SlideShare a Scribd company logo
( , 67
ts Odʼ’a
L P
( , 67
( , 67
l  s x– ~∼y
l  s
r
v v
l  s v p
l 
k  v x
y
o v
k  C6
o ~∼ fi ~∼ v v v
l 
k  –BFp v
(
( , 67
-‐‑‒ -‐‑‒ ʼ’mg
i L e
t
)
DRAM(and(Flash(Scaling:((
“The(End(is(Nigh”(
1985( 1990( 1995( 2000( 2005( 2010( 2015( 2020(
Density,
DRAM(
SLC(NAND(
M FdWTa 4FC BF( N
( , 67
Summary(
Register((
Cache(
STT&RAM,,
NV&DIMM,
ns=class(NVM(
RRAM,,PCM,,
Low(us=class(NVM(
NAND,SSD,
High(us=class(NVM(
HDD(
Low(ms=class(Mass(Storage(
Capacity( 10(
”
v
k  so ~∼
k  s –~∼
M FdWTa 4FC BF( N
( , 67
whg Ld
l  v p
– fi t
l  ” ~∼“ fi
k  p p p r
k 
l 
l  p ~∼
k  p
( , 67
ts
l  ~∼ p” t
k  ]G p v t
k  v t
l  77EpA4A7 t
k  v t
l  v
k  z v t ~∼ ~∼
k  ” ~∼ x ~∼y
l  sA4A7
,
( , 67
v x CH F8y
V X oL
oL
-‐‑‒
2015 2030
vv v
hmd
fl
v
v fi
Lcndl
v
rr
r
rr
r
( , 67
l  (,
l  )
l  JSP v
.
Vaa 0 b Wa OW a U] X OW Z aO
W[ bZ S W[ bZ S Va[Z
CPU
	
NAND
	
RAM
• 
• 
• 
• 
RAM
“ ”
( , 67
: : I
The$Machine$could$be$six$%mes$more$powerful$than$an$
equivalent$conven2onal$design,$while$using$just$1.25$
percent$of$the$energy$and$being$around$1/100$the$size.
h:p://www.hpl.hp.com/research/systems@research/themachine/
( , 67
l  “ •~∼
k  B S 6][ baS C ]XSQa   9OQSP]] ”
k  EOQ FQOZS 6][ baW U    aSZ
k  8ea S[SZf FV W W U 6][ baW U    5
k  GVS OQVW S    C
k  9W S5]e   H65
k  6GE 6] ] aWb[    G
l 
k  v x p p v y
k  v
k  v
l  – p p p ~∼–
( , 67 (
IoTpCPSpHPCp p
v
•  v v v
•  v v
•  SoC (eFlash?)
•  v v
•  v v v v
•  v
•  DRAM
( , 67
gin a
 
