a rt,ra,rb

2

0

Add words in ra and rb

absdb rt,ra,rb

4

0

Subtract bytes in ra from rb, returns absolute value

addx rt,ra,rb

2

0

Add words in ra and rb to LSB of rt

ah rt,ra,rb

2

0

Add halfwords in ra and rb

ahi rt,ra,imm

2

0

Add halfwords in ra and imm value

ai rt,ra,imm

2

0

Add words in ra to imm value

and rt,ra,rb

2

0

AND the values of ra and rb

andbi rt,ra,imm

2

0

AND the bytes of ra with the imm value

andc rt,ra,rb

2

0

AND the values of ra and the complement of rb

andhi rt,ra,imm

2

0

AND the halfwords of ra with the imm value

andi rt,ra,imm

2

0

AND the words of ra with the imm value

avgb rt,ra,rb

4

0

Average of bytes in ra and rb

bg rt,ra,rb

2

0

Generate borrow from ra and rb

bgx rt,ra,rb

2

0

Generate borrow from ra, rb, and the LSB of rt

bi ra

4

1

Branch to ra

bid ra

4

1

Branch to ra, disable interrupts

bie ra

4

1

Branch to ra, enable interrupts

bihnz rt,ra

4

1

Branch to ra if rt halfword doesn’t equal 0

bihnzd rt,ra

4

1

Branch to ra if rt halfword doesn’t equal 0, disable

bihnze rt,ra

4

1

Branch to ra if rt halfword doesn’t equal 0, enable

bihz rt,ra

4

1

Branch to ra if rt halfword equals 0

bihzd rt,ra

4

1

Branch to ra if rt halfword equals 0, disable

bihze rt,ra

4

1

Branch to ra if rt halfword equals 0, enable

binz rt,ra

4

1

Branch to ra if rt word doesn’t equal 0

binzd rt,ra

4

1

Branch to ra if rt word doesn’t equal 0, disable

binze rt,ra

4

1

Branch to ra if rt word doesn’t equal 0, enable

bisl rt,ra

4

1

Branch to ra and set link

bisld rt,ra

4

1

Branch to ra and set link, enable

bisle rt,ra

4

1

Branch to ra and set link, disable

bisled rt,ra

4

1

Branch to ra and set link if an event occurs

bisledd rt,ra

4

1

Branch to ra and set link if an event occurs, disable

bislede rt,ra

4

1

Branch to ra and set link if an event occurs, enable

biz rt,ra

4

1

Branch to ra if rt word equals 0

bizd rt,ra

4

1

Branch to ra if rt word equals 0, disable

bize rt,ra

4

1

Branch to ra if rt word equals 0, enable

br imm

4

1

Branch to sum of imm and PC

bra imm

4

1

Branch to the imm address

brasl rt,imm

4

1

Branch to imm and set link

brhnz rt,imm

4

1

Branch to sum of imm and PC if rt halfword doesn’t equal 0

brhz rt,imm

4

1

Branch to sum of imm and PC if rt halfword equals 0

brnz rt,imm

4

1

Branch to sum of imm and PC if rt word doesn’t equal 0

brsl rt,imm

4

1

Branch to sum of imm and PC, set link

brz rt,imm

4

1

Branch to sum of imm and PC if rt word equals 0

cbd rt,index(ra)

4

1

Create mask for byte insertion

cbx rt,ra,rb

4

1

Create mask for byte insertion

cdd rt,index(ra)

