wu-arch/kernels
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added warp independance

Browse Source
This commit is contained in:
Demo
2019-03-10 03:43:25 -04:00
parent 56706e1f38
commit f82a938267
18 changed files with 1323 additions and 298451 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/Makefile
View File

@@ -5,7 +5,7 @@ CXXFLAGS ?= -std=c++11 -fPIC -O3 # -g -DUSE_DEBUG=3 -DPRINT_ACTIVE_THREADS
LDLIBS ?= -pthread LDLIBS ?= -pthread
PREFIX ?= /usr/local PREFIX ?= /usr/local
LIB_OBJS=args.o obj.o mem.o core.o instruction.o enc.o util.o lex.yy.o LIB_OBJS=args.o obj.o mem.o core.o instruction.o enc.o util.o
all: harptool libharplib.so libharplib.a #libqsim-harp.so all: harptool libharplib.so libharplib.a #libqsim-harp.so
@@ -45,8 +45,6 @@ core.o : core.cpp include/types.h include/util.h include/mem.h \
# include/mem.h # include/mem.h
# $(CXX) $(CXXFLAGS) $(QSIM_CXXFLAGS) -shared -o $@ $^ # $(CXX) $(CXXFLAGS) $(QSIM_CXXFLAGS) -shared -o $@ $^
lex.yy.cc: scanner.lex
flex scanner.lex
install: install:
cp libharplib.so $(PREFIX)/lib cp libharplib.so $(PREFIX)/lib

117
src/enc.cpp
View File

@@ -403,120 +403,3 @@ Instruction *WordDecoder::decode(const std::vector<Byte> &v, Size &idx) {
return &inst; return &inst;
} }
// WordEncoder::WordEncoder(const ArchDef &arch) {
// getSizes(arch, n, o, r, p, i1, i2, i3);
// if (p > r) r = p;
// oMask = mask(o); rMask = mask(r); pMask = mask(p);
// i1Mask = mask(i1); i2Mask = mask(i2); i3Mask = mask(i3);
// }
// Size WordEncoder::encode(Ref *&ref, std::vector<Byte> &v,
// Size idx, Instruction &i)
// {
// Word code = 0;
// Size bitsWritten = 0;
// /* Predicate/predicated bit */
// if (i.hasPred()) {
// code = 1 << p;
// code |= (i.getPred()&pMask);
// if (i.getPred() > pMask) {
// cout << "Predicate in " << i << " does not fit in encoding.\n";
// exit(1);
// }
// }
// bitsWritten += (1 + p);
// /* Opcode */
// code <<= o;
// code |= (i.getOpcode()&oMask);
// if (i.getOpcode() > oMask) {
// cout << "Opcode in " << i << " does not fit in encoding.\n";
// exit(1);
// }
// bitsWritten += o;
// if (i.hasRDest()) {
// code <<= r;
// code |= i.getRDest();
// bitsWritten += r;
// if (i.getRDest() > rMask) {
// cout << "Destination register in " << i << " does not fit in encoding.\n";
// exit(1);
// }
// }
// if (i.hasPDest()) {
// code <<= r;
// code |= i.getPDest();
// bitsWritten += r;
// if (i.getPDest() > rMask) {
// cout << "Destination predicate in " <<i<< " does not fit in encoding.\n";
// exit(1);
// }
// }
// for (Size j = 0; j < i.getNRSrc(); j++) {
// code <<= r;
// code |= i.getRSrc(j);
// bitsWritten += r;
// if (i.getRSrc(j) > rMask) {
// cout << "Source register " << j << " in " << i
// << " does not fit in encoding.\n";
// exit(1);
// }
// }
// for (Size j = 0; j < i.getNPSrc(); j++) {
// code <<= r;
// code |= i.getPSrc(j);
// bitsWritten += r;
// if (i.getPSrc(j) > rMask) {
// cout << "Source predicate " << j << " in " << i
// << " does not fit in encoding.\n";
// exit(1);
// }
// }
// if (i.hasRefLiteral()) {
// Ref *r = i.getRefLiteral();
// ref = new OffsetRef(r->name, v, idx, n - bitsWritten, n, i.hasRelImm());
// } else {
// ref = NULL;
// }
// if (i.hasImm()) {
// if (bitsWritten == n - i1) {
// code <<= i1;
// code |= (i.getImm()&i1Mask);
// bitsWritten += i1;
// Word_s ws(i.getImm());
// if ((ws >> i1) != 0 && (ws >> i1) != -1) goto tooBigImm;
// } else if (bitsWritten == n - i2) {
// code <<= i2;
// code |= (i.getImm()&i2Mask);
// bitsWritten += i2;
// Word_s ws(i.getImm());
// if ((ws >> i2) != 0 && (ws >> i2) != -1) goto tooBigImm;
// } else if (bitsWritten == n - i3) {
// code <<= i3;
// code |= (i.getImm()&i3Mask);
// bitsWritten += i3;
// Word_s ws(i.getImm());
// if ((ws >> i3) != 0 && (ws >> i3) != -1) goto tooBigImm;
// } else {
// cout << "WordEncoder::encode() could not encode: " << i << '\n';
// exit(1);
// }
// }
// if (bitsWritten < n) code <<= (n - bitsWritten);
// writeWord(v, idx, n/8, code);
// return n/8;
// tooBigImm:
// cout << "Immediate in " << i << " too large to encode.\n";
// exit(1);
// }

6
src/harptool.cpp
View File

@@ -262,12 +262,6 @@ int emu_main(int argc, char **argv) {
if (showStats) core.printStats(); if (showStats) core.printStats();
Addr base_addr = 0x81000000;
for (Addr i = 0; i < 16; i++)
{
Addr new_addr = base_addr + (i *4);
std::cout << std::hex << new_addr << " = " << std::hex << old_ram.read(new_addr) << "\n";
}
std::cout << "\n"; std::cout << "\n";
return 0; return 0;

2
src/include/debug.h
View File

@@ -4,7 +4,7 @@
#ifndef __DEBUG_H #ifndef __DEBUG_H
#define __DEBUG_H #define __DEBUG_H
#define USE_DEBUG 9 //#define USE_DEBUG 9
#ifdef USE_DEBUG #ifdef USE_DEBUG
#include <iostream> #include <iostream>

