🚀 go-pugleaf

Excuuuuuuuuuuse me! Three errors, one very slight, two not so slight!

In article <6903@pdn.paradyne.com> alan@oz.paradyne.com (Alan Lovejoy) writes:
>Sounds like an interesting contest to me!  Here's my try (for multiply, 
>anyway) using the mc88k instruction set:
>
>(NOTE: r0 is the constant 0 (a hardware protocol); jsr sets r1 with the 
>return addrss)

;dmulu -- multiply unsigned 32x32=<64
;r2 and r3 contain two 32-bit unsigned integers to be multiplied
;r12 will contain the high word (32 bits) of the product (r2 * r3)
;r13 will contain the low word (32 bits) of the product (r2 * r3)

	.proc	dmulu

dmulu:
;divide the two 32-bit numbers into 4 16-bit parts:
extu 	r4, r2, 16<16>; 	r4 = r2 >> 16; (1 cycle)

*** errror #1:
The original posting consistently had "16<16<" instead of "16<16>".  The
wonders of global replace... (The very slight error).

extu	r5, r3, 16<16>;		r5 = r3 >> 16; (1 cycle)
mask 	r2, r2, $FFFF;		r2 = r2 & 0xFFFF (1 cycle)
mask 	r3, r3, $FFFF;		r3 = r3 & 0xFFFF (1 cycle)

;calculate partial products:
>mul	r6, r3, r3;		r6 = r2 * r3; (4 cycles!)
 
*** error #2:
One of the r3's in the above instruction must be an r2 (as the comment shows).

>mul	r7, r3, r4;             r7 = r3 * r4; (4 cycles!)
>mul    r8, r2, r5;             r8 = r2 * r5; (4 cycles!)
>mul	r9, r4, r5;             r9 = r4 * r5; (4 cycles!)

*** error #3:
I completely forgot that the mul instruction is fully pipelined!
It takes 4 cycles to complete, yes, but a new instruction can be
issued on the very next cycle. Upto six mul and/or fmul instructions
can be in the pipeline at one time.  So, (after some reordering to
avoid stalls) we have instead:

;calculate partial products:
mul	r6, r2, r3;		r6 = r2 * r3; (1 cycle)
mul	r8, r2, r5;             r8 = r2 * r5; (1 cycle)
mul	r7, r3, r4;             r7 = r3 * r4; (1 cycle)
mul	r9, r4, r5;             r9 = r4 * r5; (1 cycle)

;sum partial products:
extu	r10, r6, 16<16>;	r10 = r6 >> 16; (1 cycle)
addu	r8, r8, r10;		r8 = r8 + r10; (1 cycle)
*** note that r7 changed to r8 to avoid a stall
addu.co	r10, r7, r8;		r10 = r7 + r8; generate carry bit (1 cycle)
addu.ci	r12, r0, r0;		r12 = carry from previous addu (1 cycle)
mak	r12, r12, 16<16>;	r12 = r12 << 16; (1 cycle)
addu  	r12, r12, r9;		r12 = r12 + r9; (1 cycle)
mak	r10, r10, 16<16>;	r10 = r10 << 16; (1 cycle) 
mask	r13, r6, $FFFF;		r13 = r6 & 0xFFFF; (1 cycle)
jmp.n	r1;                     return to caller after next instruction (1cycle)
or	r13, r13, r10;		r13 = r13 | r10; (1 cycle)
; 	done: 			18 cycles total (without short circuit code)
 	.end

; invoking dmulu:
;       it is the caller's responsibility to save registers r2-13, which
;       the caller may or may not need to do...
or	r25, r0, #dmuluLo16;    r25 = low 16 bits of dmulu address (1 cycle) 
or.u	r25, r25, #dmuluHi16;   r25 = r25 | high 16 bits of dmulu address (1 c)
ld	r2, r30, #factor1;	r2 = *(framePtr + offset of factor1) (1 cycle) 
jsr.n	r25;			call dmulu after next instruction (1 cycle) 
ld	r3, r30, #factor2;	r3 = *(framePtr + offset of factor2) (1 cycle) 
st.d	r12, r30, #product;	*(frametPtr + offset of product) = r12, r13

; register to register: 1 cycle call excluding execution time for dmulu 
; memory to memory: 6 cycle call excluding execution time for dmulu
; GRAND TOTAL register to register: 19 cycles (was 30)
; GRAND TOTAL memory to memory: 24 cycles (was 36) 

*** A significant performance improvement of 33 per cent.



____"Congress shall have the power to prohibit speech offensive to Congress"____
Alan Lovejoy; alan@pdn; 813-530-2211; AT&T Paradyne: 8550 Ulmerton, Largo, FL.
Disclaimer: I do not speak for AT&T Paradyne.  They do not speak for me. 
Mottos:  << Many are cold, but few are frozen. >>     << Frigido, ergo sum. >>