k 
(  
k  Z]OR a] S
k  v v
)  
k  v
)
NVM)SSD)Challenges)
•  So:ware)overheads)in)
kernel)
7)
0
5
10
15
20
25
BaseLatency(us)
PCM
Ring
DMA
Wait
Interrupt
Issue
Copy
Schedule
OS/User
Software is Critical
• Baseline Latencies:
– Hardware: 8.2 us
– Software: 13.4 us
Hardware costs
11[Caulfield,)SC’10])
M6ObZjSZR F6 N
v
( , 67
-‐‑‒ -‐‑‒
An application using NVM.FILE mode may or may not be using memory-mapped file
behavior.
The NVM.FILE mode describes NVM extensions including:
• Discovery and use of atomic write features
• The discovery of granularities (length or alignment characteristics)
4.3.3 NVM.PM.VOLUME mode overview
NVM.PM.VOLUME mode describes the behavior for operating system components (
file systems) accessing persistent memory. NVM.PM.VOLUME mode provides a soft
abstraction for Persistent Memory hardware and profiles functionality for operating sy
components including:
• the list of physical address ranges associated with each PM volume
• the capability to determine whether PM errors have been reported
Figure 5 NVM.PM.VOLUME and NVM.PM.FILE mode examples
Application
PM device PM device PM device. . .
User space
Kernel space
MMU
MappingsPM-aware file system
NVM PM capable driver
Load/
store
Native file
API
PM-aware kernel module
PM device
NVM.PM.VOLUME mode
NVM.PM.FILE mode
4.3.4 NVM.PM.FILE mode overview
NVM.PM.FILE mode describes the behavior for applications accessing persistent me
The commands implementing NVM.PM.FILE mode are similar to those using NVM.F
l  AI 5Z]Q []RS
l  AI 9WZS []RS
l  CS W aS a S[] f I]Zb[S []RS
l  CS W aS a S[] f 9WZS []RS
Note that there are other models for connecting a non-PM file system to PM hardware.
4.3 NVM programming modes
4.3.1 NVM.BLOCK mode overview
NVM.BLOCK and NVM.FILE modes are used when NVM devices provide block storage
behavior to software (in other words, emulation of hard disks). The NVM may be exposed as a
single or as multiple NVM volumes. Each NVM volume supporting these modes provides a
range of logically-contiguous blocks. NVM.BLOCK mode is used by operating system
components (for example, file systems) and by applications that are aware of block storage
characteristics and the block addresses of application data.
This specification does not document existing block storage software behavior; the
NVM.BLOCK mode describes NVM extensions including:
• Discovery and use of atomic write and discard features
• The discovery of granularities (length or alignment characteristics)
• Discovery and use of ability for applications or operating system components to mark
blocks as unreadable
Figure 4 NVM.BLOCK and NVM.FILE mode examples
Application
NVM block capable driver
File system
Application
NVM device NVM device
User space
Kernel space
Native file
API
NVM.BLOCK mode
NVM.FILE mode
4.3.2 NVM.FILE mode overview
NVM.FILE mode is used by applications that are not aware of details of block storage
hardware or addresses. Existing applications written using native file I/O behavior should work
FA 4 C ]U O[[W U ]RSZ I]Z
( , 67
l  v CS W aS a S[] f   C
t
l  – 4C p
C fi p 4C
k  [OZZ]Q – fSa O ]aVS [OZZ]Q
l  v v t
k 
k  v
( , 67
Lcndl
,
6CH
AIE4 7E4
6CH
AIE4
7E4
6CH
AIE4
  4 ES ZOQS RW   5 FVO SR ORR S OQS   6 8 aW SZf AIE4
5Z]Q 9
S[] f 9
QOQVS QOQVSQOQVS
AIE4 B AIE4
( , 67
c mg M N
l  AIE4
k  Fa] S • AIE4 ~∼
k  r r v
-‐‑‒
• Recovery depends on write ordering
CPU
Write-back
Cache
NVM
V D
VD
STORE data[0] = 0xFOOD
STORE data[1] = 0xBEEF
STORE valid = 1
Crash
D
D
CPU
Persistent Memory (PM) Ordering
M FdWTa 4FC BF( N
( , 67
c mg MRN
l 
k  ”“ t
k  aSZ CS W aS a S[] f v
l  6 9 HF BCG
l  6 J5 C6B G
l  ” t
.
dering with Existing Hardware
er writes by flushing cachelines via CLFLUSH
CLFLUSH:
talls the CPU pipeline and serializes execution
STORE data[0] = 0xFOOD
STORE data[1] = 0xBEEF
CLFLUSH data[0]
CLFLUSH data[1]
STORE valid = 1
ata[0] ST CLFLUSH
CLFLUSHOPT
• Provides unordered version of CLFLUSH
• Supports efficient cache flushing
data[1]
valid
ST CLFLUSHOPT
ST
data[0] ST CLFLUSHOPT
time
data[1]
valid
ST CLFLUSH
ST
data[0] ST CLFLUSH
time 20
M FdWTa 4FC BF( N
aOZZ
  h(
( , 67
X
/
F]b QS0 GEF ( )
l  u x p p p
r y
l  – ~∼
k  y •
( , 67
X
l  AI T WS RZf OQQS OaaS
l  p
k  J WaS SOR ~∼–
v •
l  v
k  BF
k  v
(
MRAMDRAM
VM
$
VM $
$
ff
v
VMVMVM
( , 67
S T
a
l  )
k  s
x yo o o o
k  p – fi
l  • p
l  ff ~∼
k  Q T Gb P] 5]] a
(
•~∼
p –
v
( , 67
l  C OdO S 9WZS Ff aS[
k  C 9F M8b ]Ff N F6 9F MF6 N
5C9F MFBFC /N
l  C WP O f
k  AI SO M4FC BF N 677F MHF8A K N
S[] f S M4FC BF N 4aZO MBBCF 4 N
l  7OaOPO S
k  9B87HF MF : B7 N
l  a b[S aOaW]
k  AI F64I8A:8E M C7CF (N
((
( , 67
l  4FC BF ( Gba] WOZ
k  C ]U O[[W U O R H OUS ]RSZ T] A] I]ZOaWZS
S[] f
k  Vaa 0 S SO QV Q dW Q SRb ] O aba] WOZ
l  6EB ( Gba] WOZ
k  7OaOQS aS FW[bZOaW] SaV]R]Z]UWS
k  Vaa 0 S] ZS Rb S SRb hPQZ aba] WOZ R [
()
( , 67
C
l  W be
k  7E4 –~∼
k 
–~∼
k  K C   SKSQbaW] CZOQS
l  74K   7W SQa 4QQS
k  AIE4 v
–~∼
k  K C
(
Linux – PDA
Agenda VR3 (2001)

