C o m m e n t s on "The Case for the R e d u c e d I n s t r u c t i o n Set
C o m p u t e r , " by P a t t e r s o n and D i t z e l
D o u g l a s W. C l a r k and W i l l i a m D. S t r e c k e r V A X S y s t e m s A r c h i t e c t u r e
D i g i t a l E q u i p m e n t C o r p o r a t i o n 1 9 2 5 A n d o v e r S t r e e t
T e w k s b u r y , MA 01876
S e p t e m b e r 1980
P a t t e r s o n a n d D i t z e l ' s p a p e r [3] a r g u e s t h a t a R e d u c e d I n s t r u c t i o n Set C o m p u t e r (RISC) c a n be as c o s t - e f f e c t i v e as a C o m p l e x I n s t r u c t i o n Set C o m p u t e r (CISC). In this note we s u g g e s t that s e v e r a l of their p o i n t s are m i s l e a d i n g , and p r e s e n t some e v i d e n c e on the o t h e r side of the a r g u m e n t . We r e l y h e a v i l y , as did they, on the V A X - i I a r c h i t e c t u r e [5] for e x a m p l e s .
The s u p e r i o r i t y of a RISC o v e r a c o r r e s p o n d i n g CISC w i l l be v e r y d i f f i c u l t to prove. C a s u a l e v a l u a t i o n of c o s t and p e r f o r m a n c e will not be s u f f i c i e n t u n l e s s the d i f f e r e n c e s b e t w e e n a C I S C and a R I S C are e x t r e m e , w h i c h is u n l i k e l y . P a p e r d e s i g n s will not be e n o u g h . A c a r e f u l c o m p a r i s o n b e t w e e n a RISC and a C I S C w o u l d seem to us to r e q u i r e a c o m p l e t e d e s i g n of the h a r d w a r e and m i c r o c o d e for both, c o n s t r u c t i o n or s i m u l a t i o n of the p r o c e s s o r s ,
the w r i t i n g of c o m p i l e r s and p o s s i b l y an o p e r a t i n g s y s t e m , and p e r f o r m a n c e m e a s u r e m e n t a c r o s s a v a r i e t y o f a p p l i c a t i o n s . W i t h o u t t h i s l e v e l of e f f o r t , c l a i m s o f i n c r e a s e d c o s t - e f f e c t i v e n e s s for a RISC are h a r d to s u p p o r t .
T h i s is not to say, h o w e v e r , that this s u b j e c t is u n w o r t h y of d i s c u s s i o n or a r g u m e n t until such a c o m p r e h e n s i v e e x p e r i m e n t is p e r f o r m e d . The i s s u e s r a i s e d in the P a t t e r s o n - D i t z e l p a p e r are i n t e r e s t i n g and i m p o r t a n t and a r c h i t e c t s of e i t h e r s t y l e of c o m p u t e r can p r o f i t from a d i s c u s s i o n of them.
C o m p l e x i t y vs. S i z e
n u m b e r of d a t a t y p e s ? D o e s c o m p l e t e n e s s (e.g. o r t h o g o n a l i t y o f o p e r a t o r and d a t a type) i n c r e a s e or d e c r e a s e c o m p l e x i t y ? Our m o s t s e r i o u s c r i t i c i s m of the p a p e r is t h a t it c o n t a i n s no f o r m a l d e f i n i t i o n o f a R I S C or a C I S C . In the a b s e n c e o f a d e f i n i t i o n of c o m p l e x i t y the s t a t e m e n t " . . . w e s h a l l a r g u e t h a t in m a n y c a s e s
the c o m p l e x i n s t r u c t i o n s e t s a r e m o r e d e t r i m e n t a l t h a n u s e f u l " is m e a n i n g l e s s .
C o d e D e n s i t y
P a t t e r s o n and D i t z e l c l a i m t h a t c o d e c o m p a c t i o n , n o r m a l l y t h o u g h t of as an a d v a n t a g e of a C I S C s u c h as V A X [i] , is as e a s i l y a c h i e v e d on a RISC, and t h a t in a n y c a s e d e n s e c o d e is n o t as i m p o r t a n t as it o n c e was, t h a n k s to c h e a p m e m o r y . But t h e i r c a s e for the R I S C b o i l s d o w n to the s e n t e n c e "We s u s p e c t t h a t c o d e c o m p a c t i o n c a n be as e a s i l y a c h i e v e d b y c l e a n i n g up the o r i g i n a l [ s i m p l e ] i n s t r u c t i o n s e t , " w h i c h is n o t c o n v i n c i n g w i t h o u t s u p p o r t i n g e v i d e n c e . A n d w h i l e the c o s t of m e m o r y d e c r e a s e s o v e r t i m e , it w i l l r e m a i n true t h a t a s m a l l a m o u n t of c h e a p m e m o r y c o s t s l e s s t h a n a l a r g e a m o u n t . F u r t h e r m o r e , for a g iven c o m p u t e r m o d e l m e m o r y is a p e r - s y s t e m c o s t , w h i l e m i c r o c o d e d e v e l o p m e n t (for a CISC) is a o n e - t i m e c o s t .
