IA-32 implementation AMD K7
General Details |
Name |
Athlon |
Codename |
Argon, K7 |
Family/Generation |
80686, 7th Generation, MMX, SSE (partial) |
Vendor |
AMD |
Manufacturer |
AMD |
First Introduction |
Oct 13, 1998 (MPF: 500 MHz Samples)
Jun 23, 1999 (500, 550, and 600 MHz)
Aug 9, 1999 (650 MHz)
|
Physical Details |
Package Type |
242 Contact Slot A Cartridge
575 Pin BGA
|
Package Size |
13.99 cm x 6.27 cm x 1.64 cm (Slot A Cartridge)
3.10 cm x 3.10 cm x 0.20 cm (BGA)
|
Socket or Slot |
Slot A |
Transistors |
22,000,000 (includes 152 KB L1 Cache) |
Process Technology |
6Mi, 0.25 µm, CMOS |
Die Size |
184 mm² |
Electrical Details |
Split Voltage |
N/A |
Core Voltage |
1.6 V |
L2 Voltage |
3.3 V
2.5 V
|
I/O Voltage |
1.6 V |
Typical Power |
500 MHz: 38 W
550 MHz: 41 W
600 MHz: 45 W
650 MHz: 48 W
|
Maximum Power |
500 MHz: 42 W
550 MHz: 46 W
600 MHz: 50 W
650 MHz: 54 W
|
Cooling |
Required |
Clock Frequencies |
CPU Core Speed |
500, 550, 600, 650 MHz |
L1 Cache Speed |
1.0x Core Speed |
L2 Cache Speed |
1/3x, 1/2x, or 2/3x Core Speed |
External Bus Speed |
100 MHz Double-Data Rate, DEC Alpha EV6 Protocol,
up to 24 Outstanding Transactions per Processor
|
Core/Bus Ratio |
5.0x, 5.5x, 6.0x, 6.5x |
Miscellaneous |
usual Motherboard |
Single or Dual Processor Slot A |
usual Chipset |
AMD 751/756, AMD 752/756, or non-AMD |
Pictures |
0.25 µm Die (364 KB JPG)
0.25 µm FPU Die (34 KB JPG)
BGA Top (27 KB JPG) and Bottom (68 KB JPG)
|
Processor Core |
Generic Details |
RISC, Out-of-order and Speculative Execution |
Specific Details |
72 MOP Entry ROB/ICU
18 MOP Entry Integer Scheduler
36 MOP Entry FP Scheduler
|
Registers |
32 Bit Integer, 80 Bit FP, 64 Bit MM
24 Entry 32 Bit 9xRead 8xWrite IFFRF (Integer RAT)
88 Entry 90 Bit 5xRead 5xWrite FPRF (FP RAT)
|
Pipeline Depth |
10 (Integer), 15 (FP) |
Instruction Decoder |
Direct and Vector Path
up to 3x IA-32/Cycle
up to 3x MOPs/Cycle
up to 6x ROPs/Cycle
|
Execution Units |
3x IEU (2x w/ MUL, 1x w/o MUL), 3x AGU (1x/IEU),
3x Pipelined FP (1x ADD, 1x MUL, 1x STORE)
|
Execution Speed |
up to 3x Integer MOPs/Cycle
up to 6x Integer ROPs/Cycle
up to 3x FP ROPs/Cycle
|
Processor Buses |
Address/Request Bus Width |
13 Bit |
Snoop Bus Width |
13 Bit |
Data Bus Width |
64+8 Bit, separate 64+8 Bit Backside L2 Cache Bus |
Physical Memory |
2^43 Bit = 8 TB |
Virtual Memory |
(8,190 + 8,192) x 4 GB = 65,528 GB (~64 TB) |
Multiprocessing |
SMP, up to 14 Processors |
Power Management |
HLT, STPCLK, SMI/SMM |
Processor Caches |
Level 0 |
N/A |
Level 1 |
Code |
64 KB, 2-Way, 64 Byte/Line, SI, LRU,
3 Pre-decode Bits/Byte (adds 24 KB)
|
Data |
64 KB, 2-Way, 64 Byte/Line, MOESI, LRU,
Dual-ported, Write-Allocate, Multi-banked
|
Level 2 |
Unified |
512 KB..8 MB, 64 Byte/Line
On-Die Tags for 512 KB
2-Way (512 KB, 1 MB, or 2 MB)
Direct-Mapped (4 MB and larger)
|
Processor Buffers |
Buffers |
2x I-miss Address
6x D-miss Address
8x Victim Address
8x Snoop Address
4x Write Address
|
Load/Store Queue |
12 Entries for L1 Cache
32 Entries for L2 Cache and Memory
|
Prefetch Queue |
16 Byte/Cycle from Fetch to Scan Stage
3x 8 Byte Queue in Direct Path Decoder
|
Branch Prediction |
Static |
Yes |
Dynamic |
2,048 Entry BHT, 2-Bit Smith Algorithm |
2,048 Entry BTAC |
RSB |
12 Entries |
TLB |
4 KB Code L1 |
16 Entries, Fully, LRU |
4/2 MB Code L1 |
8 Entries, Fully, LRU |
4 KB Data L1 |
24 Entries, Fully, LRU |
4/2 MB Data L1 |
8 Entries, 4-Way, LRU |
4 KB Code L2 |
256 Entries, 4-Way, LRU |
4/2 MB Code L2 |
n/a |
4 KB Data L2 |
256 Entries, 4-Way, LRU |
4/2 MB Data L2 |
n/a |
Instruction Set |
Regular |
IA-32 |
Floating Point |
Integrated |
Multi Media |
MMX, 3DNow!, Extended 3DNow!, SSE-MMX, SSE-MEM |
Processor Modes |
Real, Protected, Virtual, Paging, SMM |
|