4

1

Create mask for doubleword insertion

cdx rt,ra,rb

4

1

Create mask for doubleword insertion

ceq rt,ra,rb

2

0

Compare equality of words in ra and rb

ceqb rt,ra,rb

2

0

Compare equality of bytes in ra and rb

ceqbi rt,ra,imm

2

0

Compare equality of bytes in ra to imm value

ceqh rt,ra,rb

2

0

Compare equality of halfwords in ra and rb

ceqhi rt,ra,imm

2

0

Compare equality of halfwords in ra to imm value

ceqi rt,ra,imm

2

0

Compare equality of words in ra to imm value

cflts rt,ra,imm

7

0

Convert float in ra to signed integer in rt, scaled by imm

cfltu rt,ra,imm

7

0

Convert float in ra to unsigned integer in rt, scaled by imm

cg rt,ra,rb

2

0

Generate carry vector from ra and rb

cgt rt,ra,rb

2

0

Return if words in ra are greater than words in rb

cgtb rt,ra,rb

2

0

Return if bytes in ra are greater than bytes in rb

cgtbi rt,ra,imm

2

0

Return if bytes in ra are greater than imm

cgth rt,ra,rb

2

0

Return if halfwords in ra are greater than rb

cgthi rt,ra,imm

2

0

Return if halfwords in ra are greater than imm

cgti rt,ra,imm

2

0

Return if words in ra are greater than imm

cgx rt,ra,rb

2

0

Generate carry vector from ra, rb, and the LSB of rt

chd rt,index(ra)

4

1

Create mask for halfword insertion

chx rt,ra,rb

4

1

Create mask for halfword insertion

clgt rt,ra,rb

2

0

Return if words in ra are logically greater than rb

clgtb rt,ra,rb

2

0

Return if bytes in ra are logically greater than rb

clgtbi rt,ra,imm

2

0

Return if bytes in ra are logically greater than imm

clgth rt,ra,rb

2

0

Return if halfwords in ra are logically greater than rb

clgthi rt,ra,imm

2

0

Return if halfwords in ra are logically greater than imm

clgti rt,ra,imm

2

0

Return if words in ra are logically greater than imm

clz rt,ra

2

0

Count 0s preceding the first 1 in ra

cntb rt,ra

4

0

Count number of 1s in each byte of ra

csflt rt,ra,imm

7

0

Convert signed integer in ra to float in rt, scaled by imm

cuflt rt,ra,imm

7

0

Convert unsigned integer in ra to float in rt, scaled by imm

cwd rt,index(ra)

4

1

Create mask for word insertion

cwx rt,ra,rb

4

1

Create mask for word insertion

dfa^[1] rt,ra,rb

13

0

Add double-precision values in ra and rb

dfm1 rt,ra,rb

13

0

Multiply double-precision values in ra and rb

dfma1 rt,ra,rb

13

0

Multiply double-precision values in ra and rb, add to rt

dfms1 rt,ra,rb

13

0

Multiply double-precision values in ra and rb, subtract values in rt

dfnma1 rt,ra,rb

13

0

Multiply double-precision values in ra and rb, add values in rt, negate result

dfnms1 rt,ra,rb

13

0

Multiply double-precision values in ra and rb, subtract values in rt, negate result

dfs1 rt,ra,rb

13

0

Subtract double-precision value in ra from rb

dsync

4

1

Ensures LS data is current before external accessing

eqv rt,ra,rb

2

0

Return 1 if ra and rb are the same, 0 otherwise

fa rt,ra,rb

6

0

Add single-precision values in ra and rb

fceq rt,ra,rb

2

0

Compare floating-point equality of ra and rb

fcgt rt,ra,rb

2

0

Return if floating-point ra is greater than floating-point rb

fcmeq rt,ra,rb

2

0

Compare floating-point equality of ra and rb magnitudes

fcmgt rt,ra,rb

2

0

Return if floating-point magnitude of ra is greater than that of rb

fesd1 rt,ra

13

0

Convert float in ra to double in rt

fi rt,ra,rb

7

0

Floating-point interpolate between ra and rb

fm rt,ra,rb

6

0

Multiply floating-point values in ra and rb

fma rt,ra,rb,rc

6

0

Multiply floating-point values in ra and rb, add to values in rc

fms rt,ra,rb,rc

6

0

Multiply floating-point values in ra and rb, subtract values in rc

fnms rt,ra,rb,rc

6

0

Multiply floating-point values in ra and rb, subtract values in rc, negate result