D e n s e c o d e , of c o u r s e , o f f e r s o t h e r a d v a n t a g e s as w e l l . C a c h e p e r f o r m a n c e and p a g i n g p e r f o r m a n c e will be b e t t e r if t h e r e are m o r e i n s t r u c t i o n s per c a c h e b l o c k and per p a g e .
D i f f e r e n t L a n g u a g e s U s e D i f f e r e n t I n s t r u c t i o n s
T h e p a p e r m a k e s the p o i n t ( c i t i n g S h u s t e k [4] and o t h e r s ) t h a t
" ' v e r y f e w o p c o d e s a c c o u n t for m o s t of a p r o g r a m ' s e x e c u t i o n . ' "
T r u e e n o u g h . But w h a t a b o u t d i f f e r e n t l a n g u a g e s ?
The top 20 i n s t r u c t i o n s for, say, F o r t r a n m a y n o t m a k e the C O B O L h i t - p a r a d e at a l l . The w e l l - k n o w n F o r t r a n b e n c h m a r k W h e t s t o n e ,
for e x a m p l e , d i s p l a y s the u s u a l kind of i n s t r u c t i o n f r e q u e n c y d i s t r i b u t i o n o n the V A X - i i / 7 8 0 : the t o p 10 i n s t r u c t i o n s a c c o u n t for 60% of all i n s t r u c t i o n e x e c u t i o n s , the top 20 g e t o v e r 75%, the t o p 40 g e t o v e r 90%. But t h o s e s a m e t o p 10 a c c o u n t for a m e r e 8% of the i n s t r u c t i o n e x e c u t i o n s in a c o m p a r a b l e s y n t h e t i c C O B O L b e n c h m a r k . W h e t s t o n e ' s t o p 20 g e t 21% (as o p p o s e d to 75%) of the C O B O L b e n c h m a r k ' s e x e c u t i o n s . M o r e t e l l i n g s t i l l , the t o p 20 of W h e t s t o n e a c c o u n t for o n l y 4% of the t i m e t a k e n b y the C O B O L b e n c h m a r k .
So a m u l t i p l i c i t y o f i n s t r u c t i o n s c a n h e l p to s u p p o r t a m u l t i p l i c i t y of l a n g u a g e s . As a r e c e n t p a p e r [2] p o i n t s out,
"For m o s t p r o g r a m m i n g e n v i r o n m e n t s , a s y s t e m m u s t be able to e f f e c t i v e l y s u p p o r t m u l t i p l e l a n g u a g e s . . . A s i n g l e i n s t r u c t i o n s e t t a i l o r e d to o n e p a r t i c u l a r l a n g u a g e is c o n s t r i c t i v e , as it can m a k e i m p l e m e n t a t i o n of o t h e r l a n g u a g e s d i f f i c u l t and i n e f f i c i e n t . "
T i m e is of the E s s e n c e
As S h u s t e k a r g u e s in his t h e s i s [4], the a m o u n t of t i m e s p e n t e x e c u t i n g an i n s t r u c t i o n is m o r e i m p o r t a n t for p e r f o r m a n c e t h a n its f r e q u e n c y of e x e c u t i o n . He g i v e s e x a m p l e s of r a r e l y - e x e c u t e d i n s t r u c t i o n s that c o n s u m e a large a m o u n t of e x e c u t i o n time.
R e p l a c i n g s u c h an i n s t r u c t i o n b y a m u l t i - i n s t r u c t i o n s e q u e n c e can m a k e t h i s m u c h w o r s e , and o p t i m i z i n g o n l y t h o s e i n s t r u c t i o n s t h a t are e x e c u t e d f r e q u e n t l y h a s the o b v i o u s h a z a r d .
Here is an e x a m p l e from the V A X - i i / 7 8 0 : in one t i m e - s h a r i n g b e n c h m a r k the i n s t r u c t i o n M O V C 3 (a c h a r a c t e r - m o v e i n s t r u c t i o n ) a c c o u n t s for less t h a n 0.4% of the i n s t r u c t i o n e x e c u t i o n s , b u t for 13% of the time; it is 60th in the f r e q u e n c y r a n k i n g , ist in the time r a n k i n g .
