When we talk about processors, it is common to find the word «Instructions for AVX«, To the extent that they are sometimes required to run certain programs or games, and we may find ourselves in a situation where the game or system does not work for us when the processor does not have them. But whatwhat are the AVX instructions and what does the processor need?? Is your processor compatible with them? Let's see it.
What are the instructions for AVX
AVX Commands is a word from English «Advanced Vector Extensions«,, Also known as Sandy Bridge New Extensions since it was proposed by Intel in 2008 and started operating on Sand & # 39; s Bridge processors.
They contain a specific set of instructions for Intel and AMD processors that add to the construction of x86s for their associations, in the case of Intel from Sandy Bridge in 2011 and in the case of AMD from Bulldozer, also in 2011. AVX offers new features and new governing orders how the processor manages the data.
AVX commands can be varied, such as AVX2 expanded by most of the processor to 256 bits and including multiple cluster compilation (FMA) operations. By itself, AVX-512 extends AVX compatibility to 512 bits using the new EVEX prefix code proposed by Intel in 2013 and integrated for the first time in Knights Landing processors in 2016.
How do AVX commands use processors?
These instructions serve to tell the processor how to handle the incoming data, depending on the data type. In other words, it gives the processor greater administrative power, independently of the ability to decide which method is best to operate.
To be educated | Description |
VBROADCASTSS, VBROADCASTSD, VBROADCASTF128 | Copy 32, 64 or 128 bit memory to XMM or YMM directory |
VINSERTF128 | It replaces the upper part of the 256-YMM register with a 128 operand. The other part of your location has not been changed. |
VEXTRACTF128 | Extract the bottom half from the upper part of the 256 YMM register and copy the value to the 128-end function. |
VMASKMOVPS, VMASKMOVPD | Usually it reads any number of SIMD veth memory items in the destination directory, leaving the remaining vector items blank and setting the corresponding items in the local directory up to zero. Alternatively, write any number of elements from the vest SIMD register to the memory veector, leaving the remaining working memory items unchanged. In the AMD Jaguar processor text, this memory-based message takes more than 300 clock cycles when the mask is zero, in which case the instructions should do nothing. This seems to be a mark of design. |
VPERMILPS, VPERMILPD | Change lane. Include 32-bit or 64-bit veterans in the input operand. These are 256-bit commands in the queue, which means they work on all 256 bits with different random data of 128 bits, so they can't generate random runs on 128-bit lines. |
VPERM2F128 | Combine the four quadrants of the 128 vector of two 256-bit running sources into one 256-bit destination, always as a selector. |
VZEROALL | Set all YMM records to zero and mark them as used. It is used for switching between 128 and 256. |
VZEROUPPER | Set the upper half of all YMM records to zero. It is used for switching between 128 and 256. |
Does your processor have these commands?
There are times when, if you have a very old processor (before Sandy Bridge or Bulldozer), certain apps or games might work and you don't know why. It could be because the software needs to use any of the above AVX commands to work, and if your processor doesn't, it won't work.
These are the AVX active processes:
- Intel: Sandy Bridge (2011), Sandy Bridge-E (2011), Ivy Bridge (2012), Ivy Bridge-E (2013), Haswell (2013), Haswell-E (2014), Broadwell (2014), Broadwell- E (2016), Skylake (2015), Kaby Lake (2016-2017), Skylake-X (2017), Coffee Lake (2017), Cannon Lake (2018), Whiskey Lake (2018), Cascade Lake (2018) , Ice Lake (2019), Comet Lake (Core only, 2019), Tiger Lake (2020).
- AMD: Jaguar and Youth, New Puma, Bulldozer (2011), Piledriver (2012), Steamroller (2014), Excavator (2015), Zen (2017), Zen + (2018), Zen 2 (2019).