More Related Content

PDF
USENIX NSDI 2016 (Session: Resource Sharing)
Ryousei Takano
 
PDF
Exploring the Performance Impact of Virtualization on an HPC Cloud
Ryousei Takano
 
PDF
User-space Network Processing
Ryousei Takano
 
PDF
IEEE CloudCom 2014参加報告
Ryousei Takano
 
PDF
Flow-centric Computing - A Datacenter Architecture in the Post Moore Era
Ryousei Takano
 
PDF
Expectations for optical network from the viewpoint of system software research
Ryousei Takano
 
PDF
クラウド環境におけるキャッシュメモリQoS制御の評価
Ryousei Takano
 
PDF
From Rack scale computers to Warehouse scale computers
Ryousei Takano
 
USENIX NSDI 2016 (Session: Resource Sharing)
Ryousei Takano
 
Exploring the Performance Impact of Virtualization on an HPC Cloud
Ryousei Takano
 
User-space Network Processing
Ryousei Takano
 
IEEE CloudCom 2014参加報告
Ryousei Takano
 
Flow-centric Computing - A Datacenter Architecture in the Post Moore Era
Ryousei Takano
 
Expectations for optical network from the viewpoint of system software research
Ryousei Takano
 
クラウド環境におけるキャッシュメモリQoS制御の評価
Ryousei Takano
 
From Rack scale computers to Warehouse scale computers
Ryousei Takano
 

What's hot (20)

PDF
Iris: Inter-cloud Resource Integration System for Elastic Cloud Data Center
Ryousei Takano
 
PDF
HPC Cloud: Clouds on supercomputers for HPC
Ryousei Takano
 
PDF
Bruno Silva - eMedLab: Merging HPC and Cloud for Biomedical Research
Danny Abukalam
 
PDF
Stig Telfer - OpenStack and the Software-Defined SuperComputer
Danny Abukalam
 
PPTX
Exascale Capabl
Sagar Dolas
 
PDF
Evolving Virtual Networking with IO Visor
Larry Lang
 
PDF
On heap cache vs off-heap cache
rgrebski
 
PDF
Hands on MapR -- Viadea
viadea
 
PPTX
GPGPU programming with CUDA
Savith Satheesh
 
PDF
SQL+GPU+SSD=∞ (English)
Kohei KaiGai
 
PDF
POWER10 innovations for HPC
Ganesan Narayanasamy
 
PDF
dCUDA: Distributed GPU Computing with Hardware Overlap
inside-BigData.com
 
PDF
Programming Trends in High Performance Computing
Juris Vencels
 
PDF
pgconfasia2016 plcuda en
Kohei KaiGai
 
PDF
Achitecture Aware Algorithms and Software for Peta and Exascale
inside-BigData.com
 
PDF
Designing High Performance Computing Architectures for Reliable Space Applica...
Fisnik Kraja
 
PDF
20160407_GTC2016_PgSQL_In_Place
Kohei KaiGai
 
PDF
LizardFS-WhitePaper-Eng-v4.0 (1)
Pekka Männistö
 
PPTX
Sun jdk 1.6 gc english version
bluedavy lin
 
PDF
Lustre Generational Performance Improvements & New Features
inside-BigData.com
 
Iris: Inter-cloud Resource Integration System for Elastic Cloud Data Center
Ryousei Takano
 
HPC Cloud: Clouds on supercomputers for HPC
Ryousei Takano
 
Bruno Silva - eMedLab: Merging HPC and Cloud for Biomedical Research
Danny Abukalam
 