E a s e of C o m p i l e r - W r i t i n g
A l a r g e i n s t r u c t i o n set can be j u s t i f i e d by the d e s i r e to k e e p o p e r a t o r s and data t y p e s o r t h o g o n a l . T h u s the V A X a r c h i t e c t u r e i n c l u d e s , for e x a m p l e , six d i f f e r e n t e x c l u s i v e - O R i n s t r u c t i o n s : t w o - and t h r e e - o p e r a n d v e r s i o n s for b y t e s , w o r d s , and l o n g w o r d s . Some of t h e s e are u n d o u b t e d l y l i t t l e used. But c o d e g e n e r a t i o n in VAX c o m p i l e r s is s i m p l i f i e d by h a v i n g t h e m all (this is a t t e s t e d to b y V A X c o m p i l e r - w r i t e r s ) . F u r t h e r m o r e , o n c e y o u h a v e m i c r o c o d e for some of them, the o t h e r s c a n be i m p l e m e n t e d v e r y c h e a p l y .
M i c r o c o d e S i z e
The p a p e r ' s P D P - i i / 4 0 - V A X - i i / 7 8 0 m i c r o c o d e size c o m p a r i s o n s are s p e c i o u s . The a m o u n t of m i c r o c o d e in the 11/60 is n e a r l y ten t i m e s the 11/40, e v e n t h o u g h the i n s t r u c t i o n sets are the same.
The i n c r e a s e in m i c r o c o d e r e f l e c t s (a) i n c r e a s e d p e r f o r m a n c e , (b) r e p l a c e m e n t of h a r d w a r e b y m i c r o c o d e , and (c) m o r e e l a b o r a t e d i a g n o s t i c and c o n s o l e f u n c t i o n s . The 1 1 / 7 8 0 also s u p p o r t s t h r e e
I n c r e a s e d D e s i g n T i m e
The p a p e r ' s PDP-i - V A X - i I / 7 8 0 d e s i g n time c o m p a r i s o n s are also s p e c i o u s . I g n o r i n g t h e i n s t r u c t i o n s e t a l t o g e t h e r , t h e V A X - i i / 7 8 0 h a r d w a r e s y s t e m and the V M S s o f t w a r e are e n o r m o u s l y m o r e c o m p l e x than the P D P - i p r o c e s s o r . F u r t h e r m o r e , t h e r e are n u m e r o u s t i m e - c o n s u m i n g p r o c e s s e s in a l a r g e c o m p a n y d e s i g n i n g p r o d u c t s for h i g h - v o l u m e m a n u f a c t u r i n g that do not e x i s t in a small c o m p a n y d e s i g n i n g p r o d u c t s for l o w - v o l u m e m a n u f a c t u r i n g .
I n c r e a s e d D e s i g n E r r o r s
It is u n a r g u a b l e that t h e r e will be m o r e m i c r o c o d e e r r o r s in a large a m o u n t of m i c r o c o d e than in a small a m o u n t , all o t h e r t h i n g s b e i n g e q u a l . But if t h e r e is a small a m o u n t (RISC) , s o m e b o d y has to i m p l e m e n t the c o m p l e x f u n c t i o n s s o m e w h e r e . In a RISC, the c o m p i l e r and r u n - t i m e s y s t e m w o u l d bear the b u r d e n f o r m e r l y b o r n e b y m i c r o c o d e , and i m p l e m e n t a t i o n e r r o r s w o u l d p r e s u m a b l y t u r n up in this s o f t w a r e .
N o w one m i g h t argue that s o f t w a r e is e a s i e r to w r i t e t h a n m i c r o c o d e , and e a s i e r to c h a n g e . And one m i g h t c o u n t e r - a r g u e
t h a t m i c r o p r o g r a m d e v e l o p m e n t tools s u c h as m i c r o c o d e c o m p i l e r s will c h a n g e this s i t u a t i o n . But a c o m p l e x f u n c t i o n d e m a n d e d by a user p r o g r a m m u s t be i m p l e m e n t e d s o m e h o w , and e r r o r s will o c c u r . The d e t e c t i o n of d e s i g n e r r o r s d e p e n d s in l a r g e m e a s u r e on formal and i n f o r m a l test p r o c e s s e s . T h i s is as true for c o m p i l e r s as it
is for p r o c e s s o r s . Such p r o c e s s e s can r e a d i l y be d e v e l o p e d for p r o c e s s o r s . ( H o w w o u l d P a t t e r s o n a n d D i t z e l c o m p a r e t h e c o m p l e x i t y of the VAX-i i/780 m i c r o c o d e to that of, say, an o p t i m i z i n g c o m p i l e r ? )
I n t e r f a c e L e v e l
O n e o f the a d v a n t a g e s o f a h i g h e r l e v e l h a r d w a r e - s o f t w a r e i n t e r f a c e (CISC) over a l o w e r level i n t e r f a c e (RISC) is t h a t there is m o r e o p p o r t u n i t y to use s p e c i a l i z e d h a r d w a r e to a c h i e v e i m p r o v e d c o s t / p e r f o r m a n c e . For e x a m p l e , c o n s i d e r a C I S C w i t h a m u l t i p l y i n s t r u c t i o n and a RISC w i t h o u t one. The m u l t i p l y f u n c t i o n w o u l d be p e r f o r m e d on the R I S C w i t h a s e q u e n c e of m o v e s , b r a n c h e s , s h i f t s , and adds. To speed up the m u l t i p l y f u n c t i o n on the R I S C w o u l d r e q u i r e a s p e e d - u p of the w h o l e p r o c e s s o r w h i l e s p e e d i n g up the m u l t i p l y i n s t r u c t i o n on the CISC c o u l d be a c c o m p l i s h e d by a d d i n g s p e c i a l i z e d d a t a p a t h s and c o n t r o l . For some t e c h n o l o g i e s and p e r f o r m a n c e l e v e l s , the l a t t e r m a y be far less e x p e n s i v e than the former.
