Improve Verilation memory by reducing V3Number size (#3521)

This commit is contained in:
Mariusz Glebocki
2022-09-20 16:46:47 -04:00
committed by GitHub
parent bba800f2d6
commit fc3ce29845
4 changed files with 532 additions and 220 deletions
+141 -128
View File
@@ -44,7 +44,7 @@ constexpr int MAX_SPRINTF_DOUBLE_SIZE
"Number operation called with same source and dest")
#define NUM_ASSERT_LOGIC_ARGS1(arg1) \
UASSERT((!(arg1).isDouble() && !(arg1).isString()), \
UASSERT(((arg1).dataType() == V3NumberData::V3NumberDataType::LOGIC), \
"Number operation called with non-logic (double or string) argument: '" << (arg1) \
<< '"')
#define NUM_ASSERT_LOGIC_ARGS2(arg1, arg2) \
@@ -79,8 +79,10 @@ void V3Number::v3errorEnd(std::ostringstream& str) const {
nsstr << str.str();
if (m_nodep) {
m_nodep->v3errorEnd(nsstr);
} else {
} else if (m_fileline) {
m_fileline->v3errorEnd(nsstr);
} else {
V3Error::v3errorEnd(nsstr);
}
}
@@ -96,13 +98,11 @@ void V3Number::v3errorEndFatal(std::ostringstream& str) const {
V3Number::V3Number(VerilogStringLiteral, AstNode* nodep, const string& str) {
// Create a number using a verilog string as the value, thus 8 bits per character.
// cppcheck bug - doesn't see init() resets these
// cppcheck: Member variable 'm_sized/m_width' is not initialized in the constructor
init(nodep, str.length() * 8);
m_fromString = true;
init(nodep, std::max<int>(str.length() * 8, 1));
m_data.m_fromString = true;
for (unsigned pos = 0; pos < str.length(); ++pos) {
const int topos = str.length() - 1 - pos;
ValueAndX& v = m_value[topos / 4];
ValueAndX& v = m_data.num()[topos / 4];
for (int bit = 0; bit < 8; ++bit) {
if (str[pos] & (1UL << bit)) { v.m_value |= (1UL << (bit + (topos % 4) * 8)); }
}
@@ -112,7 +112,7 @@ V3Number::V3Number(VerilogStringLiteral, AstNode* nodep, const string& str) {
V3Number::V3Number(AstNode* nodep, const AstNodeDType* nodedtypep) {
if (nodedtypep->isString()) {
init(nodep, 0);
init(nodep);
setString("");
} else if (nodedtypep->isDouble()) {
init(nodep, 64);
@@ -124,6 +124,7 @@ V3Number::V3Number(AstNode* nodep, const AstNodeDType* nodedtypep) {
void V3Number::V3NumberCreate(AstNode* nodep, const char* sourcep, FileLine* fl) {
init(nodep, 0);
m_data.setLogic();
m_fileline = fl;
const char* value_startp = sourcep;
for (const char* cp = sourcep; *cp; cp++) {
@@ -172,35 +173,35 @@ void V3Number::V3NumberCreate(AstNode* nodep, const char* sourcep, FileLine* fl)
base = 'd';
}
for (int i = 0; i < words(); ++i) m_value[i] = {0, 0};
for (int i = 0; i < words(); ++i) m_data.num()[i] = {0, 0};
// Special SystemVerilog unsized constructs
if (base == '0') {
width(1, false); // So we extend it
setBit(0, 0);
width(1, false); // So we extend it
m_autoExtend = true;
m_data.m_autoExtend = true;
} else if (base == '1') {
width(1, false); // So we extend it
setBit(0, 1);
width(1, false); // So we extend it
m_autoExtend = true;
m_data.m_autoExtend = true;
} else if (tolower(base) == 'z') {
width(1, false); // So we extend it
setBit(0, 'z');
width(1, false); // So we extend it
m_autoExtend = true;
m_data.m_autoExtend = true;
} else if (tolower(base) == 'x') {
setBit(0, 'x');
width(1, false); // So we extend it
m_autoExtend = true;
setBit(0, 'x');
m_data.m_autoExtend = true;
}
// Otherwise...
else if (!m_sized) {
else if (!sized()) {
width(32, false); // Says IEEE 1800-2012 5.7.1
if (unbased) isSigned(true); // Also says the spec.