Stig Telfer - OpenStack and the Software-Defined SuperComputer
Danny Abukalam
 
Exascale Capabl
Sagar Dolas
 
Evolving Virtual Networking with IO Visor
Larry Lang
 
On heap cache vs off-heap cache
rgrebski
 
Hands on MapR -- Viadea
viadea
 
GPGPU programming with CUDA
Savith Satheesh
 
SQL+GPU+SSD=∞ (English)
Kohei KaiGai
 
POWER10 innovations for HPC
Ganesan Narayanasamy
 
dCUDA: Distributed GPU Computing with Hardware Overlap
inside-BigData.com
 
Programming Trends in High Performance Computing
Juris Vencels
 
pgconfasia2016 plcuda en
Kohei KaiGai
 
Achitecture Aware Algorithms and Software for Peta and Exascale
inside-BigData.com
 
Designing High Performance Computing Architectures for Reliable Space Applica...
Fisnik Kraja
 
20160407_GTC2016_PgSQL_In_Place
Kohei KaiGai
 
LizardFS-WhitePaper-Eng-v4.0 (1)
Pekka Männistö
 
Sun jdk 1.6 gc english version
bluedavy lin
 
Lustre Generational Performance Improvements & New Features
inside-BigData.com
 
Ad

Similar to クラウド時代の半導体メモリー技術 (20)

PDF
Current and Future of Non-Volatile Memory on Linux
mountpoint.io
 
PPTX
The Forefront of the Development for NVDIMM on Linux Kernel
Yasunori Goto
 
PPTX
IMC Summit 2016 Keynote - Arthur Sainio - NVDIMM: Changes are Here So What’s ...
In-Memory Computing Summit
 
PDF
HKG15-The Machine: A new kind of computer- Keynote by Dejan Milojicic
Linaro
 
PDF
IMCSummit 2015 - Day 2 Developer Track - The NVM Revolution
In-Memory Computing Summit
 
PDF
What every-programmer-should-know-about-memory
xan peng
 
PPTX
IMC Summit 2016 Breakout - Ken Gibson - The In-Place Working Storage Tier
In-Memory Computing Summit
 
PPTX
Developing Software for Persistent Memory / Willhalm Thomas (Intel)
Ontico
 
PDF
What Every Programmer Should Know About Memory
Ying wei (Joe) Chou
 
PDF
Software Design for Persistent Memory Systems
C4Media
 
PPTX
Introduction to Memory-Style Storage in Linux
Clay (Chih-Hao) Chang
 
PPTX
Persistent memory
Benoit Hudzia
 
PDF
The Forefront of the Development for NVDIMM on Linux Kernel (Linux Plumbers c...
Yasunori Goto
 
PDF
C++ Programming and the Persistent Memory Developers Kit
Intel® Software
 
PDF
Towards Application Driven Storage
Javier González
 
PDF
Analytics, Big Data and Nonvolatile Memory Architectures – Why you Should Car...
StampedeCon
 
PPT
Cache memory
MohanChimanna
 
PDF
NVMe Takes It All, SCSI Has To Fall
inside-BigData.com
 
PDF
Memory management
Adrien Mahieux
 
PPTX
lecture asdkvakm;bk;dv;advvAVHD;KASV;DVKHSVDK
officeaiotfab
 
Current and Future of Non-Volatile Memory on Linux
mountpoint.io
 
The Forefront of the Development for NVDIMM on Linux Kernel
Yasunori Goto
 
IMC Summit 2016 Keynote - Arthur Sainio - NVDIMM: Changes are Here So What’s ...
In-Memory Computing Summit
 
HKG15-The Machine: A new kind of computer- Keynote by Dejan Milojicic
Linaro
 
IMCSummit 2015 - Day 2 Developer Track - The NVM Revolution
In-Memory Computing Summit
 
What every-programmer-should-know-about-memory
xan peng
 
IMC Summit 2016 Breakout - Ken Gibson - The In-Place Working Storage Tier
In-Memory Computing Summit
 
Developing Software for Persistent Memory / Willhalm Thomas (Intel)
Ontico
 
What Every Programmer Should Know About Memory
Ying wei (Joe) Chou
 
Software Design for Persistent Memory Systems
C4Media
 
Introduction to Memory-Style Storage in Linux
Clay (Chih-Hao) Chang
 
Persistent memory
Benoit Hudzia
 
The Forefront of the Development for NVDIMM on Linux Kernel (Linux Plumbers c...
Yasunori Goto
 