37
RISC and V L S I
In the a b s e n c e of m e t r i c s , t h i s s e c t i o n of the P a t t e r s o n - D i t z e l paper is u n c o n v i n c i n g . S u r e l y a s i m p l e i n s t r u c t i o n set c a n be i m p l e m e n t e d in less s i l i c o n than a c o m p l e x one. T h i s d o e s n ' t m e a n , h o w e v e r , that s y s t e m c o s t - e f f e c t i v e n e s s is i n c r e a s e d b y
r e d u c i n g i n s t r u c t i o n set c o m p l e x i t y .
T h e V A X I N D E X I n s t r u c t i o n
A n e c d o t a l a c c o u n t s of i r r a t i o n a l i m p l e m e n t a t i o n s are c e r t a i n l y i n t e r e s t i n g . Is it t y p i c a l , h o w e v e r , t h a t c o m p o s i t e i n s t r u c t i o n s run m o r e ~ l o w l y t h a n e q u i v a l e n t s e q u e n c e s of s i m p l e i n s t r u c t i o n s ? The p a p e r r e p o r t s that a s e q u e n c e of s e v e r a l s i m p l e i n s t r u c t i o n s can r e p l a c e the V A X I N D E X i n s t r u c t i o n w i t h a 45% speed g a i n on the 780. T h i s is a p r o b l e m of i m p l e m e n t a t i o n , not a r c h i t e c t u r e . F u n d a m e n t a l l y , a f t e r a l l , t h e i m p l e m e n t a t i o n o f t h e I N D E X f u n c t i o n w i t h m o r e t h a n one i n s t r u c t i o n s i m p l y c a n n o t take less time than the o n e - i n s t r u c t i o n v e r s i o n , a s s u m i n g equal h a r d w a r e in b o t h cases. The e x p l a n a t i o n of t h i s a n o m a l y is t h a t the 7 8 0 ' s F l o a t i n g P o i n t A c c e l e r a t o r s p e e d s up t h e m u l t i p l y in the m u l t i - i n s t r u c t i o n i m p l e m e n t a t i o n , b u t d o e s n ' t see I N D E X at all.
A F i n a l V A X F a c t
P a t t e r s o n and D i t z e l s u g g e s t t h a t m a r k e t i n g s t r a t e g y can i n c r e a s e the size or c o m p l e x i t y of an i n s t r u c t i o n set. We can s t a t e f r o m f i r s t - h a n d k n o w l e d g e t h a t t h i s is n o t t r u e for the V A X a r c h i t e c t u r e .
[i]
[2]
[3]
R e f e r e n c e s
D i e t z , W . B . , a n d S z e w e r e n k o , L. C o m p u t e r F a m i l y A r c h i t e c t u r e S e l e c t i o n , Phase IV F i n a l R e p o r t : C o m p a r a t i v e E v a l u a t i o n of the C a n d i d a t e C o m p u t e r A r c h i t e c t u r e s . Tech.
R e p o r t , C o m p u t e r S c i e n c e Dept., C a r n e g i e - M e l l o n U n i v e r s i t y , Nov. 1979.
D i t z e l , D . R . , a n d P a t t e r s o n , D . A . R e t r o s p e c t i v e o n H i g h - L e v e l L a n g u a g e C o m p u t e r A r c h i t e c t u r e . S e v e n t h
I n t e r n a t i o n a l S y m p o s i u m on C o m p u t e r A r c h i t e c t u r e , La B a u l e , France, May 1980.
P a t t e r s o n , D.A., and D i t z e l , D.R. The Case for the R e d u c e d I n s t r u c t i o n C o m p u t e r . A C M S I G A R C H C o m p u t e r A r c h i t e c t u r e N e w s , t h i s issue.