110
src/instruction.cpp
View File

@@ -59,17 +59,17 @@ bool checkUnanimous(unsigned p, const std::vector<std::vector<Reg<Word> > >& m,
} }
if (i == m.size()) throw DivergentBranchException(); if (i == m.size()) throw DivergentBranchException();
std::cout << "same: " << same << " with -> "; //std::cout << "same: " << same << " with -> ";
for (; i < m.size(); ++i) { for (; i < m.size(); ++i) {
if (tm[i]) { if (tm[i]) {
std::cout << " " << (bool(m[i][p])); //std::cout << " " << (bool(m[i][p]));
if (same != (bool(m[i][p]))) { if (same != (bool(m[i][p]))) {
std::cout << " FALSE\n"; //std::cout << " FALSE\n";
return false; return false;
} }
} }
} }
std::cout << " TRUE\n"; //std::cout << " TRUE\n";
return true; return true;
} }
@@ -84,7 +84,7 @@ void Instruction::executeOn(Warp &c) {
/* If I try to execute a privileged instruction in user mode, throw an /* If I try to execute a privileged instruction in user mode, throw an
exception 3. */ exception 3. */
if (instTable[op].privileged && !c.supervisorMode) { if (instTable[op].privileged && !c.supervisorMode) {
std::cout << "INTERRUPT SUPERVISOR\n"; //std::cout << "INTERRUPT SUPERVISOR\n";
c.interrupt(3); c.interrupt(3);
return; return;
} }
@@ -126,8 +126,8 @@ void Instruction::executeOn(Warp &c) {
vector<Reg<bool> > &pReg(c.pred[t]); vector<Reg<bool> > &pReg(c.pred[t]);
stack<DomStackEntry> &domStack(c.domStack); stack<DomStackEntry> &domStack(c.domStack);
std::cout << std::hex << "opcode: " << op << " func3: " << func3 << "\n"; //std::cout << std::hex << "opcode: " << op << " func3: " << func3 << "\n";
if (op == GPGPU) std::cout << "OPCODE MATCHED GPGPU\n"; if (op == GPGPU) //std::cout << "OPCODE MATCHED GPGPU\n";
// If this thread is masked out, don't execute the instruction, unless it's // If this thread is masked out, don't execute the instruction, unless it's
// a split or join. // a split or join.
@@ -151,10 +151,10 @@ void Instruction::executeOn(Warp &c) {
switch (op) { switch (op) {
case NOP: case NOP:
std::cout << "NOP_INST\n"; //std::cout << "NOP_INST\n";
break; break;
case R_INST: case R_INST:
std::cout << "R_INST\n"; //std::cout << "R_INST\n";
switch (func3) switch (func3)
{ {
case 0: case 0:
@@ -221,11 +221,11 @@ void Instruction::executeOn(Warp &c) {
break; break;
case L_INST: case L_INST:
std::cout << "L_INST\n"; //std::cout << "L_INST\n";
memAddr = ((reg[rsrc[0]] + immsrc) & 0xFFFFFFFC); memAddr = ((reg[rsrc[0]] + immsrc) & 0xFFFFFFFC);
shift_by = ((reg[rsrc[0]] + immsrc) & 0x00000003) * 8; shift_by = ((reg[rsrc[0]] + immsrc) & 0x00000003) * 8;
data_read = c.core->mem.read(memAddr, c.supervisorMode); data_read = c.core->mem.read(memAddr, c.supervisorMode);
// std::cout <<std::hex<< "EXECUTE: " << reg[rsrc[0]] << " + " << immsrc << " = " << memAddr << " -> data_read: " << data_read << "\n"; // //std::cout <<std::hex<< "EXECUTE: " << reg[rsrc[0]] << " + " << immsrc << " = " << memAddr << " -> data_read: " << data_read << "\n";
#ifdef EMU_INSTRUMENTATION #ifdef EMU_INSTRUMENTATION
Harp::OSDomain::osDomain-> Harp::OSDomain::osDomain->
do_mem(0, memAddr, c.core->mem.virtToPhys(memAddr), 8, true); do_mem(0, memAddr, c.core->mem.virtToPhys(memAddr), 8, true);
@@ -239,7 +239,7 @@ void Instruction::executeOn(Warp &c) {
break; break;
case 1: case 1:
// LH // LH
// std::cout << "shifting by: " << shift_by << " final data: " << ((data_read >> shift_by) & 0xFFFF, 16, 0xFFFF) << "\n"; // //std::cout << "shifting by: " << shift_by << " final data: " << ((data_read >> shift_by) & 0xFFFF, 16, 0xFFFF) << "\n";
reg[rdest] = signExt((data_read >> shift_by) & 0xFFFF, 16, 0xFFFF); reg[rdest] = signExt((data_read >> shift_by) & 0xFFFF, 16, 0xFFFF);
break; break;
case 2: case 2:
@@ -259,7 +259,7 @@ void Instruction::executeOn(Warp &c) {
} }
break; break;
case I_INST: case I_INST:
std::cout << "I_INST\n"; //std::cout << "I_INST\n";
switch (func3) switch (func3)
{ {
@@ -312,7 +312,7 @@ void Instruction::executeOn(Warp &c) {
if ((func7 == 0)) if ((func7 == 0))
{ {
// SRLI // SRLI
// std::cout << "WTF\n"; // //std::cout << "WTF\n";
bool isNeg = ((0x80000000 & reg[rsrc[0]])) > 0; bool isNeg = ((0x80000000 & reg[rsrc[0]])) > 0;
Word result = Word_u(reg[rsrc[0]]) >> Word_u(immsrc); Word result = Word_u(reg[rsrc[0]]) >> Word_u(immsrc);
// if (isNeg) // if (isNeg)
@@ -332,7 +332,7 @@ void Instruction::executeOn(Warp &c) {
else else
{ {
// SRAI // SRAI
// std::cout << "WOHOOOOO\n"; // //std::cout << "WOHOOOOO\n";
op1 = reg[rsrc[0]]; op1 = reg[rsrc[0]];
op2 = immsrc; op2 = immsrc;
reg[rdest] = op1 >> op2; reg[rdest] = op1 >> op2;
@@ -345,23 +345,28 @@ void Instruction::executeOn(Warp &c) {
} }
break; break;
case S_INST: case S_INST:
std::cout << "S_INST\n"; //std::cout << "S_INST\n";
++c.stores; ++c.stores;
memAddr = reg[rsrc[0]] + immsrc; memAddr = reg[rsrc[0]] + immsrc;
// std::cout << "STORE MEM ADDRESS: " << std::hex << reg[rsrc[0]] << " + " << immsrc << "\n"; // //std::cout << "STORE MEM ADDRESS: " << std::hex << reg[rsrc[0]] << " + " << immsrc << "\n";
// std::cout << "FUNC3: " << func3 << "\n"; // //std::cout << "FUNC3: " << func3 << "\n";
if (memAddr == 0x00010000)
{
std::cout << (char) reg[rsrc[1]];
break;
}
switch (func3) switch (func3)
{ {
case 0: case 0:
// std::cout << "SB\n"; // //std::cout << "SB\n";
c.core->mem.write(memAddr, reg[rsrc[1]] & 0x000000FF, c.supervisorMode, 1); c.core->mem.write(memAddr, reg[rsrc[1]] & 0x000000FF, c.supervisorMode, 1);
break; break;
case 1: case 1:
// std::cout << "SH\n"; // //std::cout << "SH\n";
c.core->mem.write(memAddr, reg[rsrc[1]], c.supervisorMode, 2); c.core->mem.write(memAddr, reg[rsrc[1]], c.supervisorMode, 2);
break; break;
case 2: case 2:
// std::cout << std::hex << "SW: about to write: " << reg[rsrc[1]] << " to " << memAddr << "\n"; // //std::cout << std::hex << "SW: about to write: " << reg[rsrc[1]] << " to " << memAddr << "\n";
c.core->mem.write(memAddr, reg[rsrc[1]], c.supervisorMode, 4); c.core->mem.write(memAddr, reg[rsrc[1]], c.supervisorMode, 4);
break; break;
default: default:
@@ -375,7 +380,7 @@ void Instruction::executeOn(Warp &c) {
#endif #endif
break; break;
case B_INST: case B_INST:
std::cout << "B_INST\n"; //std::cout << "B_INST\n";
switch (func3) switch (func3)
{ {
case 0: case 0:
@@ -429,17 +434,17 @@ void Instruction::executeOn(Warp &c) {
} }
break; break;
case LUI_INST: case LUI_INST:
std::cout << "LUI_INST\n"; //std::cout << "LUI_INST\n";
reg[rdest] = (immsrc << 12) & 0xfffff000; reg[rdest] = (immsrc << 12) & 0xfffff000;