}
// Ignore leading blanks
while (*value_startp == '_' || isspace(*value_startp)) value_startp++;
if (!*value_startp && !m_autoExtend) {
if (!*value_startp && !m_data.m_autoExtend) {
v3error("Number is missing value digits: " << sourcep);
}
@@ -229,7 +230,7 @@ void V3Number::V3NumberCreate(AstNode* nodep, const char* sourcep, FileLine* fl)
if (olen <= 7) { // 10000000 fits in 32 bits, so ok
// Constants are common, so for speed avoid wide math until we need it
val = val * 10 + (*cp - '0');
m_value[0].m_value = val;
m_data.num()[0].m_value = val;
} else { // Wide; all previous digits are already in m_value[0]
// this = (this * 10)/*product*/ + (*cp-'0')/*addend*/
// Assumed rare; lots of optimizations are possible here
@@ -237,12 +238,12 @@ void V3Number::V3NumberCreate(AstNode* nodep, const char* sourcep, FileLine* fl)
const V3Number ten(this, width() + 4, 10);
const V3Number addend(this, width(), (*cp - '0'));
product.opMul(*this, ten);
this->opAdd(product, addend);
opAdd(product, addend);
if (product.bitsValue(width(), 4)) { // Overflowed
static int warned = 0;
v3error("Too many digits for "
<< width() << " bit number: " << sourcep << '\n'
<< ((!m_sized && !warned++) ? (
<< ((!sized() && !warned++) ? (
V3Error::warnMore() + "... As that number was unsized"
+ " ('d...) it is limited to 32 bits (IEEE 1800-2017 "
"5.7.1)\n"
@@ -387,72 +388,72 @@ int V3Number::log2b(uint32_t num) {
// Setters
V3Number& V3Number::setZero() {
for (int i = 0; i < words(); i++) m_value[i] = {0, 0};
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, 0};
return *this;
}
V3Number& V3Number::setQuad(uint64_t value) {
for (int i = 0; i < words(); i++) m_value[i] = {0, 0};
m_value[0].m_value = value & 0xffffffffULL;
if (width() > 32) m_value[1].m_value = (value >> 32ULL) & 0xffffffffULL;
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, 0};
m_data.num()[0].m_value = value & 0xffffffffULL;
if (width() > 32) m_data.num()[1].m_value = (value >> 32ULL) & 0xffffffffULL;
opCleanThis();
return *this;
}
V3Number& V3Number::setLong(uint32_t value) {
for (int i = 0; i < words(); i++) m_value[i] = {0, 0};
m_value[0].m_value = value;
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, 0};
m_data.num()[0].m_value = value;
opCleanThis();
return *this;
}
V3Number& V3Number::setLongS(int32_t value) {
for (int i = 0; i < words(); i++) m_value[i] = {0, 0};
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, 0};
union {
uint32_t u;
int32_t s;
} u;
u.s = value;
if (u.s) {}
m_value[0].m_value = u.u;
m_data.num()[0].m_value = u.u;
opCleanThis();
return *this;
}
V3Number& V3Number::setDouble(double value) {
if (VL_UNCOVERABLE(width() != 64)) v3fatalSrc("Real operation on wrong sized number");
m_double = true;
m_data.setDouble();
union {
double d;
uint32_t u[2];
} u;
u.d = value;
if (u.d != 0.0) {}
for (int i = 2; i < words(); i++) m_value[i] = {0, 0};
m_value[0].m_value = u.u[0];
m_value[1].m_value = u.u[1];
for (int i = 2; i < words(); i++) m_data.num()[i] = {0, 0};
m_data.num()[0].m_value = u.u[0];
m_data.num()[1].m_value = u.u[1];
return *this;
}
V3Number& V3Number::setSingleBits(char value) {
for (int i = 1 /*upper*/; i < words(); i++) m_value[i] = {0, 0};
m_value[0] = {(value == '1' || value == 'x' || value == 1 || value == 3),
(value == 'z' || value == 'x' || value == 2 || value == 3)};
