Semiconductor & Computer Engineering
From WikiChip
Difference between revisions of "intel/dl boost"
(→Overview) |
|||
| Line 5: | Line 5: | ||
'''DL Boost''' is a term used by [[Intel]] to describe a set of features on their microprocessors designed to accelerate AI workloads. The term was first introduced with {{intel|Cascade Lake|l=arch}}. DL Boost includes the following features: | '''DL Boost''' is a term used by [[Intel]] to describe a set of features on their microprocessors designed to accelerate AI workloads. The term was first introduced with {{intel|Cascade Lake|l=arch}}. DL Boost includes the following features: | ||
| − | * {{x86| | + | * {{x86|AVX512_VNNI|AVX-512 Vector Neural Network Instructions}} (AVX512_VNNI), first introduced with {{intel|Cascade Lake|l=arch}} (server) and {{intel|Ice Lake (Client)|Ice Lake|l=arch}} (Client) |
| − | * | + | * {{x86|AVX512_BF16|AVX-512 BFloat16 Instructions}} (AVX512_BF16), first introduced with {{intel|Cooper Lake|l=arch}}. This extension offers instructions for converting to [[bfloat16]] and then performing multiply-accumulate on two bfloat16 values. |
| + | == Implementations == | ||
| + | {| class="wikitable" | ||
| + | |- | ||
| + | ! Microarchitecture !! {{x86|AVX512_VNNI}} !! {{x86|AVX512_BF16}} | ||
| + | |- | ||
| + | ! colspan="3" | Client | ||
| + | |- | ||
| + | | {{intel|Ice Lake (Client)|l=arch}} || {{tchk|yes}} || {{tchk|no}} | ||
| + | |- | ||
| + | ! colspan="3" | Server | ||
| + | |- | ||
| + | | {{intel|Cascade Lake|l=arch}} || {{tchk|yes}} || {{tchk|no}} | ||
| + | |- | ||
| + | | {{intel|Cooper Lake|l=arch}} || {{tchk|yes}} || {{tchk|yes}} | ||
| + | |- | ||
| + | | {{intel|Ice Lake (Server)|l=arch}} || {{tchk|yes}} || {{tchk|no}} | ||
| + | |- | ||
| + | | {{intel|Sapphire Rapids|l=arch}} || {{tchk|yes}} || {{tchk|yes}} | ||
| + | |- | ||
| + | | {{intel|Granite Rapids|l=arch}} || {{tchk|some|TBD}} || {{tchk|some|TBD}} | ||
| + | |} | ||
| + | |||
| + | == See also == | ||
| + | * [[bfloat16]] | ||
| + | * [[Neural processors]] | ||
| + | * [[Acceleration]] | ||
[[category:intel]] | [[category:intel]] | ||
Revision as of 23:42, 21 June 2020
DL Boost (deep learning boost) is a name used by Intel to encompass a number of x86 technologies designed to accelerate AI workloads, specifically for the acceleration of inference.
Overview
DL Boost is a term used by Intel to describe a set of features on their microprocessors designed to accelerate AI workloads. The term was first introduced with Cascade Lake. DL Boost includes the following features:
- AVX-512 Vector Neural Network Instructions (AVX512_VNNI), first introduced with Cascade Lake (server) and Ice Lake (Client)
- AVX-512 BFloat16 Instructions (AVX512_BF16), first introduced with Cooper Lake. This extension offers instructions for converting to bfloat16 and then performing multiply-accumulate on two bfloat16 values.
Implementations
| Microarchitecture | AVX512_VNNI | AVX512_BF16 |
|---|---|---|
| Client | ||
| Ice Lake (Client) | ✔ | ✘ |
| Server | ||
| Cascade Lake | ✔ | ✘ |
| Cooper Lake | ✔ | ✔ |
| Ice Lake (Server) | ✔ | ✘ |
| Sapphire Rapids | ✔ | ✔ |
| Granite Rapids | TBD | TBD |