break; break;
case AUIPC_INST: case AUIPC_INST:
std::cout << "AUIPC_INST\n"; //std::cout << "AUIPC_INST\n";
reg[rdest] = ((immsrc << 12) & 0xfffff000) + (c.pc - 4); reg[rdest] = ((immsrc << 12) & 0xfffff000) + (c.pc - 4);
break; break;
case JAL_INST: case JAL_INST:
std::cout << "JAL_INST\n"; //std::cout << "JAL_INST\n";
if (!pcSet) nextPc = (c.pc - 4) + immsrc; if (!pcSet) nextPc = (c.pc - 4) + immsrc;
if (!pcSet) std::cout << "JAL... SETTING PC: " << nextPc << "\n"; if (!pcSet) //std::cout << "JAL... SETTING PC: " << nextPc << "\n";
if (rdest != 0) if (rdest != 0)
{ {
reg[rdest] = c.pc; reg[rdest] = c.pc;
@@ -447,9 +452,9 @@ void Instruction::executeOn(Warp &c) {
pcSet = true; pcSet = true;
break; break;
case JALR_INST: case JALR_INST:
std::cout << "JALR_INST\n"; //std::cout << "JALR_INST\n";
if (!pcSet) nextPc = reg[rsrc[0]] + immsrc; if (!pcSet) nextPc = reg[rsrc[0]] + immsrc;
if (!pcSet) std::cout << "JALR... SETTING PC: " << nextPc << "\n"; if (!pcSet) //std::cout << "JALR... SETTING PC: " << nextPc << "\n";
if (rdest != 0) if (rdest != 0)
{ {
reg[rdest] = c.pc; reg[rdest] = c.pc;
@@ -457,7 +462,7 @@ void Instruction::executeOn(Warp &c) {
pcSet = true; pcSet = true;
break; break;
case SYS_INST: case SYS_INST:
std::cout << "SYS_INST\n"; //std::cout << "SYS_INST\n";
temp = reg[rsrc[0]]; temp = reg[rsrc[0]];
switch (func3) switch (func3)
{ {
@@ -513,7 +518,7 @@ void Instruction::executeOn(Warp &c) {
case 0: case 0:
if (immsrc < 2) if (immsrc < 2)
{ {
std::cout << "INTERRUPT ECALL/EBREAK\n"; //std::cout << "INTERRUPT ECALL/EBREAK\n";
nextActiveThreads = 0; nextActiveThreads = 0;
c.interrupt(0); c.interrupt(0);
} }
@@ -523,16 +528,16 @@ void Instruction::executeOn(Warp &c) {
} }
break; break;
case TRAP: case TRAP:
std::cout << "INTERRUPT TRAP\n"; //std::cout << "INTERRUPT TRAP\n";
nextActiveThreads = 0; nextActiveThreads = 0;
c.interrupt(0); c.interrupt(0);
break; break;
case FENCE: case FENCE:
std::cout << "FENCE_INST\n"; //std::cout << "FENCE_INST\n";
break; break;
case PJ_INST: case PJ_INST:
// pred jump reg // pred jump reg
std::cout << "pred jump... src: " << rsrc[0] << std::hex << " val: " << reg[rsrc[0]] << " dest: " << reg[rsrc[1]] << "\n"; //std::cout << "pred jump... src: " << rsrc[0] << std::hex << " val: " << reg[rsrc[0]] << " dest: " << reg[rsrc[1]] << "\n";
if (reg[rsrc[0]]) if (reg[rsrc[0]])
{ {
if (!pcSet) nextPc = reg[rsrc[1]]; if (!pcSet) nextPc = reg[rsrc[1]];
@@ -540,24 +545,24 @@ void Instruction::executeOn(Warp &c) {
} }
break; break;
case GPGPU: case GPGPU:
std::cout << "GPGPU\n"; //std::cout << "GPGPU\n";
switch(func3) switch(func3)
{ {
case 0: case 0:
// WSPAWN // WSPAWN
std::cout << "WSPAWN\n"; //std::cout << "WSPAWN\n";
if (sjOnce) if (sjOnce)
{ {
sjOnce = false; sjOnce = false;
D(0, "Spawning a new warp."); D(0, "Spawning a new warp.");
// std::cout << "SIZE: " << c.core->w.size() << "\n"; // //std::cout << "SIZE: " << c.core->w.size() << "\n";
for (unsigned i = 0; i < c.core->w.size(); ++i) for (unsigned i = 0; i < c.core->w.size(); ++i)
{ {
// std::cout << "WHATTT\n"; // //std::cout << "WHATTT\n";
Warp &newWarp(c.core->w[i]); Warp &newWarp(c.core->w[i]);
// std::cout << "STARTING\n"; // //std::cout << "STARTING\n";
if (newWarp.spawned == false) { if (newWarp.spawned == false) {
// std::cout << "ABOUT TO START\n"; // //std::cout << "ABOUT TO START\n";
newWarp.pc = reg[rsrc[0]]; newWarp.pc = reg[rsrc[0]];
newWarp.reg[0] = reg; newWarp.reg[0] = reg;
newWarp.csr = c.csr; newWarp.csr = c.csr;
@@ -572,12 +577,12 @@ void Instruction::executeOn(Warp &c) {
case 2: case 2:
{ {
// SPLIT // SPLIT
std::cout << "SPLIT\n"; //std::cout << "SPLIT\n";
if (sjOnce) if (sjOnce)
{ {
sjOnce = false; sjOnce = false;
if (checkUnanimous(pred, c.reg, c.tmask)) { if (checkUnanimous(pred, c.reg, c.tmask)) {
std::cout << "Unanimous pred: " << pred << " val: " << reg[pred] << "\n"; //std::cout << "Unanimous pred: " << pred << " val: " << reg[pred] << "\n";
DomStackEntry e(c.tmask); DomStackEntry e(c.tmask);
e.uni = true; e.uni = true;
c.domStack.push(e); c.domStack.push(e);
@@ -595,7 +600,7 @@ void Instruction::executeOn(Warp &c) {
break; break;
case 3: case 3:
// JOIN // JOIN
std::cout << "JOIN\n"; //std::cout << "JOIN\n";
if (sjOnce) if (sjOnce)
{ {
sjOnce = false; sjOnce = false;
@@ -615,33 +620,34 @@ void Instruction::executeOn(Warp &c) {
break; break;
case 4: case 4:
// JMPRT // JMPRT
std::cout << "JMPRT\n"; //std::cout << "JMPRT\n";
nextActiveThreads = 1; nextActiveThreads = 1;
if (!pcSet) nextPc = reg[rsrc[0]]; if (!pcSet) nextPc = reg[rsrc[0]];
pcSet = true; pcSet = true;
break; break;
case 5: case 5:
// CLONE // CLONE
std::cout << "CLONE\n"; //std::cout << "CLONE\n";
// std::cout << "CLONING REG: " << rsrc[0] << " lane: " << reg[rsrc[0]] << "\n"; // //std::cout << "CLONING REG: " << rsrc[0] << " lane: " << reg[rsrc[0]] << "\n";
c.reg[reg[rsrc[0]]] = reg; c.reg[reg[rsrc[0]]] = reg;
break; break;
case 6: case 6:
// JALRS // JALRS
std::cout << "JALRS\n"; //std::cout << "JALRS\n";
nextActiveThreads = reg[rsrc[1]]; nextActiveThreads = reg[rsrc[1]];
reg[rdest] = c.pc; reg[rdest] = c.pc;
if (!pcSet) nextPc = reg[rsrc[0]]; if (!pcSet) nextPc = reg[rsrc[0]];
pcSet = true; pcSet = true;
// std::cout << "ACTIVE_THREDS: " << rsrc[1] << " val: " << reg[rsrc[1]] << "\n"; // //std::cout << "ACTIVE_THREDS: " << rsrc[1] << " val: " << reg[rsrc[1]] << "\n";
// std::cout << "nextPC: " << rsrc[0] << " val: " << std::hex << reg[rsrc[0]] << "\n"; // //std::cout << "nextPC: " << rsrc[0] << " val: " << std::hex << reg[rsrc[0]] << "\n";
break; break;
default: default:
cout << "ERROR: UNSUPPORTED GPGPU INSTRUCTION " << *this << "\n"; cout << "ERROR: UNSUPPORTED GPGPU INSTRUCTION " << *this << "\n";
} }
break; break;
default: default:
cout << "aERROR: Unsupported instruction: " << *this << "\n"; cout << "pc: " << hex << (c.pc) << "\n";
cout << "aERROR: Unsupported instruction: " << *this << "\n" << flush;
exit(1); exit(1);
} }
} }
@@ -660,9 +666,9 @@ void Instruction::executeOn(Warp &c) {
// std::cout << "new thread mask: "; // //std::cout << "new thread mask: ";
// for (int i = 0; i < c.tmask.size(); ++i) std::cout << " " << c.tmask[i]; // for (int i = 0; i < c.tmask.size(); ++i) //std::cout << " " << c.tmask[i];
// std::cout << "\n"; // //std::cout << "\n";
// This way, if pc was set by a side effect (such as interrupt), it will // This way, if pc was set by a side effect (such as interrupt), it will
// retain its new value. // retain its new value.

