From WikiChip
Difference between revisions of "intrinsity/fastmath/fastmath-lp"
< intrinsity‎ | fastmath

m (Bot: moving all {{mpu}} to {{chip}})
 
(5 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
{{intrinsity title|FastMATH-LP}}
 
{{intrinsity title|FastMATH-LP}}
{{mpu
+
{{chip
 
| name                = FastMATH-LP
 
| name                = FastMATH-LP
| no image            = Yes
+
| no image            =  
| image              =  
+
| image              = fastmath-lp chip.gif
 
| image size          =  
 
| image size          =  
 
| caption            =  
 
| caption            =  
Line 10: Line 10:
 
| model number        = FastMATH-LP
 
| model number        = FastMATH-LP
 
| part number        =  
 
| part number        =  
| part number 1       =  
+
| part number 2       =  
 
| market              = Embedded
 
| market              = Embedded
 
| first announced    = 2002
 
| first announced    = 2002
Line 41: Line 41:
 
| thread count        = 1
 
| thread count        = 1
 
| max cpus            =  
 
| max cpus            =  
| max memory          = 1 GB
+
| max memory          = 1 GiB
 
| max memory addr    =  
 
| max memory addr    =  
  
| electrical          = Yes
+
 
 
| power              = 6 W
 
| power              = 6 W
 
| v core              = 0.85 V
 
| v core              = 0.85 V
Line 79: Line 79:
 
{{main|intrinsity/microarchitectures/fastmath#Memory_Hierarchy|l1=FastMATH § Cache}}
 
{{main|intrinsity/microarchitectures/fastmath#Memory_Hierarchy|l1=FastMATH § Cache}}
 
{{cache info
 
{{cache info
|l1i cache=16 KB
+
|l1i cache=16 KiB
|l1i break=1x16 KB
+
|l1i break=1x16 KiB
 
|l1i desc=256 blocks × 16 words/block
 
|l1i desc=256 blocks × 16 words/block
|l1i extra=
+
|l1d cache=16 KiB
|l1d cache=16 KB
+
|l1d break=1x16 KiB
|l1d break=1x16 KB
 
 
|l1d desc=256 blocks × 16 words/block
 
|l1d desc=256 blocks × 16 words/block
 
|l1d extra=write-through or write-back mode
 
|l1d extra=write-through or write-back mode
|l2 cache=1 MB
+
|l2 cache=1 MiB
|l2 break=1x1 MB
+
|l2 break=1x1 MiB
 
|l2 desc=4-way set associative
 
|l2 desc=4-way set associative
|l2 extra=(configurable as SRAM in 256 KB increments)
+
|l2 extra=(configurable as SRAM in 256 KiB increments)
|l3 cache=
 
|l3 break=
 
|l3 desc=
 
|l3 extra=
 
 
}}
 
}}
  

Latest revision as of 15:31, 13 December 2017

Edit Values
FastMATH-LP
fastmath-lp chip.gif
General Info
DesignerIntrinsity
ManufacturerTSMC
Model NumberFastMATH-LP
MarketEmbedded
Introduction2002 (announced)
2003 (launched)
General Specs
FamilyFastMATH
Frequency1,000 MHz
Bus typeRapidIO
Bus speed500 MHz
Bus rate4 GT/s
Microarchitecture
MicroarchitectureFashMATH
Process130 nm
TechnologyDynamic CMOS
Word Size32 bit
Cores1
Threads1
Max Memory1 GiB
Electrical
Power dissipation6 W
Vcore0.85 V

The FastMATH-LP was a microprocessor developed by Intrinsity operating at 1 GHz. The processor incorporates a high-performance MIPS CPU along with a powerful matrix and vector math unit. This mode was a low-power (LP) version of the normal FastMATH processor, operating at half the speed.

Cache[edit]

Main article: FastMATH § Cache
Cache Info [Edit Values]
L1I$ 16 KiB
16,384 B
0.0156 MiB
1x16 KiB 256 blocks × 16 words/block
L1D$ 16 KiB
16,384 B
0.0156 MiB
1x16 KiB 256 blocks × 16 words/block write-through or write-back mode
L2$ 1 MiB
1,024 KiB
1,048,576 B
9.765625e-4 GiB
1x1 MiB 4-way set associative (configurable as SRAM in 256 KiB increments)

Graphics[edit]

This SoC has no integrated graphics processing unit.

Memory controller[edit]

Integrated Memory Controller
Type DDR-400
Controllers 1
Channels 2
Max memory 1 GB

Matrix and Vector Unit[edit]

  • SIMD architecture
  • Operates on 4x4 array of 32-bit elements
  • Fixed-point matrix, vector, and scalar data types

Features[edit]

  • JTAG interface
  • 8-bit or 32-bit wide bus operates up to 66 MHz

Documents[edit]

Manuals[edit]


has featureJTAG +
l1d$ description256 blocks × 16 words/block +
l1i$ description256 blocks × 16 words/block +
l2$ description4-way set associative +