Code name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
for value desktop systems
Willamette
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
Frontside bus
Memory addressability
Frontside bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Transistors
Package and connector
Frequency (MHz)
and available date
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
128KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.18u
~42 million with die size of 217 square millimeters
Flip-Chip Pin Grid Array-2 (FC-PGA2) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket; used with SDRAM-based chipset (such as 845 chipset)
(26INTEL) Compiled by Roger Dodson, IBM. June 2002
PC Processors (Celeron - Northwood)
®
Intel® Celeron
Code name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
for value desktop (and mobile) systems
Celeron Northwood
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
Frontside bus
Memory addressability
Frontside bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Package and connector
Frequency (MHz)
and available date
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
128KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.13u
Flip-Chip Pin Grid Array-2 (FC-PGA2) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket; used with SDRAM-based chipset (such as 845 chipset)
2.0GHz available September 2002
2.1GHz available November 2002
2.2GHz available November 2002
2.2GHz available June 2003 for mobile systems
2.3GHz available March 2003
2.3GHz available June 2003 for mobile systems
2.4GHz available March 2003
2.4GHz available June 2003 for mobile systems
2.5GHz available June 2003 for both desktop and mobile systems (transportable processors)
2.6GHz available June 2003 for both desktop and mobile systems (transportable processors)
2.7GHz available September 2003 for both desktop and mobile systems (transportable processors)
2.8GHz available November 2003 for both desktop and mobile systems (transportable processors)
Chipset support
Intel 845 and 865 desktop family and others
Intel 852GM, 852GME, 852PM mobile chipset
System bus - parity
System bus - speed
System bus - features
Bus architecture
Execution units
Pipeline stages
Supscal dispatch/execute
Superscalar issue
Superscalar retire
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Internal processing
External data bus
External address bus
User registers
Cache line size
Power management
Multiple processors
Technology (micron)
CPU voltage
Package type
®
Intel
Value mobile PC
IA-32 / P6 microarchitecture / CISC/RISC/micro-ops
MMX (57 new instructions) /
Streaming SIMD Extensions (70 new instructions)
16KB data; 16KB instruction
Write-back or thru (data); write-thru (instruction)
4-way set associative
64-bit / full speed / non-blocking
Parity in cache and internal registers
128KB / full speed
64-bit data path / ECC
8-way set associative / non-blocking
Integrated / unified (internal die; on die)
Write-through or write-back (programmable per line),
uncacheable, write-protect
ECC on system bus; parity on address bus (frontside)
133MHz frontside bus
Nonblocking cache hierarchy
Independent backside and frontside buses operate
concurrently / Dual Independent Bus Architecture
2 integer/MMX units; 1 floating pt unit;1 load unit; 1 store unit
Decoupled, 14 stage superpipelined
5 micro-ops per cycle (3 micro-ops is typical)
6 micro-ops per cycle (3 micro-ops is typical)
3 micro-ops per cycle
Yes (called dynamic execution)
Dynamic (based on history) / 512 entry BTB
Yes
Pipelined math coprocessor
32-bits (300 bit internal bus width) / 32-bit word size
64-bit system bus with ECC
36-bits (64GB physical address space; 64TB virtual)
8 GPR, 8 FP, 40 more GPR via register renaming
32 bytes (8 bytes x 4 chunks)
Quick Start and Deep Sleep
No SMP support
MMX (57 new instructions) /
Streaming SIMD Extensions (70 new instructions)
Same
Same
Same
Same
Same
256KB / full speed (Advanced Transfer Cache)
256-bit data path / quad-wide cache line / ECC
Intelligent buffering of read and stores (called
Advanced System Buffering with 4 writeback buffers,
6 fill buffers, 8 bus queue entries) / Data Prefetch Logic
8-way set associative