24
src/obj.cpp
View File

@@ -1,21 +1,21 @@
/******************************************************************************* /*******************************************************************************
HARPtools by Chad D. Kersey, Summer 2011 HARPtools by Chad D. Kersey, Summer 2011
*******************************************************************************/ *******************************************************************************/
#include "include/types.h" // #include "include/types.h"
#include "include/obj.h" // #include "include/obj.h"
#include "include/util.h" // #include "include/util.h"
#include "include/asm-tokens.h" // #include "include/asm-tokens.h"
#include "include/debug.h" // #include "include/debug.h"
#include <iostream> // #include <iostream>
#include <stdlib.h> // #include <stdlib.h>
#include <FlexLexer.h> // #include <FlexLexer.h>
#include <cctype> // #include <cctype>
#include <cstdio> // #include <cstdio>
#include <cmath> // #include <cmath>
#include <map> // #include <map>
// using namespace std; // using namespace std;
// using namespace Harp; // using namespace Harp;

297672
src/results.txt
View File

File diff suppressed because it is too large Load Diff

33
src/riscv_gpgpu/gpgpu_test.c
View File

@@ -1,6 +1,7 @@
// #include <stdint.h> // #include <stdint.h>
#include <stdbool.h>
// #include <cstdint> // #include <cstdint>
extern void print_consol(char *);
extern void printc(char);
int main(void); int main(void);
@@ -13,24 +14,39 @@ void matAddition (unsigned, unsigned);
// unsigned y[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; // unsigned y[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
// unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; // unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
unsigned x[] = {1, 1, 6, 0, 3, 1, 1, 2, 0, 3, 6, 7, 5, 7, 7, 9}; unsigned x[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 , 1 , 1 , 1 , 1 , 1 };
unsigned y[] = {0, 2, 2, 0, 5, 0, 1, 1, 4, 2, 0, 0, 3, 2, 3, 2}; unsigned y[] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15};
unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
// unsigned x[] = {2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2}; // unsigned x[] = {2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2};
// unsigned y[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; // unsigned y[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
// unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; // unsigned z[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
#define NUM_WARPS 3 #define NUM_WARPS 2
#define NUM_THREADS 7 #define NUM_THREADS 8
int main() int main()
{ {
for (int i = 0; i < 8; i++)
{
queue_initialize(q + i);
}
createWarps(NUM_WARPS, NUM_THREADS, matAddition, x, y, z); createWarps(NUM_WARPS, NUM_THREADS, matAddition, x, y, z);
while(!queue_isEmpty()) {} wait_for_done(NUM_WARPS);
print_consol("-------------------------\n");
print_consol("FINAL Z\n");
for (int i = 0; i < 16; i++)
{
int_print(i);
print_consol(": ");
int_print(z[i]);
print_consol("\n");
}
print_consol("-------------------------------\n");
return 0; return 0;
} }
@@ -44,14 +60,11 @@ void matAddition(unsigned tid, unsigned wid)
unsigned i = (wid * NUM_THREADS) + tid; unsigned i = (wid * NUM_THREADS) + tid;
__if((i < 10)) __if((i < 11))
z_ptr[i] = x_ptr[i] + y_ptr[i]; z_ptr[i] = x_ptr[i] + y_ptr[i];
__else __else
__end_if __end_if
sleep((50 * (wid + wid))+100);
return; return;
} }

1321
src/riscv_gpgpu/gpgpu_test.dump
View File

File diff suppressed because it is too large Load Diff

BIN
src/riscv_gpgpu/gpgpu_test.elf
View File

Binary file not shown.