for (int i = 1 /*upper*/; i < words(); i++) m_data.num()[i] = {0, 0};
m_data.num()[0] = {(value == '1' || value == 'x' || value == 1 || value == 3),
(value == 'z' || value == 'x' || value == 2 || value == 3)};
return *this;
}
V3Number& V3Number::setAllBits0() {
for (int i = 0; i < words(); i++) m_value[i] = {0, 0};
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, 0};
return *this;
}
V3Number& V3Number::setAllBits1() {
for (int i = 0; i < words(); i++) m_value[i] = {~0U, 0};
for (int i = 0; i < words(); i++) m_data.num()[i] = {~0U, 0};
opCleanThis();
return *this;
}
V3Number& V3Number::setAllBitsX() {
// Use setAllBitsXRemoved if calling this based on a non-X/Z input value such as divide by zero
for (int i = 0; i < words(); i++) m_value[i] = {~0U, ~0U};
for (int i = 0; i < words(); i++) m_data.num()[i] = {~0U, ~0U};
opCleanThis();
return *this;
}
V3Number& V3Number::setAllBitsZ() {
for (int i = 0; i < words(); i++) m_value[i] = {0, ~0U};
for (int i = 0; i < words(); i++) m_data.num()[i] = {0, ~0U};
opCleanThis();
return *this;
}
@@ -493,7 +494,7 @@ string V3Number::ascii(bool prefixed, bool cleanVerilog) const {
} else if (isString()) {
return '"' + toString() + '"';
} else {
if (VL_UNCOVERABLE((m_value[words() - 1].m_value | m_value[words() - 1].m_valueX)
if (VL_UNCOVERABLE((m_data.num()[words() - 1].m_value | m_data.num()[words() - 1].m_valueX)
& ~hiWordMask())) {
out << "%E-hidden-bits"; // LCOV_EXCL_LINE
}
@@ -688,7 +689,7 @@ string V3Number::displayed(FileLine* fl, const string& vformat) const {
}
case 's': {
// Spec says always drop leading zeros, this isn't quite right, we space pad.
int bit = this->width() - 1;
int bit = width() - 1;
bool start = true;
while ((bit % 8) != 7) bit++;
for (; bit >= 0; bit -= 8) {
@@ -709,7 +710,7 @@ string V3Number::displayed(FileLine* fl, const string& vformat) const {
case 'd': { // Unsigned decimal
const bool issigned = (code == '~');
if (fmtsize == "") {
const double mantissabits = this->width() - (issigned ? 1 : 0);
const double mantissabits = width() - (issigned ? 1 : 0);
// To get the number of digits required, we want to compute
// log10(2**mantissabits) and round it up. To be able to handle
// a very wide mantissa, we use log2(2**mantissabits)/log2(10),
@@ -765,7 +766,7 @@ string V3Number::displayed(FileLine* fl, const string& vformat) const {
// 'p' // Packed - converted to another code by V3Width
case 'u': { // Packed 2-state
for (int i = 0; i < words(); i++) {
const uint32_t v = m_value[i].m_value;
const uint32_t v = m_data.num()[i].m_value;
str += static_cast<char>((v >> 0) & 0xff);
str += static_cast<char>((v >> 8) & 0xff);
str += static_cast<char>((v >> 16) & 0xff);
@@ -775,7 +776,7 @@ string V3Number::displayed(FileLine* fl, const string& vformat) const {
}
case 'z': { // Packed 4-state
for (int i = 0; i < words(); i++) {
const ValueAndX v = m_value[i];
const ValueAndX v = m_data.num()[i];
str += static_cast<char>((v.m_value >> 0) & 0xff);
str += static_cast<char>((v.m_value >> 8) & 0xff);
str += static_cast<char>((v.m_value >> 16) & 0xff);
@@ -872,12 +873,12 @@ uint32_t V3Number::toUInt() const {
UASSERT(!isFourState(), "toUInt with 4-state " << *this);
// We allow wide numbers that represent values <= 32 bits
for (int i = 1; i < words(); ++i) {
if (m_value[i].m_value) {