C++ Programming and the Persistent Memory Developers Kit
Intel® Software
 
Towards Application Driven Storage
Javier González
 
Analytics, Big Data and Nonvolatile Memory Architectures – Why you Should Car...
StampedeCon
 
Cache memory
MohanChimanna
 
NVMe Takes It All, SCSI Has To Fall
inside-BigData.com
 
Memory management
Adrien Mahieux
 
lecture asdkvakm;bk;dv;advvAVHD;KASV;DVKHSVDK
officeaiotfab
 
Ad

More from Ryousei Takano (20)

PDF
Error Permissive Computing
Ryousei Takano
 
PDF
Opportunities of ML-based data analytics in ABCI
Ryousei Takano
 
PDF
ABCI: An Open Innovation Platform for Advancing AI Research and Deployment
Ryousei Takano
 
PDF
ABCI Data Center
Ryousei Takano
 
PDF
A Look Inside Google’s Data Center Networks
Ryousei Takano
 
PDF
AIST Super Green Cloud: lessons learned from the operation and the performanc...
Ryousei Takano
 
PDF
不揮発メモリとOS研究にまつわる何か
Ryousei Takano
 
PDF
High-resolution Timer-based Packet Pacing Mechanism on the Linux Operating Sy...
Ryousei Takano
 
PDF
クラウドの垣根を超えた高性能計算に向けて~AIST Super Green Cloudでの試み~
Ryousei Takano
 
PDF
高性能かつスケールアウト可能なHPCクラウド AIST Super Green Cloud
Ryousei Takano
 
PDF
IEEE/ACM SC2013報告
Ryousei Takano
 
PDF
A Scalable and Distributed Electrical Power Monitoring System Utilizing Cloud...
Ryousei Takano
 
PDF
伸縮自在なデータセンターを実現するインタークラウド資源管理システム
Ryousei Takano
 
PDF
SoNIC: Precise Realtime Software Access and Control of Wired Networks
Ryousei Takano
 
PDF
異種クラスタを跨がる仮想マシンマイグレーション機構
Ryousei Takano
 
PDF
動的ネットワーク切替を用いた省電力指向トラフィックオフロード方式
Ryousei Takano
 
PDF
Ninja Migration: An Interconnect transparent Migration for Heterogeneous Data...
Ryousei Takano
 
PDF
インタークラウドにおける仮想インフラ構築システム
Ryousei Takano
 
PDF
Preliminary Experiment of Disaster Recovery based on Interconnect-transparent...
Ryousei Takano
 
PDF
動的ネットワークパス構築と連携したエッジオーバレイ帯域制御
Ryousei Takano
 
Error Permissive Computing
Ryousei Takano
 
Opportunities of ML-based data analytics in ABCI
Ryousei Takano
 
ABCI: An Open Innovation Platform for Advancing AI Research and Deployment
Ryousei Takano
 
ABCI Data Center
Ryousei Takano
 
A Look Inside Google’s Data Center Networks
Ryousei Takano
 
AIST Super Green Cloud: lessons learned from the operation and the performanc...
Ryousei Takano
 
不揮発メモリとOS研究にまつわる何か
Ryousei Takano
 
High-resolution Timer-based Packet Pacing Mechanism on the Linux Operating Sy...
Ryousei Takano
 
クラウドの垣根を超えた高性能計算に向けて~AIST Super Green Cloudでの試み~
Ryousei Takano
 
高性能かつスケールアウト可能なHPCクラウド AIST Super Green Cloud
Ryousei Takano
 
IEEE/ACM SC2013報告
Ryousei Takano
 
A Scalable and Distributed Electrical Power Monitoring System Utilizing Cloud...
Ryousei Takano
 
伸縮自在なデータセンターを実現するインタークラウド資源管理システム
Ryousei Takano
 
SoNIC: Precise Realtime Software Access and Control of Wired Networks
Ryousei Takano
 
異種クラスタを跨がる仮想マシンマイグレーション機構
Ryousei Takano
 
動的ネットワーク切替を用いた省電力指向トラフィックオフロード方式
Ryousei Takano
 
Ninja Migration: An Interconnect transparent Migration for Heterogeneous Data...
Ryousei Takano
 