279
src/riscv_gpgpu/gpgpu_test.hex
View File

@@ -1,104 +1,189 @@
:0200000480007A :0200000480007A
:1000000037F1FF7FEF00401DEF0080067300000016 :1000000037F1FF7FEF00C00C73000000938B0600F8
:10001000938B0600130C0700938C0700130F01004D :10001000130C0700938C0700130D0800130F010049
:100020009303050013051000635C75001301018044 :100020009303050013051000635C75001301018044
:10003000130305006B500300130515006FF0DFFE7E :10003000130305006B500300130515006FF0DFFE7E
:1000400013010F0013050000930F0600938D0300AA :1000400013010F0013050000930F0600938D0300AA
:10005000EBE0BF01170500001305052F6B400500FD :10005000EBE0BF01170500001305854B6B40050061
:1000600017030000130303FB6B000300678000000D :10006000B708010023A0B800678000001703000054
:10007000130101FF23261100232481001304010131 :10007000130303FA6B00030067800000130141FFC4
:10008000B707008193870700370700811307C76010 :10008000232011002322B100834505006388050069
:10009000B70600819386C65C370600801306460DBE :10009000EFF01FFD130515006FF01FFF8320010017
:1000A0009305700013053000EF00803D1300000041 :1000A000832541001301C10067800000130141FF57
:1000B000EF00C02593070500E38C07FE93070000BF :1000B000232011002322B10093050503EFF05FFA1E
:1000C000138507008320C10003248100130101016F :1000C00083200100832541001301C10067800000E7
:1000D00067800000130101FC232E1102232C8102F2 :1000D000130101FE232E1100232C810013040102C1
:1000E000130401042326A4FC2324B4FCEF000045E0 :1000E000232604FE6F00C0040327C4FE9307070005
:1000F0002326A4FEEF0000472324A4FEEF000049BE :1000F00093973700B387E74093972700B387E7408C
:100100002322A4FE032784FC93070700939737005C :10010000139747003307F74093172700138707001B
:10011000B387E7400327C4FCB307F7002320F4FEAE :10011000B707008193874724B307F70013850700CB
:10012000832704FE93B7A70093C71700A30FF4FC1F :10012000EF00401C8327C4FE938717002326F4FEAC
:100130008347F4FD138F0700B7070080938F47199B :100130000327C4FE93077000E3D8E7FAB7070081EE
:100140006B200F007B70FF01832704FE939727002D :100140009387871F3707008113078717B70600813A
:100150000327C4FEB307F70083A60700832704FE26 :1001500093868613370600801306062193058000D8
:1001600093972700032784FEB307F70003A7070030 :1001600013052000EF00406013052000EF00007130
:10017000832704FE93972700032644FEB307F60067 :10017000B707008113850704EFF05FF0B707008130
:100180003387E60023A0E700B7070080138E8719A6 :100180001385C705EFF09FEF232404FE6F004005A1
:1001900067000E00130000006B300000832784FC12 :10019000832784FE13850700EF00002BB70700813B
:1001A000138717009307070093971700B387E7009B :1001A00013858706EFF09FEDB7070081032784FED4
:1001B00093973700B387E7009397270013850700CD :1001B000131727009387871FB307F70083A707004C
:1001C000EF000028130000008320C10303248103F3 :1001C00013850700EF004028B70700811385C70695
:1001D000130101046780000097020001938282E608 :1001D000EFF0DFEA832784FE938717002324F4FEE1
:1001E000130300009303700023A0620023A26200A7 :1001E000032784FE9307F000E3D4E7FAB707008102
:1001F00023A4620023A6720023A862006780000087 :1001F00013850707EFF09FE89307000013850700BA
:1002000097020001938202E403A382001303130008 :100200008320C101032481011301010267800000E2
:1002100023A462001383420183AE420093935E00E5 :10021000130101FC232E1102232C8102130401047B
:1002200033037300032E05002320C301032E450072 :100220002326A4FC2324B4FCEF00C06E2326A4FEE6
:100230002322C301032E85002324C301032EC500FE :10023000EF00C0702324A4FEEF00C0722322A4FEAE
:100240002326C301032E05012328C301032E4501E4 :10024000832784FC939737000327C4FCB307F70088
:10025000232AC301032E8501232CC301938E1E0084 :100250002320F4FE832704FE93B7B70093C717004B
:10026000130F20036394EE01930E000023A2D2012A :10026000A30FF4FC8347F4FD138F0700B70700804A
:1002700067800000970200019382C2DC03A3820022 :10027000938F872C6B200F007B70FF01832704FE78
:100280001303F3FF23A462001383420183AE020031 :10028000939727000327C4FEB307F70083A6070050
:10029000930F2003138F0E00130F1F006314FF0131 :10029000832704FE93972700032784FEB307F70004
:1002A000130F000023A0E20193935E003303730059 :1002A00003A70700832704FE93972700032644FE35
:1002B000032E03002320C501032E43002322C50182 :1002B000B307F6003387E60023A0E700B707008006
:1002C000032E83002324C501032EC3002326C5016A :1002C000138EC72C67000E00130000006B30000077
:1002D000032E03012328C501032E4301232AC50150 :1002D000130000008320C1030324810313010104E0
:1002E000032E8301232CC5016780000097020001C3 :1002E0006780000093020500130300009303700071
:1002F000938242D503A3820013050000130E20034E :1002F00023A0620023A2620023A4620023A672004E
:1003000063146E001305150067800000970200015A :1003000023A86200678000009302050003A3820017
:10031000938242D303A3820013050000130E000052 :100310001303130023A462001383420183AE42003F
:1003200063146E001305150067800000970200013A :1003200093935E003303730003AE05002320C301E3
:10033000938242D103A3C20083A3020133B56300B9 :1003300003AE45002322C30103AE85002324C3017D
:1003400067800000130101FD23261102232481028E :1003400003AEC5002326C30103AE05012328C30164
:1003500013040103EFF09FFB93070500638407007C :1003500003AE4501232AC30103AE8501232CC3014B
:1003600073000000930744FD13850700EFF09FF032 :1003600003AEC501232EC301938E1E00130F20037D
:100370008327C4FD13810700032584FD832544FDE5 :100370006394EE01930E000023A2D2016780000077
:10038000032604FE832644FE032784FE8327C4FE3F :100380009302050003A382001303F3FF23A462007A
:10039000EFF01FC873000000130000008320C102AB :100390001383420183AE0200930F2003138F0E00DC
:1003A000032481021301010367800000130101FD92 :1003A000130F1F006314FF01130F000023A0E201CD
:1003B000232611022324810213040103930901005F :1003B00093935E0033037300032E030023A0C50153
:1003C0006F000005B70700819387070483A7070123 :1003C000032E430023A2C501032E830023A4C501ED
:1003D00013871700B70700819387070423A8E70056 :1003D000032EC30023A6C501032E030123A8C501D4
:1003E000930744FD13850700EFF0DFE88327C4FD82 :1003E000032E430123AAC501032E830123ACC501BB
:1003F00013810700032584FD832544FD032604FEA5 :1003F000032EC30123AEC5016780000093020500F0
:10040000832644FE032784FE8327C4FEEFF05FC5E6 :1004000003A3820013050000130E200363146E0083
:10041000EFF0DFEF9307050063980700EFF01FF19F :1004100013051500678000009302050003A3820006
:1004200093070500E39007FA138109001300000009 :1004200013050000130E000063146E001305150081
:100430008320C102032481021301010367800000AD :10043000678000009302050003A3C20083A30201AA
:10044000130101FD2326810213040103232EA4FCC2 :1004400033B5630067800000130101FD232611020C
:10045000232604FE6F0000018327C4FE9387170044 :100450002324810213040103232EA4FC0327C4FDDB
:100460002326F4FE0327C4FE8327C4FDE346F7FEDC :100460009307F00063E4E702B70700810327C4FDA8