Integrated / unified (internal die; on die)
Write-through or write-back (programmable per line),
uncacheable, write-protect
Non-blocking / pipelined burst synchronous
Same
100MHz or 133MHz frontside bus
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
Same
0.13u
1.1 volts for Ultra Low Voltage processors
1.15 volts for Low Voltage processors
1.4 or 1.45 volts for others
Micro-Flip Chip Ball Grid Array (Micro-FCBGA)
Micro-Flip Chip Pin Grid Array (Micro-FCPGA)
The "A" is added to the "800A" in Micro-FCBGA and
Micro-FCPGA to distinguish it from the Mobile Intel
Celeron Processor 800MHz in Micro-BGA2 and
Micro-PGA2 packages
650/100MHz Ultra Low Voltage (January 2002)
650/100MHz Low Voltage (October 2001)
700MHz/100MHz Ultra Low Voltage (September 2002)
733MHz/133MHz Low Voltage (April 2002)
733MHz/133MHz Ultra Low Voltage (September 2002)
800MHz/133MHz Ultra Low Voltage (January 2003)
866MHz/133MHz Low Voltage (January 2003)
1GHz/133MHz (April 2002)
1.06GHz/133MHz (January 2002)
1.13GHz/133MHz (January 2002)
1.2GHz/133MHz (January 2002)
1.33GHz/133MHz (June 2002)
(20INTEL) Compiled by Roger Dodson, IBM. January 2003
PC Processors (Mobile Celeron - Northwood)
®
Mobile Intel® Celeron
Code name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
Power mgmt technology
for value mobile systems
Northwood
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
AutoHALT, Stop-Grant, Sleep, Deep Sleep
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
System bus
Memory addressability
System bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Voltage
Package and connector
Frequency (MHz)
and available date
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
256KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.13u
1.3 volts
Micro Flip-Chip Pin Grid Array (uFCPGA) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
1.26GHz available April 2003
1.4GHz available June 2002
1.5GHz available June 2002
1.6GHz available September 2002
1.7GHz available September 2002
1.8GHz available September 2002
2.0GHz available January 2003
2.2GHz available April 2003
2.4GHz available June 2003
2.5GHz available November 2003
Chipset support
Intel 845MZ with DDR-SDRAM memory
Intel 845MP with DDR-SDRAM memory
Intel 852GM, 852GME, 852PM with DDR-SDRAM memory
Other compatible chipsets
(27INTEL) Compiled by Roger Dodson, IBM. November 2003
PC Processors (Intel Celeron M)
®
Intel® Celeron
Code name
Messaging
Micro-architecture
MMX™ / Streaming SIMD
SSE2
Power mgmt technology
M processor for mobile systems
Banias Celeron or ICP-M
Based on an architecture designed specifically for mobile computing, the Intel Celeron M processor delivers a balanced
level of mobile processor technology and exceptional value in sleeker, lighter notebook designs
IA-32 / micro-op fusion, dedicated stack manager, advanced branch prediction, power-optimized processor system bus
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Auto Halt, Stop Grant, Deep Sleep, Deeper Sleep
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
L3 cache
System bus
Memory addressability
System bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Multiple processors
Technology (micron)
Package and connector
Frequency (MHz/GHz)
and available date
256-bit data path / full speed
32KB data cache / integrated
32KB instruction cache / integrated
512KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / 64 byte cache line size / 8-way set associative / integrated / unified (internal die; on die)
None
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Yes (out-of-order instruction execution)
Dynamic (based on history)
Yes (Advanced Dynamic Execution)
Pipelined floating point unit
Compatible with IA-32 software
No SMP support
0.13u
Micro Flip-Chip Pin Grid Array (Micro-FCPGA) requires 479-pin surface mount Zero Insertion Force (ZIF) socket
(mPGA479M socket) or Micro Flip-Chip Ball Grid Array (Micro-FCBGA) for surface mount (479-ball)
64-bit / full speed
16KB data; 16KB instruction / integrated / non-blocking
Write-back or thru (data); write-thru (instruction)
4 way set associative (data); 2 way set associative (instruction)
256 or 512KB / full speed (Advanced Transfer Cache)