インタークラウドにおける仮想インフラ構築システム
Ryousei Takano
 
Preliminary Experiment of Disaster Recovery based on Interconnect-transparent...
Ryousei Takano
 
動的ネットワークパス構築と連携したエッジオーバレイ帯域制御
Ryousei Takano
 

Recently uploaded (20)

PDF
Trying to figure out MCP by actually building an app from scratch with open s...
Julien SIMON
 
PPTX
cloud computing vai.pptx for the project
vaibhavdobariyal79
 
PDF
Economic Impact of Data Centres to the Malaysian Economy
flintglobalapac
 
PDF
CIFDAQ's Market Wrap : Bears Back in Control?
CIFDAQ
 
PDF
Oracle AI Vector Search- Getting Started and what's new in 2025- AIOUG Yatra ...
Sandesh Rao
 
PDF
Presentation about Hardware and Software in Computer
snehamodhawadiya
 
PDF
Using Anchore and DefectDojo to Stand Up Your DevSecOps Function
Anchore
 
PDF
How Open Source Changed My Career by abdelrahman ismail
a0m0rajab1
 
PPTX
New ThousandEyes Product Innovations: Cisco Live June 2025
ThousandEyes
 
PDF
NewMind AI Weekly Chronicles - July'25 - Week IV
NewMind AI
 
PDF
The Future of Mobile Is Context-Aware—Are You Ready?
iProgrammer Solutions Private Limited
 
PDF
Automating ArcGIS Content Discovery with FME: A Real World Use Case
Safe Software
 
PPTX
AI in Daily Life: How Artificial Intelligence Helps Us Every Day
vanshrpatil7
 
PPTX
Introduction to Flutter by Ayush Desai.pptx
ayushdesai204
 
PDF
Accelerating Oracle Database 23ai Troubleshooting with Oracle AHF Fleet Insig...
Sandesh Rao
 
PDF
OFFOFFBOX™ – A New Era for African Film | Startup Presentation
ambaicciwalkerbrian
 
PDF
Research-Fundamentals-and-Topic-Development.pdf
ayesha butalia
 
PPTX
The-Ethical-Hackers-Imperative-Safeguarding-the-Digital-Frontier.pptx
sujalchauhan1305
 
PPTX
The Future of AI & Machine Learning.pptx
pritsen4700
 
PPTX
Agile Chennai 18-19 July 2025 | Emerging patterns in Agentic AI by Bharani Su...
AgileNetwork
 
Trying to figure out MCP by actually building an app from scratch with open s...
Julien SIMON
 
cloud computing vai.pptx for the project
vaibhavdobariyal79
 
Economic Impact of Data Centres to the Malaysian Economy
flintglobalapac
 
CIFDAQ's Market Wrap : Bears Back in Control?
CIFDAQ
 
Oracle AI Vector Search- Getting Started and what's new in 2025- AIOUG Yatra ...
Sandesh Rao
 
Presentation about Hardware and Software in Computer
snehamodhawadiya
 
Using Anchore and DefectDojo to Stand Up Your DevSecOps Function
Anchore
 
How Open Source Changed My Career by abdelrahman ismail
a0m0rajab1
 
New ThousandEyes Product Innovations: Cisco Live June 2025
ThousandEyes
 
NewMind AI Weekly Chronicles - July'25 - Week IV
NewMind AI
 
The Future of Mobile Is Context-Aware—Are You Ready?
iProgrammer Solutions Private Limited
 
Automating ArcGIS Content Discovery with FME: A Real World Use Case
Safe Software
 
AI in Daily Life: How Artificial Intelligence Helps Us Every Day
vanshrpatil7
 
Introduction to Flutter by Ayush Desai.pptx
ayushdesai204
 
Accelerating Oracle Database 23ai Troubleshooting with Oracle AHF Fleet Insig...
Sandesh Rao
 
OFFOFFBOX™ – A New Era for African Film | Startup Presentation
ambaicciwalkerbrian
 
Research-Fundamentals-and-Topic-Development.pdf
ayesha butalia
 
The-Ethical-Hackers-Imperative-Safeguarding-the-Digital-Frontier.pptx
sujalchauhan1305
 
The Future of AI & Machine Learning.pptx
pritsen4700
 
Agile Chennai 18-19 July 2025 | Emerging patterns in Agentic AI by Bharani Su...
AgileNetwork
 