:10047000130000000324C102130101036780000080 :10047000131727009387871BB307F70083A707008D
:10048000130101FB23261104232481041304010515 :1004800013850700EFF09FBF6F004007930700023E
:100490002326A4FC2324B4FC2322C4FC2320D4FC64 :100490002326F4FEA30504FE8327C4FE9387C7FF2B
:1004A000232EE4FA232CF4FA13090100232604FE78 :1004A0000327C4FDB357F70093F7F7002322F4FEA8
:1004B0006F00C005B709FFFF330131018327C4FE78 :1004B000832744FE6386070093071000A305F4FE1C
:1004C0002328F4FC832784FC232AF4FC93070100EF :1004C0008347B4FE63820702B7070081032744FE17
:1004D000232CF4FC832744FC232EF4FC832704FC08 :1004D000131727009387871BB307F70083A707002D
:1004E0002320F4FE8327C4FB2322F4FE832784FB0E :1004E00013850700EFF09FB98327C4FE9387C7FFEA
:1004F0002324F4FE930704FD13850700EFF05FD07B :1004F0002326F4FE8327C4FEE340F0FA8320C102E2
:100500008327C4FE938717002326F4FE0327C4FE27 :10050000032481021301010367800000130101FD30
:100510008327C4FCE360F7FA13010900EFF01FE939 :1005100023261102232481022322A10313040103B1
:1005200013054006EFF0DFF1130000008320C10443 :1005200013070D009307070093973700B387E74041
:10053000032481041301010567800000130101FFFA :1005300093972700B387E740139747003307F740A7
:1005400023268100232471011304010193870B00EA :100540009317270013870700B70700819387472475
:10055000138507000324C100832B810013010101CF :10055000B307F70013850700EFF01FEC93070500C2
:1005600067800000130101FF2326810023248101FD :1005600063820704B70700811385470DEFF01FB1C1
:100570001304010193070C00138507000324C10035 :1005700093070D0013850700EFF01FEDB70700810B
:10058000032C81001301010167800000130101FFAA :100580001385C70DEFF09FAF13070D00B73700813C
:1005900023268100232491011304010193870C0079 :100590009387474EB307F700130710002380E70047
:1005A000138507000324C100832C8100130101017E :1005A0007300000013070D009307070093973700AF
:0405B0006780000060 :1005B000B387E74093972700B387E7401397470037
:1005C0003307F7409317270013870700B707008109
:1005D00093874724B307F700130704FD930507002B
:1005E00013850700EFF0DFD9832784FD138107000F
:1005F000032544FD832504FD0326C4FD832604FE54
:10060000032744FE832784FE0328C4FEEFF01FA0C7
:1006100073000000130000008320C1020324810244
:10062000032D41021301010367800000130101FC47
:10063000232E1102232C81021304010493090100CB
:10064000232604FE6F00000C0327C4FE9307070057
:1006500093973700B387E74093972700B387E74026
:10066000139747003307F7409317270013870700B6
:10067000B707008193874724B307F7001385070066
:10068000EFF09FD993070500639807060327C4FE80
:100690009307070093973700B387E74093972700A6
:1006A000B387E740139747003307F74093172700B6
:1006B00013870700B707008193874724B307F70024
:1006C0001307C4FC9305070013850700EFF05FCB09
:1006D000832744FD13810700032504FD8325C4FC03
:1006E000032684FD8326C4FD032704FE832744FEDE
:1006F000032884FEEFF09F978327C4FE938717009B
:100700002326F4FE0327C4FE93076000E3DEE7F22E
:1007100013810900130000008320C1030324810317
:100720001301010467800000130101FD23268102EB
:1007300013040103232EA4FC232604FE6F000001F2
:100740008327C4FE938717002326F4FE0327C4FEE5
:100750008327C4FDE346F7FE130000000324C10213
:100760001301010367800000130101FA232E110415
:10077000232C810413040106232EA4FA232CB4FA9B
:10078000232AC4FA2328D4FA2326E4FA2324F4FAE9
:1007900013090100232604FE232404FE6F00800BAE
:1007A000B709FFFF33013101832784FE2324F4FCC2
:1007B000832784FB2326F4FC930701002328F4FC01
:1007C000832744FB232AF4FC832704FB232CF4FC1B
:1007D0008327C4FA232EF4FC832784FA2320F4FE13
:1007E0008327C4FE2322F4FE0327C4FE93070700D9
:1007F00093973700B387E74093972700B387E74085
:10080000139747003307F740931727001387070014
:10081000B707008193874724B307F700130784FCC9
:100820009305070013850700EFF01FAE8327C4FE72
:10083000938717002326F4FE0327C4FE9307600066
:1008400063D4E700232604FE832784FE93871700E2
:100850002324F4FE032784FE8327C4FBE362F7F41A
:1008600013010900EFF09FDC130000008320C10595
:10087000032481051301010667800000130101FDB7
:10088000232611022324810213040103232EA4FC36
:10089000B70700811385070EEFF04FFEA30704FE94
:1008A0006F00000593071000A307F4FE232404FE45
:1008B0006F0040038347F4FE373700819306474EAD
:1008C000032784FE3387E60003470700B3F7E700FA
:1008D000B337F000A307F4FE832784FE9387170045
:1008E0002324F4FE832784FE0327C4FDE3E4E7FC0E
:1008F0008347F4FE93C7170093F7F70FE39407FAC3
:10090000130000008320C1020324810213010103AC
:1009100067800000130101FF232681002324710159
:100920001304010193870B00138507000324C10002
:10093000832B81001301010167800000130101FF77
:1009400023268100232481011304010193070C0055
:10095000138507000324C100032C8100130101014A
:1009600067800000130101FF2326810023249101E9
:100970001304010193870C00138507000324C100B1
:0C098000832C810013010101678000003E
:02000004810079 :02000004810079
:1005CC000100000001000000060000000000000017 :10000000300000003100000032000000330000002A
:1005DC000300000001000000010000000200000008 :10001000340000003500000036000000370000000A
:1005EC0000000000030000000600000007000000EF :10002000380000003900000061000000620000009C
:1005FC0005000000070000000700000009000000D3 :10003000630000006400000065000000660000002E
:10060C0000000000020000000200000000000000DA :100040002D2D2D2D2D2D2D2D2D2D2D2D2D2D2D2DE0
:10061C0005000000000000000100000001000000C7 :100050002D2D2D2D2D2D2D2D2D0A000046494E41E3
:10062C0004000000020000000000000000000000B8 :100060004C205A0A000000003A2000000A0000005C
:10063C0003000000020000000300000002000000A4 :100070002D2D2D2D2D2D2D2D2D2D2D2D2D2D2D2DB0
:100080002D2D2D2D2D2D2D2D2D2D2D2D2D2D2D0AC3
:1000900000000000300000003100000032000000CD
:1000A000330000003400000035000000360000007E
:1000B0003700000038000000390000006100000037
:1000C00062000000630000006400000065000000A2
:1000D00066000000646F6E653A2000000A000000B0
:1000E00061626F757420746F207761697420666F28
:0800F0007220646F6E650A00C6
:1000F8000000008104000081080000810C000081DC
:100108001000008114000081180000811C0000818B
:100118002000008124000081280000812C0000813B
:100128003000008134000081380000813C000081EB
:1001380001000000010000000100000001000000B3
:1001480001000000010000000100000001000000A3
:100158000100000001000000010000000100000093
:100168000100000001000000010000000100000083
:100178000000000001000000020000000300000071
:100188000400000005000000060000000700000051
:1001980008000000090000000A0000000B00000031
:1001A8000C0000000D0000000E0000000F00000011
:1001B80094000081980000819C000081A0000081CB
:1001C800A4000081A8000081AC000081B00000817B
:1001D800B4000081B8000081BC000081C00000812B
:1001E800C4000081C8000081CC000081D0000081DB
:040000058000000077 :040000058000000077
:00000001FF :00000001FF

