vAlign=top width=475 colSpan=3>
align=center>AMD Architecture Comparison
|
vAlign=top width=139>
align=left>
|
vAlign=top width=168>
align=left>K8
|
vAlign=top width=168>
align=left>Barcelona
|
vAlign=top width=139>
align=left>SSE Execution Width
|
vAlign=top width=168>
align=left>64-bit
|
vAlign=top width=168>
align=left>128-bit
|
vAlign=top width=139>
align=left>Instruction Fetch Bandwidth
|
vAlign=top width=168>
align=left>16 bytes/cycle
|
vAlign=top width=168>
align=left>32 bytes/cycle
|
vAlign=top width=139>
align=left>Data Cache Bandwidth
|
vAlign=top width=168>
align=left>2 x 64-bit loads/cycle
|
vAlign=top width=168>
align=left>2 x 128-bit loads/cycle
|
vAlign=top width=139>
align=left>L2/Northbridge Bandwidth
|
vAlign=top width=168>
align=left>64 bits/cycle
|
vAlign=top width=168>
align=left>128 bits/cycle
|
vAlign=top width=139>
align=left>FP Scheduler Depth
|
vAlign=top width=168>
align=left>36 Dedicated x 64-bit ops
|
vAlign=top width=168>
align=left>36 Dedicated x 128-bit ops
|