クラウド時代の半導体メモリー技術

  • 1. ( , 67 ts Odʼ’a L P ( , 67
  • 2. ( , 67 l  s x– ~∼y l  s r v v l  s v p l  k  v x y o v k  C6 o ~∼ fi ~∼ v v v l  k  –BFp v (
  • 3. ( , 67 -‐‑‒ -‐‑‒ ʼ’mg i L e t ) DRAM(and(Flash(Scaling:(( “The(End(is(Nigh”( 1985( 1990( 1995( 2000( 2005( 2010( 2015( 2020( Density, DRAM( SLC(NAND( M FdWTa 4FC BF( N
  • 5. ( , 67 whg Ld l  v p – fi t l  ” ~∼“ fi k  p p p r k  l  l  p ~∼ k  p
  • 6. ( , 67 ts l  ~∼ p” t k  ]G p v t k  v t l  77EpA4A7 t k  v t l  v k  z v t ~∼ ~∼ k  ” ~∼ x ~∼y l  sA4A7 ,
  • 7. ( , 67 v x CH F8y V X oL oL -‐‑‒ 2015 2030 vv v hmd fl v v fi Lcndl v rr r rr r
  • 8. ( , 67 l  (, l  ) l  JSP v . Vaa 0 b Wa OW a U] X OW Z aO W[ bZ S W[ bZ S Va[Z
  • 10. ( , 67 : : I The$Machine$could$be$six$%mes$more$powerful$than$an$ equivalent$conven2onal$design,$while$using$just$1.25$ percent$of$the$energy$and$being$around$1/100$the$size. h:p://www.hpl.hp.com/research/systems@research/themachine/
  • 11. ( , 67 l  “ •~∼ k  B S 6][ baS C ]XSQa  9OQSP]] ” k  EOQ FQOZS 6][ baW U   aSZ k  8ea S[SZf FV W W U 6][ baW U   5 k  GVS OQVW S   C k  9W S5]e  H65 k  6GE 6] ] aWb[   G l  k  v x p p v y k  v k  v l  – p p p ~∼–
  • 12. ( , 67 ( IoTpCPSpHPCp p v •  v v v •  v v •  SoC (eFlash?) •  v v •  v v v v •  v •  DRAM
  • 13. ( , 67 gin a   k  (   k  Z]OR a] S k  v v )   k  v ) NVM)SSD)Challenges) •  So:ware)overheads)in) kernel) 7) 0 5 10 15 20 25 BaseLatency(us) PCM Ring DMA Wait Interrupt Issue Copy Schedule OS/User Software is Critical • Baseline Latencies: – Hardware: 8.2 us – Software: 13.4 us Hardware costs 11[Caulfield,)SC’10]) M6ObZjSZR F6 N v
  • 14. ( , 67 -‐‑‒ -‐‑‒ An application using NVM.FILE mode may or may not be using memory-mapped file behavior. The NVM.FILE mode describes NVM extensions including: • Discovery and use of atomic write features • The discovery of granularities (length or alignment characteristics) 4.3.3 NVM.PM.VOLUME mode overview NVM.PM.VOLUME mode describes the behavior for operating system components ( file systems) accessing persistent memory. NVM.PM.VOLUME mode provides a soft abstraction for Persistent Memory hardware and profiles functionality for operating sy components including: • the list of physical address ranges associated with each PM volume • the capability to determine whether PM errors have been reported Figure 5 NVM.PM.VOLUME and NVM.PM.FILE mode examples Application PM device PM device PM device. . . User space Kernel space MMU MappingsPM-aware file system NVM PM capable driver Load/ store Native file API PM-aware kernel module PM device NVM.PM.VOLUME mode NVM.PM.FILE mode 4.3.4 NVM.PM.FILE mode overview NVM.PM.FILE mode describes the behavior for applications accessing persistent me The commands implementing NVM.PM.FILE mode are similar to those using NVM.F l  AI 5Z]Q []RS l  AI 9WZS []RS l  CS W aS a S[] f I]Zb[S []RS l  CS W aS a S[] f 9WZS []RS Note that there are other models for connecting a non-PM file system to PM hardware. 