81
src/riscv_gpgpu/lib/lib.c
View File

@@ -1,22 +1,49 @@
#include "lib.h" #include "lib.h"
extern void createThreads(unsigned, unsigned, unsigned, unsigned *, unsigned *, unsigned *); extern void createThreads(unsigned, unsigned, unsigned, unsigned *, unsigned *, unsigned *, unsigned);
extern void wspawn(unsigned, unsigned, unsigned, unsigned *, unsigned *, unsigned *); extern void wspawn(unsigned, unsigned, unsigned, unsigned *, unsigned *, unsigned *, unsigned);
extern void print_consol(char *);
extern void printc(char);
void int_print(unsigned f)
{
if (f < 16)
{
print_consol(hextoa[f]);
return;
}
int temp;
int sf = 32;
bool start = false;
do
{
temp = (f >> (sf - 4)) & 0xf;
if (temp != 0) start = true;
if (start) print_consol(hextoa[temp]);
sf -= 4;
} while(sf > 0);
}
void reschedule_warps() void reschedule_warps()
{ {
if (queue_isEmpty()) register unsigned curr_warp asm("s10");
if (queue_isEmpty(q+curr_warp))
{ {
print_consol("done: ");
int_print(curr_warp);
print_consol("\n");
done[curr_warp] = true;
ECALL; ECALL;
} }
Job j; Job j;
queue_dequeue(&j); queue_dequeue(q+curr_warp,&j);
asm __volatile__("mv sp,%0"::"r" (j.base_sp):); asm __volatile__("mv sp,%0"::"r" (j.base_sp):);
createThreads(j.n_threads, j.wid, j.func_ptr, j.x, j.y, j.z); createThreads(j.n_threads, j.wid, j.func_ptr, j.x, j.y, j.z, j.assigned_warp);
ECALL; ECALL;
@@ -25,16 +52,20 @@ void reschedule_warps()
void schedule_warps() void schedule_warps()
{ {
asm __volatile__("mv s3, sp"); asm __volatile__("mv s3, sp");
while (!queue_isEmpty() && queue_availableWarps())
{
++q.active_warps;
Job j;
queue_dequeue(&j);
asm __volatile__("mv sp,%0"::"r" (j.base_sp):); for (int curr_warp = 0; curr_warp < 7; ++curr_warp)
wspawn(j.n_threads, j.wid, j.func_ptr, j.x, j.y, j.z); {
if (!queue_isEmpty(q+curr_warp))
{
Job j;
queue_dequeue(q+curr_warp,&j);
asm __volatile__("mv sp,%0"::"r" (j.base_sp):);
wspawn(j.n_threads, j.wid, j.func_ptr, j.x, j.y, j.z, j.assigned_warp);
}
} }
asm __volatile__("mv sp, s3"); asm __volatile__("mv sp, s3");
} }
void sleep(int t) void sleep(int t)
@@ -46,8 +77,8 @@ void sleep(int t)
void createWarps(unsigned num_Warps, unsigned num_threads, FUNC, unsigned * x_ptr, unsigned * y_ptr, unsigned * z_ptr) void createWarps(unsigned num_Warps, unsigned num_threads, FUNC, unsigned * x_ptr, unsigned * y_ptr, unsigned * z_ptr)
{ {
asm __volatile__("addi s2, sp, 0"); asm __volatile__("addi s2, sp, 0");
int warp = 0;
for (unsigned i = 0; i < num_Warps; i++) for (unsigned i = 0; i < num_Warps; i++)
{ {
asm __volatile__("lui s3, 0xFFFF0"); asm __volatile__("lui s3, 0xFFFF0");
@@ -62,23 +93,33 @@ void createWarps(unsigned num_Warps, unsigned num_threads, FUNC, unsigned * x_pt
j.x = x_ptr; j.x = x_ptr;
j.y = y_ptr; j.y = y_ptr;
j.z = z_ptr; j.z = z_ptr;
j.assigned_warp = warp;
queue_enqueue(&j); queue_enqueue(q + warp,&j);
++warp;
if (warp >= 7) warp = 0;
} }
asm __volatile__("addi sp, s2, 0"); asm __volatile__("addi sp, s2, 0");
schedule_warps(); schedule_warps();
sleep(100);
} }
void wait_for_done(unsigned num_wait)
{
print_consol("about to wait for done\n");
bool temp = false;
while (!temp)
{
temp = true;
for (int i = 0; i < num_wait; i++)
{
temp &= done[i];
}
}
}
// unsigned get_wid()
// {
// register unsigned ret asm("s7");
// return ret;
// }
unsigned * get_1st_arg(void) unsigned * get_1st_arg(void)
{ {

9
src/riscv_gpgpu/lib/lib.h
View File

@@ -1,7 +1,7 @@
#ifndef __RISCV_GP_ #ifndef __RISCV_GP_
#define __RISCV_GP_ #define __RISCV_GP_
#include <stdbool.h>
#include "queue.h" #include "queue.h"
#define WSPAWN asm __volatile__(".word 0x3006b"::); #define WSPAWN asm __volatile__(".word 0x3006b"::);
@@ -29,9 +29,16 @@
JOIN; JOIN;
static char * hextoa[] = {"0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "a", "b", "c", "d", "e", "f"};
static bool done[] = {false, false, false, false, false, false, false};
static int main_sp[1];
#define FUNC void (func)(unsigned, unsigned) #define FUNC void (func)(unsigned, unsigned)
void createWarps(unsigned num_Warps, unsigned num_threads, FUNC, unsigned *, unsigned *, unsigned *); void createWarps(unsigned num_Warps, unsigned num_threads, FUNC, unsigned *, unsigned *, unsigned *);
void reschedule_warps(void); void reschedule_warps(void);
void int_print(unsigned);
void wait_for_done(unsigned);
unsigned * get_1st_arg(void); unsigned * get_1st_arg(void);
unsigned * get_2nd_arg(void); unsigned * get_2nd_arg(void);

49
src/riscv_gpgpu/lib/lib.s
View File

@@ -7,18 +7,18 @@
.global _start .global _start
_start: _start:
lui sp, 0x7ffff lui sp, 0x7ffff
jal queue_initialize
jal main jal main
ecall ecall
.type createThreads, @function .type createThreads, @function
.global createThreads .global createThreads
createThreads: createThreads:
mv s7,a3 # Moving x_ptr to s7 mv s7 ,a3 # Moving x_ptr to s7
mv s8,a4 # Moving y_ptr to s8 mv s8 ,a4 # Moving y_ptr to s8
mv s9,a5 # Moving z_ptr to s9 mv s9 ,a5 # Moving z_ptr to s9
mv t5,sp # Saving the current stack pointer to t5 mv s10,a6 # Moving assigned_warp to s10
mv t2, a0 # t2 = num_threads mv t5 ,sp # Saving the current stack pointer to t5
mv t2 , a0 # t2 = num_threads
loop_init: loop_init:
li a0,1 # i = 0 li a0,1 # i = 0
loop_cond: loop_cond:
@@ -39,6 +39,12 @@ loop_done:
la a0, reschedule_warps la a0, reschedule_warps
.word 0x5406b .word 0x5406b
.type printc, @function
.global printc
printc:
la a7, 0x00010000
sw a1, 0(a7)
ret
.type wspawn, @function .type wspawn, @function
@@ -47,3 +53,34 @@ wspawn:
la t1, createThreads la t1, createThreads
.word 0x3006b # WSPAWN instruction .word 0x3006b # WSPAWN instruction
ret ret
.type print_consol, @function
.global print_consol
print_consol:
addi sp, sp, -12
sw ra, 0(sp)
sw a1, 4(sp)
bl:
lbu a1,0(a0)
beqz a1,be
jal printc
addi a0, a0, 1
j bl
be:
lw ra, 0(sp)
lw a1, 4(sp)
addi sp, sp, 12
ret
.type print_int, @function
.global print_int
print_int:
addi sp, sp, -12
sw ra, 0(sp)
sw a1, 4(sp)
addi a1, a0, 48
jal printc
lw ra, 0(sp)
lw a1, 4(sp)
addi sp, sp, 12
ret

15
src/riscv_gpgpu/lib/queue.h
View File

@@ -18,6 +18,7 @@ typedef struct Job_t
unsigned * x; unsigned * x;
unsigned * y; unsigned * y;
unsigned * z; unsigned * z;
unsigned assigned_warp;
} Job; } Job;
@@ -32,17 +33,17 @@ typedef struct Queue_t
} Queue; } Queue;
Queue q; Queue q[8];
void queue_initialize(void); void queue_initialize(Queue *);
void queue_enqueue(Job *); void queue_enqueue(Queue *, Job *);
void queue_dequeue(Job *); void queue_dequeue(Queue *, Job *);
int queue_isFull(void); int queue_isFull(Queue *);
int queue_isEmpty(void); int queue_isEmpty(Queue *);
int queue_availableWarps(); int queue_availableWarps(Queue *);
void func(); void func();