256-bit data path / quad-wide cache line / ECC
Intelligent buffering of read and stores (called Advanced
System Buffering with 4 writeback buffers, 6 fill buffers,
8 bus queue entries)
8-way set associative
Integrated / unified (internal die; on die)
Write-through or write-back (programmable per line), uncacheable, write-protect
Non-blocking / pipelined burst synchronous
133MHz
64GB memory addressability
64-bit system bus with ECC
ECC on system bus; parity on address bus (frontside bus)
2 integer/MMX units; 1 floating point unit; 1 load unit; 1 store unit
5 micro-ops per cycle (3 micro-ops is typical); Pipeline stages: decoupled, 14 stage superpipelined
6 micro-ops per cycle (3 micro-ops is typical)
3 micro-ops per cycle
Yes (called dynamic execution)
Dynamic (based on history) / 512 entry BTB / typically
predicts 10 to 15 nested branches
L2 cache bus also called Backside Bus
Memory or system bus also called
Frontside Bus
Yes (typically 20 to 30 instructions beyond counter
with an average of 5 branches)
Pipelined math coprocessor
Processor serial number
Serial number
Bus architecture
Internal processing
User registers
Cache line size
Power management
Multiple processors
Technology (micron)
Package type
Connector
Frequency (MHz)
Chipset support
Server blade support
None
Unique processor serial number
Independent backside and frontside buses operate concurrently / Dual Independent Bus Architecture (DIB)
32-bits (300 bit internal bus width)
8 GPR, 8 FP, 8 FPscalar and SIMD, 40 more GPR via register renaming
32 bytes (8 bytes x 4 chunks); burst mode bus of addr-data-data-data
System Management Mode (SMM)
Some support 2-way SMP with appropriate chipset support
Frontside bus - speed
Memory addressability
System bus - width
System bus - parity
Execution units
Supscal dispatch/execute
Superscalar issue/retire
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Serial number
Bus architecture
Internal processing
User registers
Cache line size
Power management
Multiple processors
Technology (micron)
Package type
Frequency (MHz)
Chipset support
Server blade support
III Processor-M for mobile systems (and server blade systems)
Tualatin (pronounced "TWO-ala-tin")
IA-32 / CISC/RISC/micro-ops
MMX (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Enhanced Intel SpeedStep™ technology
64-bit / full speed
16KB data; 16KB instruction / integrated / non-blocking
Write-back or thru (data); write-thru (instruction)
4 way set associative (data); 2 way set associative (instruction)
512KB / full speed (Advanced Transfer Cache) / integrated / unified (internal die; on die)
256-bit data path / quad-wide cache line / ECC
Intelligent buffering of read and stores (called Advanced System Buffering with 4 writeback buffers, 6 fill buffers,
8 bus queue entries) / Data Prefetch Logic
8-way set associative / non-blocking / pipelined burst synchronous
Write-through or write-back (programmable per line), uncacheable, write-protect
133MHz (some at 100MHz)
64GB memory addressability
64-bit system bus with ECC
ECC on system bus; parity on address bus (frontside bus)
2 integer/MMX units; 1 floating point unit; 1 load unit; 1 store unit
5 micro-ops per cycle (3 micro-ops is typical); Pipeline stages: decoupled, 14 stage superpipelined
Issues 6 micro-ops per cycle (3 micro-ops is typical) / retires 3 micro-ops per cycle
Yes (called dynamic execution)
Dynamic (based on history) / 512 entry BTB / typically predicts 10 to 15 nested branches
Yes (typically 20 to 30 instructions beyond counter with an average of 5 branches)
Pipelined math coprocessor
Unique processor serial number
Independent backside and frontside buses operate concurrently / Dual Independent Bus Architecture (DIB)
32-bits (300 bit internal bus width)
8 GPR, 8 FP, 8 FPscalar and SIMD, 40 more GPR via register renaming
32 bytes (8 bytes x 4 chunks); burst mode bus of addr-data-data-data
Quick Start, Deep Sleep, Deeper Sleep
No SMP support (2-way SMP for 800MHz Low Voltage for server blade systems with ServerWorks
0.13u (130-nanometer)
Micro-FCPGA (Flip-Chip Pin Grid Array) for socketable boards
Micro-FCBGA (Flip-Chip Ball Grid Array) for surface mount boards