frds1 rt,ra

13

0

Round double in ra to float in rt

frest rt,ra

4

1

Floating-point reciprocal estimate

frsqest rt,ra

4

1

Floating-point reciprocal absolute square-root estimate

fs rt,ra,rb

6

0

Subtract floating-point values in ra from rb

fscrrd1 rt

13

0

Move floating point status and control register to rt

fscrwr ra

7

0

Move rt to floating point status and control register

fsm rt,ra

4

1

Form select mask for words

fsmb rt,ra

4

1

Form select mask for bytes

fsmbi rt,imm

4

1

Form select mask for bytes with imm

fsmh rt,ra

4

1

Form select mask for halfwords

gb rt,ra

4

1

Concatenate LSBs of each word in ra

gbb rt,ra

4

1

Concatenate LSBs of each byte in ra

gbh rt,ra

4

1

Concatenate LSBs of each halfword in ra

hbr imm,ra

15

1

Hint that the branch at imm will target PC + ra

hbra imm1,imm2

15

1

Hint that the branch at imm1 will target PC + imm2

hbrp

15

1

Hint for upcoming branch, prefetch

hbrr ra,rb

15

1

Hint that the branch at PC + imm1 will target PC + imm2

heq ra,rb

2

0

Halt if ra equals rb

heqi ra,imm

2

0

Halt if ra equals imm

hgt ra,rb

2

0

Halt if ra is greater than rb

hgti ra,imm

2

0

Halt if ra is greater than imm

hlgt ra,rb

2

0

Halt if ra is logically greater than rb

hlgti ra,imm

2

0

Halt if ra is logically greater than imm

il rt,imm

2

0

Load each word in rt with the imm value

ila rt,imm

2

0

Load imm (18-bit) into the LSBs of rt

ilh rt,imm

2

0

Load each halfword in rt with the imm value

llhu rt,imm

2

0

Load the high halfword of each word in rt with imm

iohl rt,imm

2

0

OR the low halfword of each word in rt with imm

iret

4

1

Interrupt return

iretd

4

1

Interrupt return, disable

irete

4

1

Interrupt return, enable

lnop

0

1

No operation (pipeline 1)

lqa rt,lsa

6

1

Load quadword from lsa to register rt

lqd rt,index(ra)

6

1

Load quadword from ra+index to register rt

lqr rt,lsa

6

1

Load quadword from lsa+PC to register rt

lqx rt,ra,rb

6

1

Load quadword from ra+rb to register rt

mfspr rt,imm

6

1

Move special-purpose register imm to rt

mpy rt,ra,rb

7

0

Multiply low halfwords in ra and rb

mpya rt,ra,rb,rc

7

0

Multiply signed words in ra and rb, add to rc

mpyh rt,ra,rb

7

0

Multiply high halfwords of ra and low hws of rb

mpyhh rt,ra,rb

7

0

Multiply high halfwords of ra and rb

mpyhha rt,ra,rb

7

0

Multiply high halfwords of ra and rb, and add rt

mpyhhau rt,ra,rb

7

0

Multiply unsigned high halfwords of ra and rb, and add rt

mpyhhu rt,ra,rb

7

0

Multiply unsigned high halfwords of ra and rb

mpyi rt,ra,imm

7

0

Multiply low halfwords in ra by imm value

mpys rt,ra,rb

7

0

Multiply low halfwords in ra and rb and shift right

mpyu rt,ra,rb

7

0

Multiply unsigned low halfwords in ra and rb

mpyui rt,ra,imm

7

0

Multiply unsigned low halfwords in ra by imm value

mtspr imm,rt

6

1

Move rt to special-purpose register imm

nand rt,ra,rb

2

0

NAND the values of ra and rb

nop

0

0

No operation (pipeline 0)

nor rt,ra,rb

2

0

NOR the values of ra and rb

or rt,ra,rb

2

0

OR the values of ra and rb

orbi rt,ra,imm

2

0

OR the bytes of ra with the imm value

orc rt,ra,rb

2

0

OR the values of ra and the complement of rb

orhi rt,ra,imm

2

0

OR the halfwords of ra with the imm value

ori rt,ra,imm

2

0

OR the words of ra with the imm value

orx rt,ra

4

1

OR the words of ra

rchcnt rt,imm

6

1

Read capacity of channel imm into rt

rdch rt,imm

6

1

Read data from channel imm into rt

rot rt,ra,rb

4

0

Rotate bits in words of ra left according to rb

roth rt,ra,rb

4

0

Rotate bits in halfwords of ra left according to rb

rothi rt,ra,imm

4

0

Rotate bits in halfwords of ra left according to imm

rothm rt,ra,rb

4

0

Rotate bits in halfwords of ra right according to -rb

rothmi rt,ra,imm

4

0

Rotate bits in halfwords of ra right according to -imm

roti rt,ra,imm

4

0

Rotate bits in words of ra left according to imm

rotm rt,ra,rb

4

0

Shift bits in words of ra right according to -rb

rotma rt,ra,rb

4

0

Shift bits in words of ra right according to -rb (algebraic)

