github-mirrors/vortex
1
0
Fork 0
mirror of https://github.com/vortexgpgpu/vortex.git synced 2025-04-23 21:39:10 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Multithreaded matrix multiply closer to working, but segfaulting.

Browse source
This commit is contained in:
cdkersey 2013-11-07 18:45:22 -05:00
parent d7352c02d1
commit a95bb3b097
2 changed files with 33 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

11
src/test/Makefile
View file

@ -4,11 +4,13 @@ HARPEM = ../harptool -E
HARPDIS = ../harptool -D
4BARCH = 4b16/16/2
all: simple.bin sieve.bin 2thread.bin simple.4b.bin sieve.4b.bin 2thread.4b.bin bubble.bin bubble.4b.bin dotprod.bin dotprod.4b.bin matmul.bin matmul.4b.bin
all: simple.bin sieve.bin 2thread.bin simple.4b.bin sieve.4b.bin 2thread.4b.bin bubble.bin bubble.4b.bin dotprod.bin dotprod.4b.bin matmul.bin matmul.4b.bin \
matmul-mt.s
run: simple.out sieve.out 2thread.out simple.4b.out sieve.4b.out 2thread.4b.out bubble.out bubble.4b.out dotprod.out dotprod.4b.out matmul.out matmul.4b.out
run: simple.out sieve.out 2thread.out simple.4b.out sieve.4b.out 2thread.4b.out bubble.out bubble.4b.out dotprod.out dotprod.4b.out matmul.out matmul.4b.out\
matmul-mt.out
disas: simple.d sieve.d 2thread.d simple.4b.d sieve.4b.d 2thread.4b.d bubble.d bubble.4b.d dotprod.d dotprod.4b.d matmul.d matmul.4b.d
disas: simple.d sieve.d 2thread.d simple.4b.d sieve.4b.d 2thread.4b.d bubble.d bubble.4b.d dotprod.d dotprod.4b.d matmul.d matmul.4b.d matmul-mt.d
%.4b.out : %.4b.bin
$(HARPEM) -a $(4BARCH) -c $< > $@
@ -40,6 +42,9 @@ dotprod.bin : boot.HOF lib.HOF dotprod.HOF
matmul.bin : boot.HOF lib.HOF matmul.HOF
$(HARPLD) -o $@ $^
matmul-mt.bin : boot.HOF lib.HOF matmul-mt.HOF
$(HARPLD) -o $@ $^
simple.4b.bin : boot.4b.HOF lib.4b.HOF simple.4b.HOF
$(HARPLD) --arch $(4BARCH) -o $@ $^

32
src/test/matmul-mt.s
View file

@ -41,7 +41,7 @@ ploop: ld %r7, %r1, #0;
matgen: ldi %r2, #0;
st %r5, %r2, retaddr;
ldi %r2, #1;
shli %r2, %r2, (`__WORD + 1); /* Multiply r0 by 2*__WORD */
shl %r2, %r2, %r1;
shl %r2, %r2, %r1;
ori %r3, %r0, #0;
@ -59,7 +59,8 @@ mgloop: jali %r5, randf;
/* Write the matrix product of square matrix at (%r0) and (%r1) to (%r2). The
size of these matrices is 2^Nx2^N, where N = %r3 */
matmul: ldi %r4, #1;
matmul: ori %r22, %r5, #0;
ldi %r4, #1;
ldi %r10, (`__WORD); /* ` is the log base 2 operator */
shl %r4, %r4, %r3;
add %r10, %r10, %r3;
@ -67,9 +68,25 @@ matmul: ldi %r4, #1;
shl %r14, %r14, %r10;
divi %r17, %r14, THREADS; /* Spawn threads */
sloop:
ori %r18, %r0, #0;
ori %r19, %r2, #0;
ldi %r20, #0;
sloop: add %r0, %r0, %r17;
add %r2, %r2, %r17;
addi %r20, %r20, #1;
subi %r21, %r20, THREADS;
rtop @p0, %r21;
notp @p1, @p0;
@p1 ? clone %r20;
@p0 ? jmpi sloop;
jmpr %r5;
ori %r0, %r18, #0;
ori %r2, %r19, #0;
clone %r20;
jalis %r5, matmulthd;
jmpr %r22;
/* One thread of matrix multiplication. Expected register values at start:
* %r0 - matrix a pointer (plus offset)
@ -115,10 +132,11 @@ iloop: ld %r7, %r11, #0;
jmprt %r5;
.align 4096
.perm rw
matrix_a: .space 64;
matrix_b: .space 64;
.align 4096
matrix_a: .space 64
matrix_b: .space 64
matrix_r: .space 64
retaddr: .word 0