Frontside Maximum PerformanceBattery OptimizedAnnounce
busMode Mode date
700MHz Ultra Low Voltage* 100MHz700MHz at 1.1V300MHz at 0.95VOctober 2001/Nov 2001*
733MHz Low Voltage133MHz733MHz at 1.15V466MHz at 1.05VOctober 2001
750MHz Ultra Low Voltage 100MHz750MHz at 1.1V350MHz at 0.95VJanuary 2002
750MHz Low Voltage100MHz750MHz at 1.15V450MHz at 1.05VOctober 2001
800A MHz Low Voltage100MHz800MHz at 1.15V500MHz at 1.05VOctober 2001
800MHz Low Voltage**133MHz800MHz at 1.15V533MHz at 1.05VOctober 2001/Mar 2002**
800MHz Ultra Low Voltage* 100MHz800MHz at 1.15V400MHz at 1.05VApril 2002*
800MHz Ultra Low Voltage* 133MHz800MHz at 1.15V400MHz at 1.05VApril 2002*
850MHz Low Voltage133MHz850MHz at 1.15V500MHz at 1.05VJanuary 2002
850MHz Ultra Low Voltage 100MHz850MHz at 1.1V400MHz at 0.95VSeptember 2002
866MHz Low Voltage133MHz866MHz at 1.15V533MHz at 1.05VJanuary 2002
866MHz133MHz866MHz at 1.40V667MHz at 1.15VJuly 2001
866MHz Ultra Low Voltage 133MHz866MHz at 1.1V400MHz at 0.95VSeptember 2002
900MHz Ultra Low Voltage 100MHz900MHz at 1.1V400MHz at 0.95VJanuary 2003
933MHz Ultra Low Voltage 133MHz933MHz at 1.1V400MHz at 0.95VJanuary 2003
933MHz Low Voltage133MHz933MHz at 1.15V533MHz at 1.05VApril 2002
933MHz133MHz933MHz at 1.40V733MHz at 1.15VJuly 2001
1GHz133MHz1GHz at 1.40V733MHz at 1.15VJuly 2001
1GHz Low Voltage133MHz1GHz at 1.15V533MHz at 1.05VSeptember 2002
1.06GHz133MHz1.06GHz at 1.40V733MHz at 1.15VJuly 2001
1.13GHz133MHz1.13GHz at 1.40V733MHz at 1.15VJuly 2001
1.2GHz133MHz1.2GHz at 1.40V800MHz at 1.15VOctober 2001
1.26GHz133MHz1.2GHz at 1.40V800MHz at 1.15VSeptember 2002
1.33GHz133MHz1.2GHz at 1.40V800MHz at 1.15VSeptember 2002
Intel 830MP, 830M, 830MG and others
* Supported in server blade systems; Micro-FCBGA only; uses Intel 440GX chipset
** Announced March 2002 for server blade systems; Micro-FCBGA only; supports 2-way SMP in server blade
systems with ServerWorks ServerSet III LE chipset
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L2 cache bus also called Backside Bus
Memory or system bus also called
System bus
Memory addressability
System bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Package and connector
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
512KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die)
None
400 or 533MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.13u
400MHz:
Micro Flip-Chip Pin Grid Array (uFCPGA) requires 478-pin surface mount Zero Insertion Force (ZIF)
socket
533MHz:
Micro Flip-Chip Pin Grid Array (uFCPGA2) requires 478-pin surface mount Zero Insertion Force (ZIF)
socket
SystemMaximum Performance Battery Optimized ModeHyper-Threading (HT) Announce
bus Mode Technology date
Frequency (MHz)
1.4GHz 400MHz1.4GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)April 2002
1.5GHz 400MHz1.5GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)April 2002
1.6GHz 400MHz1.6GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)March 2002
1.7GHz 400MHz1.7GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)March 2002
1.8GHz 400MHz1.8GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)April 2002
1.9GHz 400MHz1.9GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)June 2002
2.0GHz 400MHz2.0GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)June 2002
2.2GHz 400MHz2.2GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)Sept 2002
2.4GHz 400MHz2.4GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)Jan 2003
2.5GHz 400MHz2.5GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)April 2003
2.6GHz 400MHz2.6GHz at 1.3 volts 1.2GHz at 1.2v (<2 watts avg power)June 2003
2.40GHz 533MHz2.40GHz at 1.525 v 1.6GHz at 1.2v(<4.5 watts avg power)June 2003
2.66GHz 533MHz2.66GHz at 1.525 v 1.6GHz at 1.2v(<4.5 watts avg power)June 2003
2.80GHz 533MHz2.80GHz at 1.525 v 1.6GHz at 1.2v (<4.5 watts avg power)June 2003
3.06GHz 533MHz3.06GHz at 1.525 v 1.6GHz at 1.2v (<4.5 watts avg power)June 2003
2.66GHz 533MHz2.66GHz at 1.525v 1.6GHz at 1.2v (<3.0 watts avg power) with Hyper-Threading Sept 2003
2.80GHz 533MHz2.80GHz at 1.525v 1.6GHz at 1.2v (<3.0 watts avg power) with Hyper-Threading Sept 2003