rotmah rt,ra,rb

4

0

Shift bits in halfwords of ra right according to -rb (algebraic)

rotmahi rt,ra,imm

4

0

Shift bits in halfwords of ra right according to -imm (algebraic)

rotmai rt,ra,imm

4

0

Shift bits in words of ra right according to -imm (algebraic)

rotmi rt,ra,imm

4

0

Shift bits in words of ra right according to -imm

rotqbi rt,ra,rb

4

1

Rotate entire ra left by bits according to rb

rotqbii rt,ra,imm

4

1

Rotate entire ra left by bits according to imm

rotqby rt,ra,rb

4

1

Rotate entire ra left by bytes according to rb

rotqbybi rt,ra,rb

4

1

Rotate entire ra left by bytes according to rb count

rotqbyi rt,ra,imm

4

1

Rotate entire ra left by bytes according to imm

rotqmbi rt,ra,rb

4

1

Shift entire ra right by bits according to -rb

rotqmbii rt,ra,imm

4

1

Shift entire ra right by bits according to -imm

rotqmby rt,ra,rb

4

1

Shift entire ra right by bytes according to -rb

rotqmbybi rt,ra,rb

4

1

Shift entire ra right by bytes according to -rb count (algebraic)

rotqmbyi rt,ra,imm

4

1

Shift entire ra right by bytes according to -imm

selb rt,ra,rb,rc

2

0

Select bits from ra and rb according to rc

sf rt,ra,rb

2

0

Subtract words in ra from rb

sfh rt,ra,rb

2

0

Subtract halfwords in ra from rb

sfhi rt,ra,imm

2

0

Subtract halfwords in ra from imm value

sfi rt,ra,imm

2

0

Subtract words in ra from imm value

sfx rt,ra,rb

2

0

Subtract words in ra from rb and LSB of rt

shl rt,ra,rb

4

0

Shift bits in words of ra left according to rb

shlh rt,ra,rb

4

0

Shift bits in halfwords of ra left according to rb

shlhi rt,ra,imm

4

0

Shift bits in halfwords of ra left according to imm

shli rt,ra,imm

4

0

Shift bits in words of ra left according to imm

shlqbi rt,ra,rb

4

1

Shift entire ra left by bits according to rb

shlqbii rt,ra,imm

4

1

Shift entire ra left by bits according to imm

shlqby rt,ra,rb

4

1

Shift entire ra left by bytes according to rb

shlqbybi rt,ra,rb

4

1

Shift entire ra left by bytes according to rb count

shlqbyi rt,ra,imm

4

1

Shift entire ra left by bytes according to imm

shufb rt,ra,rb,rc

4

1

Form rt from the bytes of ra and rb according to rc

stop

4

1

Halt the SPU and send stop signal to PPU

stopd

4

1

Halt the SPU and send signal (can be used as breakpoint)

stqa rt,lsa

6

1

Store quadword from register rt to lsa

stqd rt,index(ra)

6

1

Store quadword from register rt to ra+index

stqr rt,lsa

6

1

Store quadword from register rt to lsa+PC

stqx rt,ra,rb

6

1

Store quadword from register rt to ra+rb

sumb rt,ra,rb

4

0

Add bytes in ra and rb, return halfword results

sync

4

1

Force SPU to complete all store operations before continuing

syncc

4

1

Force SPU to complete store and channel operations

wrch imm,rt

6

1

Write data from rt into channel imm

xor rt,ra,rb

2

0

XOR the values of ra and rb

xorbi rt,ra,imm

2

0

XOR the bytes of ra with the imm value

xorhi rt,ra,imm

2

0

XOR the halfwords of ra with the imm value

xori rt,ra,imm

2

0

XOR the words of ra with the imm value

xsbh rt,ra

2

0

Sign extend bytes in ra to halfwords

xshw rt,ra

2

0

Sign extend halfwords in ra to words

xswd rt,ra

2

0

Sign extend words in ra to doublewords