AMD Opteron™ Processor
for MP Server Systems
Fred Weber
VP & CTO,
Computation P rodu ct s G roup
AMD
Microprocessor Forum 2002 2
Agenda
•AMD Opteron™ processor overview
•Contrast to exi sti ng MP sys tem topo l ogy
•Glueless MP system topology
Microprocessor Forum 2002 3
AMD Opteron™ Processor
Technology Overview
•Processor C ore Overview
– Support for AMD’s 64-bit technology
– 12-stage int, 17-stage fp pipelines
– Enhanced TLB structures
– TLB flush filter
– Enhanced bran ch prediction
– Large L2 cache (up to 1MB)
– ECC protection
•Memory Cont ro ller Over vi ew
– Dual-channel DDR memory
– PC2700, PC2100, or PC1600 DDR memory support
– Registered or Un buffered DIMMs
– ECC and Chip Kill
– High bandwidth (up to 5.3GB/s)
•HyperTransport™ Technology Overview
– One, two, or three links
– 2, 4, 8, 16, or 32-bits full duplex
– Up to 6.4 GB/s bandwidth per link
– 19.2 GB/s aggregate bandwidth
HT = HyperTranspo rt™ technology
XBAR
HT
HT
HT
MCT CPU
SRQ
DRAM
Microprocessor Forum 2002 4
AMD Opteron™ Processor
Glueless MP System Overview
HT = HyperTranspo rt™ technology
I/O
I/O
I/OI/O
non-Coherent
HyperTransport™ Link
I/O
Coherent
HyperTransport ™
XBAR
cHT
HT
cHT
MCT CPU
SRQ
DRAM
XBAR
HT
cHT
cHT
CPU
SRQ
DRAM
MCT
XBAR
HT
cHT
cHT
CPU
SRQ
MCT
DRAM
XBAR
cHT
HT
cHT
CPU
SRQ
MCT
DRAM
Microprocessor Forum 2002 5
Existing MP System Topology
System
Logic
System
Logic
CPU CPU
I/O
DRAM
Gfx
2.1 GB/s each
4.2 GB/s total
528 MB/s total 2.1 GB/s total
System
Logic
System
Logic
CPU CPUCPU CPU
I/O
DRAM
3.2 GB/s total
132 MB/s total
I/O
I/O
I/O
1.6 GB/s each
4.8 GB/s total
6.4 GB/s total
AMD Intel
Microprocessor Forum 2002 6
P0
P3
P1
P2
0-
hop
P0
P3
P1
P2
1-hop
Local vs. Remote Memory Access
•0 Hop: Local Memory Access
•1 Hop: Remote 1 Memory Access
•2 Hop: Remote 2 Memory Access
•Diameter: maximum hop count between any pair of nodes
•Average distance: average hop count between nodes
P0
P3
P1
P2
2-hops
Microprocessor Forum 2002 7
Local vs. Crossfire Memory Bandwidth
•Local memory access bandwidth
– Each processor reads data from
its own lo cal m emo ry
•Xfire memory access bandwidth
– All processors read data from
memory at all nodes
P2 P3
P1
P0
Local
BW
P2 P3
P0 P1
Xfire BW
Microprocessor Forum 2002 8
Single Processor Population
System Parameters:
•8 DIMMs (up to 16 GB using 256Mb DRAM)
•2 HyperTransport™ links available for I/O
•Processor-to-Memory Read Bandwidth = 5.3 GB/s
•I/O Bandwidth = 6.4 GB/s (per l ink)
AMD Opteron™
AMD Opteron™
DDR
Mem
DDR
Mem
I/O
I/O
I/O
I/O
AMD Opteron™
(Proc 0)
AMD Opteron™
(Proc 0)
Microprocessor Forum 2002 9
SPEC
®
CPU 2000
•System Configurations
–AMD Opteron™ processor operating at 2.0GHz
–Registered PC2700 DDR memory
•SPECint®2000:
–Estimated base score = 1202
•SPECfp®2000:
–Estimated base score = 1170
Microprocessor Forum 2002 10
Dual Processor System Topology
System Parameters:
• 16 DIMMs (up to 32 GB using 256Mb DRAM)
• 4 HyperTransport™ link s available for I/O
• Bisection-bandwidth = 6.4GB/s
• Dia meter = 1, Avg distance=0.5
• Local Memory Read Bandwidth = 10.67 GB/s
– Local Bandwidth/processor = 5.3 GB/s
• Xfire Memory Read Bandwidth = 7.06 GB/s
– Xfire Bandwidth/processor = 3.53 GB/s
AMD Opteron™
AMD Opteron™
DDR
Mem
DDR
Mem AMD Opteron™
(Proc 0)
AMD Opteron™
(Proc 0)
I/O
I/O
AMD Opteron™
AMD Opteron™ DDR
Mem
DDR
Mem
AMD Opteron
(Proc 1)
AMD Opteron
(Proc 1)
I/O
I/O
Bisection plane
I/O
I/O I/O
I/O
Microprocessor Forum 2002 11
Quad Processor System Topology
System Parameters:
• 32 DIMMs (up to 64 GB using 256Mb DRAM)
• 4 HyperTransport™ link s available for I/O
• Bisection-bandwidth = 12.8GB/s
• Dia meter = 2, Avg distance = 1
• Local Memory Read Bandwidth = 15.59 GB/s
– Loca l Bandwidth/processor = 3.9 GB/s
• Xfire Memory Read Bandwidth = 11.23 GB/s
– Xfire Bandwidth/processor = 2.8 GB/s
AMD Opteron™
AMD Opteron™
DDR
Mem
DDR
Mem AMD Opteron™
(Proc 0)
AMD Opteron™
(Proc 0)
I/O
I/O
AMD Opteron™
AMD Opteron™ DDR
Mem
DDR
Mem
AMD Opteron
(Proc 1)
AMD Opteron
(Proc 1)
I/O
I/O
AMD Opteron™
AMD Opteron™
DDR
Mem
DDR
Mem AMD Opteron
(Proc 2)
AMD Opteron
(Proc 2)
I/O
I/O
AMD Opteron™
AMD Opteron™ DDR
Mem
DDR
Mem
AMD Opteron
(Proc 3)
AMD Opteron
(Proc 3)
I/O
I/O
Bisection plane
Microprocessor Forum 2002 12
MP System Scalability
Memory Bandwidth
Memory Bandwidth Scalability
0
2
4
6
8
10
12
14
16
18
1P 2P 4P
Number of Processors in System
GB/s
Local B/W
Xfire B/W
Microprocessor Forum 2002 13
Summary
•The AMD Opteron™ processor is designed to provide industry
leading performance for enterprise class servers
– 32-bit performance leader ship substantiated by delivering on AMD’s promise
of nearly doubling x86-based SPEC
®
CPU perfor mance from a year ago
– Simulta neous 32 and 64-bit performance
•AMD Opteron " plumbing" is designed to provide exceptional
MP scalability
– Performance advantage grows versus competitive platforms
– Memory capacity and bandwidth scales
– I/O capacity and bandwidth increases
Microprocessor Forum 2002 14
Trademark Attribution
AMD, the AMD Arrow Logo, AMD Opteron and combination s
thereof are trademarks of Advanced Micro Devices, Inc.
HyperTransport is a licensed trademark of the
HyperTransport Consortium. Othe r product names used in
this presentation are for identification purposes only and may
be trademarks of their respective companies.
SPEC, SPECint, and SPECfp are registered trademarks of the
Standard Performance Evaluation Corporation (SPEC).