3.06GHz 533MHz3.06GHz at 1.55v1.6GHz at 1.2v (<3.0 watts avg power) with Hyper-Threading Sept 2003
3.2GHz 533MHz3.2GHz at 1.55 volts 1.6GHz at 1.2v (<3.0 watts avg power) with Hyper-Threading Sept 2003
Chipset support
Intel 845MZ with DDR-SDRAM memory
Intel 845MP with DDR-SDRAM memory
Intel 852GM, 852GME, 852PM with DDR-SDRAM memory
(24INTEL) Compiled by Roger Dodson, IBM. September 2003
PC Processors (Intel Pentium M)
®
Intel® Pentium
Code name
Branding
Micro-architecture
MMX™ / Streaming SIMD
SSE2
Power mgmt technology
M processor for mobile systems
Banias
Part of the Intel Centrino™ mobile technology when included with an Intel 855 family chipset and Intel PRO/Wireless
Network Connection wireless chip
IA-32 / micro-op fusion, dedicated stack manager, advanced branch prediction, power-optimized processor system bus
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Enhanced Intel SpeedStep™ technology, Auto Halt, Stop Grant, Deep Sleep, Deeper Sleep
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
L3 cache
System bus
Memory addressability
System bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Multiple processors
Technology (micron)
Package and connector
Frequency (MHz/GHz)
and available date
256-bit data path / full speed
32KB data cache / integrated
32KB instruction cache / integrated
1MB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / 64 byte cache line size / 8-way set associative / integrated / unified (internal die; on die)
None
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Yes (out-of-order instruction execution)
Dynamic (based on history)
Yes (Advanced Dynamic Execution)
Pipelined floating point unit
Compatible with IA-32 software
No SMP support
0.13u
Micro Flip-Chip Pin Grid Array (Micro-FCPGA) requires 479-pin surface mount Zero Insertion Force (ZIF) socket
(mPGA479M socket) or Micro Flip-Chip Ball Grid Array (Micro-FCBGA) for surface mount (479-ball)
Highest Frequency Lowest FrequencyAnnounce
Mode Mode date
900MHz Ultra Low Voltage900MHz at 1.0 volts600MHz at 0.85 voltsMarch 2003
1.0GHz Ultra Low Voltage1.0GHz at 1.0 volts600MHz at 0.85 voltsJune 2003
1.1GHz Low Voltage1.1GHz at 1.18 volts 600MHz at 0.96 voltsMarch 2003
1.2GHz Low Voltage1.2GHz at 1.18 volts 600MHz at 0.96 voltsJune 2003
1.3GHz1.3GHz at 1.5 volts600MHz at 0.96 voltsMarch 2003
1.4GHz1.4GHz at 1.5 volts600MHz at 0.96 voltsMarch 2003
1.5GHz1.5GHz at 1.5 volts600MHz at 0.96 voltsMarch 2003
1.6GHz1.6GHz at 1.5 volts600MHz at 0.96 voltsMarch 2003
1.7GHz1.7GHz at 1.5 volts600MHz at 0.96 voltsJune 2003
Chipset support
Intel 855 chipset family with DDR-SDRAM memory
Other compatible chipsets
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
Front Side Bus
Memory addressability
Front Side Bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Transistors
Package and connector
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
256KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
400MHz (transfers data four times per clock) / address bus transfers at two times per clock / 64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.18u
~42 million with die size of 217 square millimeters
1. Pin Grid Array (PGA) r equires 423-pin Zero Insertion Force (ZIF) socket named Intel Socket 423 (PGA423);
used with RDRAM-based 850 chipset
2. Flip-Chip Pin Grid Array-2 (FC-PGA2) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket; used with SDRAM-based chipset (such as 845 chipset)
Frequency (MHz)
and available date
1.3GHz: 423-pin available January 2001
1.4GHz: 423-pin available November 2000
1.5GHz: 423-pin available November 2000, 478-pin available August 2001
1.6GHz: 423-pin available November 2000, 478-pin available August 2001
1.7GHz: 423-pin available November 2000, 478-pin available August 2001
1.8GHz: 423-pin available November 2000, 478-pin available August 2001
1.9GHz: 423-pin available November 2000, 478-pin available August 2001
2.0GHz: 423-pin available November 2000, 478-pin available August 2001
Chipset support
Intel 850 with dual channel RDRAM memory
Intel 845 with SDRAM memory
(14INTEL) Compiled by Roger Dodson, IBM. August 2001
PC Processors (Pentium 4 - Northwood)
®
Intel® Pentium
Code name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
