From WikiChip
Editing intel/cores/cascade lake sp
Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
This page supports semantic in-text annotations (e.g. "[[Is specified as::World Heritage Site]]") to build structured and queryable content provided by Semantic MediaWiki. For a comprehensive description on how to use annotations or the #ask parser function, please have a look at the getting started, in-text annotation, or inline queries help pages.
Latest revision | Your text | ||
Line 20: | Line 20: | ||
|successor=Ice Lake SP | |successor=Ice Lake SP | ||
|successor link=intel/cores/ice lake sp | |successor link=intel/cores/ice lake sp | ||
− | |||
− | |||
}} | }} | ||
'''Cascade Lake SP''' ('''{{intel|Cascade Lake|l=arch}} Scalable Performance''') is code name for Intel's series of server [[multiprocessors]] based on the {{intel|Cascade Lake|l=arch}} microarchitecture as part of the {{intel|Purley|l=platform}} platform serving as the successor to {{intel|Skylake SP|l=core}}. These chips support up to 8-way multiprocessing, up to [[28 cores]], incorporate a new {{x86|AVX512-VNNI|AVX512}} [[x86]] {{x86|extension}} for neural network / deep learning workloads, and introduces [[persistent memory]] support. Cascade Lake SP-based chips are manufactured on an enhanced [[14 nm process]] and utilize the {{intel|Lewisburg|l=chipset}} chipset. | '''Cascade Lake SP''' ('''{{intel|Cascade Lake|l=arch}} Scalable Performance''') is code name for Intel's series of server [[multiprocessors]] based on the {{intel|Cascade Lake|l=arch}} microarchitecture as part of the {{intel|Purley|l=platform}} platform serving as the successor to {{intel|Skylake SP|l=core}}. These chips support up to 8-way multiprocessing, up to [[28 cores]], incorporate a new {{x86|AVX512-VNNI|AVX512}} [[x86]] {{x86|extension}} for neural network / deep learning workloads, and introduces [[persistent memory]] support. Cascade Lake SP-based chips are manufactured on an enhanced [[14 nm process]] and utilize the {{intel|Lewisburg|l=chipset}} chipset. | ||
Line 29: | Line 27: | ||
== Overview == | == Overview == | ||
− | Cascade Lake SP processors are based on Intel's {{intel|Cascade Lake|l=arch}} microarchitecture which brings a modest frequency improvement due to an enhancement process technology along with {{intel|DL Boost}}, an {{x86|extension}} designed to speed up machine learning algorithms. Those processors support between two and eight-way multi-processing (the exact support depends on the Xeon family) and with all models supporting hex- | + | Cascade Lake SP processors are based on Intel's {{intel|Cascade Lake|l=arch}} microarchitecture which brings a modest frequency improvement due to an enhancement process technology along with {{intel|DL Boost}}, an {{x86|extension}} designed to speed up machine learning algorithms. Those processors support between two and eight-way multi-processing (the exact support depends on the Xeon family) and with all models supporting hex-chanel ? GiB of DDR4 ECC memory or ? TiB for extended memory models. Cascade Lake also brings support for [[persistent memory]]. |
As with {{\\|Skylake SP}}, Cascade Lake SP processors utilize the new {{intel|FCLGA-3647}} package (which makes use of "Socket P"). Those use the {{intel|Lewisburg}} chipset ({{intel|Platform Controller Hub|HUB}}) via 4 PCIe3 lanes using Intel's proprietary {{intel|Direct Media Interface}} 3.0 (DMI 3.0), allowing for 8 GT/s transfer rate per lane. When in multi-socket configuration, the microprocessor is connected to the other processors via the {{intel|Ultra Path Interconnect}} (UPI) links which Intel introduced with Skylake SP as well, replacing and obsoleting the older {{intel|QuickPath Interconnect}} (QPI) operating. Depending on the model, there may be either two or three UPI links inter-linking each socket (for more details see {{intel|cascade lake#Scalability|Cascade Lake § Scalability|l=arch}}). | As with {{\\|Skylake SP}}, Cascade Lake SP processors utilize the new {{intel|FCLGA-3647}} package (which makes use of "Socket P"). Those use the {{intel|Lewisburg}} chipset ({{intel|Platform Controller Hub|HUB}}) via 4 PCIe3 lanes using Intel's proprietary {{intel|Direct Media Interface}} 3.0 (DMI 3.0), allowing for 8 GT/s transfer rate per lane. When in multi-socket configuration, the microprocessor is connected to the other processors via the {{intel|Ultra Path Interconnect}} (UPI) links which Intel introduced with Skylake SP as well, replacing and obsoleting the older {{intel|QuickPath Interconnect}} (QPI) operating. Depending on the model, there may be either two or three UPI links inter-linking each socket (for more details see {{intel|cascade lake#Scalability|Cascade Lake § Scalability|l=arch}}). | ||
Line 47: | Line 45: | ||
** Gold and up also have Node Controller Support and offer Integrated Omni-Path Fabric Interface options | ** Gold and up also have Node Controller Support and offer Integrated Omni-Path Fabric Interface options | ||
− | {{ | + | Models that are suffixed with "''T''" have extended lifetime (10 year use) guarantees and [[NEBS]]-friendly packing specification. Additionally, models that are suffixed with "''F''" (CSL-F) integrate the {{intel|Omni-Path}} Host Fabric Interface (HFI) die on-package. |
− | |||
− | |||
− | + | {{clear}} | |
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
== Cascade Lake SP Processors == | == Cascade Lake SP Processors == | ||
Line 174: | Line 158: | ||
</table> | </table> | ||
{{comp table end}} | {{comp table end}} | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
== See also == | == See also == | ||
{{intel cascade lake core see also}} | {{intel cascade lake core see also}} |
Facts about "Cascade Lake SP - Cores - Intel"
chipset | Lewisburg + |
designer | Intel + |
first announced | April 2, 2019 + |
first launched | April 2, 2019 + |
instance of | core + |
isa | x86-64 + |
isa family | x86 + |
main image | + |
manufacturer | Intel + |
microarchitecture | Cascade Lake + |
name | Cascade Lake SP + |
package | FCLGA-3647 + |
platform | Purley + |
process | 14 nm (0.014 μm, 1.4e-5 mm) + |
socket | Socket P + and LGA-3647 + |
technology | CMOS + |
word size | 64 bit (8 octets, 16 nibbles) + |