32 bit integers, an operation of the form C = A + B can then perform 8 separate additions of 8 pairs of 32 bit integers in one instruction.

 

One should also be aware that it is not true SIMD in the sense that all of the additions (in this example) are physically being carried out at once: Rather, like all of Intels instructions, it is a highly optimized microprogram that is able to leverage parts of the processor not accessible to normal developers to improve the performance of "common" operations.


Teacher: Does anyone have a thin ruler?

Students: No.

Teacher: Ok. Does anyone have a nail file.

Students: Why do you need a nail file?
Teacher: I have a staple I need to remove.

Student: Oh. I have a staple remover...

 

People don't want power drills, they want holes.

Link to post
Share on other sites
Posted · Original PosterOP
On 2/15/2019 at 8:01 PM, straight_stewie said:

Some people seem to think that the purpose of SSE/2 and AVX/2/E is solely to allow operations on longer integers. That is actually the least of it's purposes, and is only a by product of their implementations.

These instruction set extensions provide SIMD, or Single Instruction Multiple Data, models to programmers. They are meant to process the same operation on more than one unit of data at once.

For example, each 256 bit register can be treated as 8 32 bit integers, an operation of the form C = A + B can then perform 8 separate additions of 8 pairs of 32 bit integers in one instruction.

 

One should also be aware that it is not true SIMD in the sense that all of the additions (in this example) are physically being carried out at once: Rather, like all of Intels instructions, it is a highly optimized microprogram that is able to leverage parts of the processor not accessible to normal developers to improve the performance of "common" operations.

Hold on,bro,you guy`s portrait is pretty insane😂

Link to post
Share on other sites
Posted · Original PosterOP
On 2/15/2019 at 7:36 PM, BobVonBob said:

 

AVX 512 and GPUs address similar problems in different ways. AVX 512 does more data computation with the same number of cores, GPUs just shove in a huge amount more cores. For gaming you don't need the ability to access hundreds of GB worth of RAM, so a GPU is much more useful there, whereas in something like a financial simulation you might need that much memory, which is where AVX 512 comes in since CPUs do have the ability to have that much RAM. Nvidia's NVLink is something trying to address the problem of limited memory amount on GPUs by making super high speed links so they can share RAM.

 

TL;DR:

AVX 512 no make game fast.

OK

Link to post
Share on other sites
9 minutes ago, Spirit`yL said:

Hold on,bro,you guy`s portrait is pretty insane😂

∀x i ❤️ x    ???


Teacher: Does anyone have a thin ruler?

Students: No.

Teacher: Ok. Does anyone have a nail file.

Students: Why do you need a nail file?
Teacher: I have a staple I need to remove.

Student: Oh. I have a staple remover...

 

People don't want power drills, they want holes.

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×