44
src/riscv_gpgpu/lib/queue.s
View File

@@ -7,7 +7,7 @@
.type queue_initialize, @function .type queue_initialize, @function
.global queue_initialize .global queue_initialize
queue_initialize: queue_initialize:
la t0, q # loading base address of q mv t0, a0 # loading base address of q
li t1, 0 # to initialize variables li t1, 0 # to initialize variables
li t2, A_WARPS # Num of available warps li t2, A_WARPS # Num of available warps
sw t1, 0 (t0) # start_i sw t1, 0 (t0) # start_i
@@ -23,7 +23,7 @@ queue_initialize:
.type queue_enqueue, @function .type queue_enqueue, @function
.global queue_enqueue .global queue_enqueue
queue_enqueue: queue_enqueue:
la t0, q # loading base address of q mv t0, a0 # loding base address of q
lw t1, 8 (t0) # t1 = num_j lw t1, 8 (t0) # t1 = num_j
addi t1, t1, 1 # ++t1 addi t1, t1, 1 # ++t1
sw t1, 8 (t0) # num_j = t1 sw t1, 8 (t0) # num_j = t1
@@ -31,20 +31,22 @@ queue_enqueue:
lw t4, 4 (t0) # t4 = end_i lw t4, 4 (t0) # t4 = end_i
slli t2, t4, 5 # index * 32 [log(sizeof(job))] slli t2, t4, 5 # index * 32 [log(sizeof(job))]
add t1, t1, t2 # jobs + index add t1, t1, t2 # jobs + index
lw t3, 0 (a0) # wid lw t3, 0 (a1) # wid
sw t3, 0 (t1) # sw t3, 0 (t1) #
lw t3, 4 (a0) # n_threads lw t3, 4 (a1) # n_threads
sw t3, 4 (t1) # sw t3, 4 (t1) #
lw t3, 8 (a0) # base_sp lw t3, 8 (a1) # base_sp
sw t3, 8 (t1) # sw t3, 8 (t1) #
lw t3, 12(a0) # func_ptr lw t3, 12(a1) # func_ptr
sw t3, 12(t1) # sw t3, 12(t1) #
lw t3, 16(a0) # x lw t3, 16(a1) # x
sw t3, 16(t1) # sw t3, 16(t1) #
lw t3, 20(a0) # y lw t3, 20(a1) # y
sw t3, 20(t1) # sw t3, 20(t1) #
lw t3, 24(a0) # z lw t3, 24(a1) # z
sw t3, 24(t1) # sw t3, 24(t1) #
lw t3, 28(a1) # assigned_warp
sw t3, 28(t1) #
addi t4, t4, 1 # end_i++ addi t4, t4, 1 # end_i++
li t5, SIZE # size li t5, SIZE # size
bne t4, t5, ec # if ((q.end_i + 1) == SIZE) bne t4, t5, ec # if ((q.end_i + 1) == SIZE)
@@ -58,7 +60,7 @@ ec:
.global queue_dequeue .global queue_dequeue
queue_dequeue: queue_dequeue:
la t0, q # loading base address of q mv t0, a0 # loading base address of q
lw t1, 8 (t0) # t1 = num_j lw t1, 8 (t0) # t1 = num_j
addi t1, t1, -1 # --t1 addi t1, t1, -1 # --t1
sw t1, 8 (t0) # num_j = t1 sw t1, 8 (t0) # num_j = t1
@@ -74,26 +76,28 @@ dc:
slli t2, t4, 5 # index * 32 [log(sizeof(job))] slli t2, t4, 5 # index * 32 [log(sizeof(job))]
add t1, t1, t2 # jobs + index add t1, t1, t2 # jobs + index
lw t3, 0 (t1) # wid lw t3, 0 (t1) # wid
sw t3, 0 (a0) # sw t3, 0 (a1) #
lw t3, 4 (t1) # n_threads lw t3, 4 (t1) # n_threads
sw t3, 4 (a0) # sw t3, 4 (a1) #
lw t3, 8 (t1) # base_sp lw t3, 8 (t1) # base_sp
sw t3, 8 (a0) # sw t3, 8 (a1) #
lw t3, 12(t1) # func_ptr lw t3, 12(t1) # func_ptr
sw t3, 12(a0) # sw t3, 12(a1) #
lw t3, 16(t1) # x lw t3, 16(t1) # x
sw t3, 16(a0) # sw t3, 16(a1) #
lw t3, 20(t1) # y lw t3, 20(t1) # y
sw t3, 20(a0) # sw t3, 20(a1) #
lw t3, 24(t1) # z lw t3, 24(t1) # z
sw t3, 24(a0) # sw t3, 24(a1) #
lw t3, 28(t1) # assigned_warp
sw t3, 28(a1) #
ret ret
.type queue_isFull, @function .type queue_isFull, @function
.global queue_isFull .global queue_isFull
queue_isFull: queue_isFull:
la t0, q # loading base address of q mv t0, a0 # loading base address of q
lw t1, 8 (t0) # t1 = num_j lw t1, 8 (t0) # t1 = num_j
mv a0, zero # ret_val = 0 mv a0, zero # ret_val = 0
li t3, SIZE # t3 = SIZE li t3, SIZE # t3 = SIZE
@@ -107,7 +111,7 @@ qf:
.type queue_isEmpty, @function .type queue_isEmpty, @function
.global queue_isEmpty .global queue_isEmpty
queue_isEmpty: queue_isEmpty:
la t0, q # loading base address of q mv t0, a0 # loading base address of q
lw t1, 8 (t0) # t1 = num_j lw t1, 8 (t0) # t1 = num_j
mv a0, zero # ret_val = 0 mv a0, zero # ret_val = 0
mv t3, zero # t3 = 0 mv t3, zero # t3 = 0
@@ -120,7 +124,7 @@ qe:
.type queue_availableWarps, @function .type queue_availableWarps, @function
.global queue_availableWarps .global queue_availableWarps
queue_availableWarps: queue_availableWarps:
la t0, q # loading base address of q mv t0, a0 # loading base address of q
lw t1, 12(t0) # t1 = total_warps lw t1, 12(t0) # t1 = total_warps
lw t2, 16(t0) # t2 = active_warps lw t2, 16(t0) # t2 = active_warps
sltu a0, t2, t1 sltu a0, t2, t1

4
src/riscv_gpgpu/linker.ld
View File

@@ -35,6 +35,10 @@ SECTIONS
PROVIDE (__etext = .); PROVIDE (__etext = .);
PROVIDE (_etext = .); PROVIDE (_etext = .);
PROVIDE (etext = .); PROVIDE (etext = .);
PROVIDE (_edata = .);
PROVIDE (_end = .);
PROVIDE (__global_pointer$ = .);
. = 0x81000000; . = 0x81000000;
.rodata : { *(.rodata .rodata.* .gnu.linkonce.r.*) } .rodata : { *(.rodata .rodata.* .gnu.linkonce.r.*) }
.rodata1 : { *(.rodata1) } .rodata1 : { *(.rodata1) }

4
src/test_riscv.sh
View File

@@ -1,4 +1,4 @@
echo start > results.txt echo start > results.txt
echo ./riscv_gpgpu/gpgpu_test.hex >> results.txt echo ./riscv_gpgpu/gpgpu_test.hex
./harptool -E -a rv32i --core ./riscv_gpgpu/gpgpu_test.hex -s -b ./harptool -E -a rv32i --core ./riscv_gpgpu/gpgpu_test.hex -s -b