Hyper-Threading
4 for high performance desktop systems
Northwood
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
3.06GHz with 533MHz and all 800MHz system bus processors:
for multi-threaded applications)
Personal Systems Reference (PSREF)
Created by IBM PC Institute
Hyper-Threading (HT) Technology (hardware support
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
L3 cache
System bus
Memory addressability
Frontside bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Transistors
Package and connector
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
512KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
None
400 or 533 or 800MHz (transfers data four times per clock) / address bus transfers at two times per clock /
64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200, 266, or 400MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.13u
~55 million
Flip-Chip Pin Grid Array-2 (FC-PGA2) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket
Frequency
and available date
1.6GHz sub-45W TDP (limited to under 45 watts thermal design point; for small form factor desktops); avail Jan 2002
1.8GHz sub-45W TDP (limited to under 45 watts thermal design point; for small form factor desktops); avail Jan 2002
2.0GHz sub-45W TDP (limited to under 45 watts thermal design point; for small form factor desktops); avail Jan 2002
1.8A GHz with 400MHz system bus: available July 2002
2.0A GHz with 400MHz system bus: available January 2002 ("A" signifies the 0.13 micron version, not 0.18 micron)
2.2GHzwith 400MHz system bus: available January 2002
2.26GHz with 533MHz system bus: available May 2002
2.4GHzwith 400MHz system bus: available April 2002
2.4B GHz with 533MHz system bus: available May 2002
2.4C GHz with 800MHz system bus: available May 2003with Hyper-Threading Technology
2.5GHzwith 400MHz system bus: available August 2002
2.53GHz with 533MHz system bus: available May 2002
2.6GHzwith 400MHz system bus: available August 2002
2.6C GHz with 800MHz system bus: available May 2003with Hyper-Threading Technology
2.66GHz with 533MHz system bus: available August 2002
2.8GHzwith 533MHz system bus: available August 2002
2.8C GHz with 800MHz system bus: available May 2003with Hyper-Threading Technology
3.0GHzwith 800MHz system bus: available April 2003with Hyper-Threading Technology
3.0GHzwith 400MHz system bus: available April 2003 (used in ThinkPad G40)
3.06GHz with 533MHz system bus: available November 2002 with Hyper-Threading Technology
3.2GHzwith 800MHz system bus: available June 2003with Hyper-Threading Technology
3.4GHzwith 800MHz system bus: available February 2004 with Hyper-Threading Technology
Chipset support
Intel 850 or 850E with dual channel RDRAM memory
Intel 845 with SDRAM or DDR-SDRAM memory
Intel 865 family with single or dual channel DDR-SDRAM memory (400, 533, or 800 MHz system bus)
Intel 875P with single or dual channel DDR-SDRAM memory (800 MHz system bus)
(22INTEL) Compiled by Roger Dodson, IBM. February 2004
PC Processors (Pentium 4 - Prescott)
®
Intel® Pentium
Code name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
SSE3
Hyper-Threading
4 for desktop systems
Prescott
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Streaming SIMD Extensions 3 (13 new instructions)
800MHz system bus processors:
Personal Systems Reference (PSREF)
Created by IBM PC Institute
Hyper-Threading (HT) Technology (hardware support for multi-threaded applications)
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
L3 cache
System bus
Memory addressability
System bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Other features
Technology
Package and connector
256-bit data path / full speed
16KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
1MB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
None
533 or 800MHz (transfers data four times per clock) / address bus transfers at two times per clock /
64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200, 266, or 400MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
Thermal monitoring, built-in self test, IEEE 1149.1 standard test access port and boundary scan
90nm (nanometer) or 0.09u (micron)