if (m_data.num()[i].m_value) {
v3error("Value too wide for 32-bits expected in this context " << *this);
break;
}
}
return m_value[0].m_value;
return m_data.num()[0].m_value;
}
double V3Number::toDouble() const {
@@ -888,8 +889,8 @@ double V3Number::toDouble() const {
double d;
uint32_t u[2];
} u;
u.u[0] = m_value[0].m_value;
u.u[1] = m_value[1].m_value;
u.u[0] = m_data.num()[0].m_value;
u.u[1] = m_data.num()[1].m_value;
return u.d;
}
@@ -911,14 +912,14 @@ uint64_t V3Number::toUQuad() const {
// We allow wide numbers that represent values <= 64 bits
if (isDouble()) return static_cast<uint64_t>(toDouble());
for (int i = 2; i < words(); ++i) {
if (m_value[i].m_value) {
if (m_data.num()[i].m_value) {
v3error("Value too wide for 64-bits expected in this context " << *this);
break;
}
}
if (width() <= 32) return (static_cast<uint64_t>(toUInt()));
return ((static_cast<uint64_t>(m_value[1].m_value) << 32ULL)
| (static_cast<uint64_t>(m_value[0].m_value)));
return ((static_cast<uint64_t>(m_data.num()[1].m_value) << 32ULL)
| (static_cast<uint64_t>(m_data.num()[0].m_value)));
}
int64_t V3Number::toSQuad() const {
@@ -932,8 +933,8 @@ int64_t V3Number::toSQuad() const {
string V3Number::toString() const {
UASSERT(!isFourState(), "toString with 4-state " << *this);
// Spec says always drop leading zeros, this isn't quite right, we space pad.
if (isString()) return m_stringVal;
int bit = this->width() - 1;
if (isString()) return m_data.str();
int bit = width() - 1;
bool start = true;
while ((bit % 8) != 7) bit++;
string str;
@@ -948,14 +949,18 @@ string V3Number::toString() const {
}
V3Hash V3Number::toHash() const {
V3Hash hash(m_width);
for (int i = 0; i < words(); ++i) { hash += m_value[i].m_value; }
V3Hash hash{width()};
if (isString()) {
hash += V3Hash{m_data.str()};
} else {
for (int i = 0; i < words(); ++i) { hash += m_data.num()[i].m_value; }
}
return hash;
}
uint32_t V3Number::edataWord(int eword) const {
UASSERT(!isFourState(), "edataWord with 4-state " << *this);
return m_value[eword].m_value;
return m_data.num()[eword].m_value;
}
uint8_t V3Number::dataByte(int byte) const {
@@ -963,30 +968,34 @@ uint8_t V3Number::dataByte(int byte) const {
}
bool V3Number::isAllZ() const {
if (isDouble() || isString()) return false;
for (int i = 0; i < width(); i++) {
if (!bitIsZ(i)) return false;
}
return true;
}
bool V3Number::isAllX() const {
if (isDouble() || isString()) return false;
uint32_t mask = hiWordMask();
for (int i = words() - 1; i >= 0; --i) {
const ValueAndX v = m_value[i];
const ValueAndX v = m_data.num()[i];
if ((v.m_value & v.m_valueX) ^ mask) return false;
mask = ~0U;
}
return true;
}
bool V3Number::isEqZero() const {
if (isString()) return m_data.str().empty();
for (int i = 0; i < words(); i++) {
const ValueAndX v = m_value[i];
const ValueAndX v = m_data.num()[i];
if (v.m_value || v.m_valueX) return false;
}
return true;
}
bool V3Number::isNeqZero() const {
if (isString()) return !m_data.str().empty();
for (int i = 0; i < words(); i++) {
const ValueAndX v = m_value[i];
const ValueAndX v = m_data.num()[i];
if (v.m_value & ~v.m_valueX) return true;
}
return false;
@@ -998,9 +1007,9 @@ bool V3Number::isBitsZero(int msb, int lsb) const {
return true;
}
bool V3Number::isEqOne() const {
if (m_value[0].m_value != 1 || m_value[0].m_valueX) return false;