4.3 NVM programming modes 4.3.1 NVM.BLOCK mode overview NVM.BLOCK and NVM.FILE modes are used when NVM devices provide block storage behavior to software (in other words, emulation of hard disks). The NVM may be exposed as a single or as multiple NVM volumes. Each NVM volume supporting these modes provides a range of logically-contiguous blocks. NVM.BLOCK mode is used by operating system components (for example, file systems) and by applications that are aware of block storage characteristics and the block addresses of application data. This specification does not document existing block storage software behavior; the NVM.BLOCK mode describes NVM extensions including: • Discovery and use of atomic write and discard features • The discovery of granularities (length or alignment characteristics) • Discovery and use of ability for applications or operating system components to mark blocks as unreadable Figure 4 NVM.BLOCK and NVM.FILE mode examples Application NVM block capable driver File system Application NVM device NVM device User space Kernel space Native file API NVM.BLOCK mode NVM.FILE mode 4.3.2 NVM.FILE mode overview NVM.FILE mode is used by applications that are not aware of details of block storage hardware or addresses. Existing applications written using native file I/O behavior should work FA 4 C ]U O[[W U ]RSZ I]Z
  • 15. ( , 67 l  v CS W aS a S[] f  C t l  – 4C p C fi p 4C k  [OZZ]Q – fSa O ]aVS [OZZ]Q l  v v t k  k  v
  • 16. ( , 67 Lcndl , 6CH AIE4 7E4 6CH AIE4 7E4 6CH AIE4  4 ES ZOQS RW  5 FVO SR ORR S OQS  6 8 aW SZf AIE4 5Z]Q 9 S[] f 9 QOQVS QOQVSQOQVS AIE4 B AIE4
  • 17. ( , 67 c mg M N l  AIE4 k  Fa] S • AIE4 ~∼ k  r r v -‐‑‒ • Recovery depends on write ordering CPU Write-back Cache NVM V D VD STORE data[0] = 0xFOOD STORE data[1] = 0xBEEF STORE valid = 1 Crash D D CPU Persistent Memory (PM) Ordering M FdWTa 4FC BF( N
  • 18. ( , 67 c mg MRN l  k  ”“ t k  aSZ CS W aS a S[] f v l  6 9 HF BCG l  6 J5 C6B G l  ” t . dering with Existing Hardware er writes by flushing cachelines via CLFLUSH CLFLUSH: talls the CPU pipeline and serializes execution STORE data[0] = 0xFOOD STORE data[1] = 0xBEEF CLFLUSH data[0] CLFLUSH data[1] STORE valid = 1 ata[0] ST CLFLUSH CLFLUSHOPT • Provides unordered version of CLFLUSH • Supports efficient cache flushing data[1] valid ST CLFLUSHOPT ST data[0] ST CLFLUSHOPT time data[1] valid ST CLFLUSH ST data[0] ST CLFLUSH time 20 M FdWTa 4FC BF( N aOZZ  h(
  • 19. ( , 67 X / F]b QS0 GEF ( ) l  u x p p p r y l  – ~∼ k  y •
  • 20. ( , 67 X l  AI T WS RZf OQQS OaaS l  p k  J WaS SOR ~∼– v • l  v k  BF k  v ( MRAMDRAM VM $ VM $ $ ff v VMVMVM
  • 21. ( , 67 S T a l  ) k  s x yo o o o k  p – fi l  • p l  ff ~∼ k  Q T Gb P] 5]] a ( •~∼ p – v
  • 22. ( , 67 l  C OdO S 9WZS Ff aS[ k  C 9F M8b ]Ff N F6 9F MF6 N 5C9F MFBFC /N l  C WP O f k  AI SO M4FC BF N 677F MHF8A K N S[] f S M4FC BF N 4aZO MBBCF 4 N l  7OaOPO S k  9B87HF MF : B7 N l  a b[S aOaW] k  AI F64I8A:8E M C7CF (N ((
  • 23. ( , 67 l  4FC BF ( Gba] WOZ k  C ]U O[[W U O R H OUS ]RSZ T] A] I]ZOaWZS S[] f k  Vaa 0 S SO QV Q dW Q SRb ] O aba] WOZ l  6EB ( Gba] WOZ k  7OaOQS aS FW[bZOaW] SaV]R]Z]UWS k  Vaa 0 S] ZS Rb S SRb hPQZ aba] WOZ R [ ()
  • 24. ( , 67 C l  W be k  7E4 –~∼ k  –~∼ k  K C  SKSQbaW] CZOQS l  74K  7W SQa 4QQS k  AIE4 v –~∼ k  K C ( Linux – PDA Agenda VR3 (2001)