Flip-Chip Pin Grid Array (FC-mPGA4) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket
Frequency
and available date
2.80A GHz with 533MHz system busavailable February 2004
2.80E GHz with 800MHz system bus with Hyper-Threading Technology available February 2004
3.00E GHz with 800MHz system bus with Hyper-Threading Technology available February 2004
3.20E GHz with 800MHz system bus with Hyper-Threading Technology available February 2004
3.40E GHz with 800MHz system bus with Hyper-Threading Technology available February 2004
Chipset support
Intel 865 family with single or dual channel DDR-SDRAM memory
Intel 875P with single or dual channel DDR-SDRAM memory
(34INTEL) Compiled by Roger Dodson, IBM. February 2004
PC Processors (Pentium 4 Extreme Edition)
®
Intel® Pentium
Code name
Formal name
Micro-architecture
MMX™ / Streaming SIMD
SSE2
Hyper-Threading
4 Extreme Edition for high-end gamers and power users
None
Intel Pentium 4 Processor with HT Technology Extreme Edition
IA-32 / NetBurst™ (CISC/RISC/micro-ops) / 20 stage pipeline (Hyper-pipelined technology)
MMX™ (57 new instructions) / Streaming SIMD Extensions (70 new instructions)
Streaming SIMD Extensions 2 (144 new instructions)
Hyper-Threading (HT) Technology (hardware support for multi-threaded applications)
Personal Systems Reference (PSREF)
Created by IBM PC Institute
L1 cache - bus
L1 data cache
L1 instruction cache
L2 cache - size
L2 cache - data path
L3 cache
L3 cache - data path
System bus
Memory addressability
Frontside bus - width
Execution units
Out-of-order instructions
Branch prediction
Speculative execution
Math coprocessor
Compatibility
Cache line size
Multiple processors
Technology (micron)
Transistors
Package and connector
256-bit data path / full speed
8KB data cache / 4-way set associative / write-through / 64 byte cache line / integrated
Size not published / holds 12,000 micro-ops / 8-way set associative / integrated /
called Execution Trace Cache; caches decoded x86 instructions (micro-ops)
512KB / full speed (Advanced Transfer Cache)
256-bit data path (32 bytes) / transfers on each bus clock / 128 byte cache line size (usually divided into two 64 byte
sectors) / 8-way set associative / integrated / unified (internal die; on die) / ECC
2MB / full speed
256-bit data path (32 bytes) / transfers on each bus clock / 64 byte cache line size / 8-way set associative /
write-back / parity / integrated / unified (internal die; on die)
800MHz (transfers data four times per clock) / address bus transfers at two times per clock /
64 byte cache line size
64GB memory addressability / 36-bit addressing / address bus is double clocked at 200, 266, or 400MHz
64-bit data path
2 integer units; 1 floating point units; 1 load unit; 1 store unit
Two integer units (or Arithmetic Logic Units) run at two times core frequency (Rapid Execution Engine)
Yes
Dynamic (based on history) / 4KB Branch Target Buffer
Yes (Advanced Dynamic Execution)
Pipelined floating point unit / handles 128-bit floating point registers
Compatible with IA-32 software
128 bytes (32 bytes x 4 chunks); burst mode bus of addr-data-data-data
No SMP support
0.13u
~108 million
Flip-Chip Pin Grid Array-2 (FC-PGA2) requires 478-pin surface mount Zero Insertion Force (ZIF) socket
named mPGA478B socket
(32INTEL) Compiled by Roger Dodson, IBM. February 2004
This publication could include technical
inaccuracies or typographical errors.
References herein to IBM products and
services do not imply that IBM intends
to make them available in other
countries. IBM PROVIDES THIS
PUBLICATION AS IS WITHOUT
WARRANTY OF ANY KIND, EITHER
EXPRESS OR IMPLIED, INCLUDING
THE IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR
A PARTICULAR PURPOSE. Some
jurisdictions do not allow disclaimer of
express of implied warranties;
therefore this disclaimer may not apply
to you.
IBM is a registered trademarks of the
IBM Corporation in the United States or
other countries.
i486, Intel386, Intel486, IntelDX2, MMX,
Xeon, NetBurst, Centrino, and Itanuim
are trademarks and Intel, Pentium, and
Celeron are registered trademarks of
Intel.
Other trademarks and registered
trademarks are the properties of
others.