if (m_data.num()[0].m_value != 1 || m_data.num()[0].m_valueX) return false;
for (int i = 1; i < words(); i++) {
const ValueAndX v = m_value[i];
const ValueAndX v = m_data.num()[i];
if (v.m_value || v.m_valueX) return false;
}
return true;
@@ -1015,7 +1024,7 @@ bool V3Number::isEqAllOnes(int optwidth) const {
bool V3Number::isFourState() const {
if (isDouble() || isString()) return false;
for (int i = 0; i < words(); ++i) {
if (m_value[i].m_valueX) return true;
if (m_data.num()[i].m_valueX) return true;
}
return false;
}
@@ -1036,10 +1045,10 @@ bool V3Number::isAnyZ() const {
}
bool V3Number::isLtXZ(const V3Number& rhs) const {
// Include X/Z in comparisons for sort ordering
for (int bit = 0; bit < std::max(this->width(), rhs.width()); bit++) {
if (this->bitIs1(bit) && rhs.bitIs0(bit)) return true;
if (rhs.bitIs1(bit) && this->bitIs0(bit)) return false;
if (this->bitIsXZ(bit)) return true;
for (int bit = 0; bit < std::max(width(), rhs.width()); bit++) {
if (bitIs1(bit) && rhs.bitIs0(bit)) return true;
if (rhs.bitIs1(bit) && bitIs0(bit)) return false;
if (bitIsXZ(bit)) return true;
if (rhs.bitIsXZ(bit)) return false;
}
return false;
@@ -1070,7 +1079,7 @@ int V3Number::widthMin() const {
uint32_t V3Number::countBits(const V3Number& ctrl) const {
int n = 0;
for (int bit = 0; bit < this->width(); ++bit) {
for (int bit = 0; bit < width(); ++bit) {
switch (ctrl.bitIs(0)) {
case '0':
if (bitIs0(bit)) ++n;
@@ -1101,14 +1110,14 @@ uint32_t V3Number::countBits(const V3Number& ctrl1, const V3Number& ctrl2,
uint32_t V3Number::countOnes() const {
int n = 0;
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (bitIs1(bit)) n++;
}
return n;
}
uint32_t V3Number::mostSetBitP1() const {
for (int bit = this->width() - 1; bit >= 0; bit--) {
for (int bit = width() - 1; bit >= 0; bit--) {
if (bitIs1(bit)) return bit + 1;
}
return 0;
@@ -1120,7 +1129,7 @@ V3Number& V3Number::opBitsNonX(const V3Number& lhs) { // 0/1->1, X/Z->0
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs0(bit) || lhs.bitIs1(bit)) setBit(bit, 1);
}
return *this;
@@ -1130,7 +1139,7 @@ V3Number& V3Number::opBitsOne(const V3Number& lhs) { // 1->1, 0/X/Z->0
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs1(bit)) setBit(bit, 1);
}
return *this;
@@ -1140,7 +1149,7 @@ V3Number& V3Number::opBitsXZ(const V3Number& lhs) { // 0/1->1, X/Z->0
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIsXZ(bit)) setBit(bit, 1);
}
return *this;
@@ -1150,7 +1159,7 @@ V3Number& V3Number::opBitsZ(const V3Number& lhs) { // 0/1->1, X/Z->0
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIsZ(bit)) setBit(bit, 1);
}
return *this;
@@ -1160,7 +1169,7 @@ V3Number& V3Number::opBitsNonZ(const V3Number& lhs) { // 0/1->1, X/Z->0
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (!lhs.bitIsZ(bit)) setBit(bit, 1);
}
return *this;
@@ -1226,7 +1235,7 @@ V3Number& V3Number::opCountBits(const V3Number& expr, const V3Number& ctrl1, con
NUM_ASSERT_OP_ARGS4(expr, ctrl1, ctrl2, ctrl3);
NUM_ASSERT_LOGIC_ARGS4(expr, ctrl1, ctrl2, ctrl3);
setZero();
m_value[0].m_value = expr.countBits(ctrl1, ctrl2, ctrl3);
m_data.num()[0].m_value = expr.countBits(ctrl1, ctrl2, ctrl3);
opCleanThis();
return *this;
}
@@ -1235,7 +1244,7 @@ V3Number& V3Number::opCountOnes(const V3Number& lhs) {
NUM_ASSERT_LOGIC_ARGS1(lhs);
if (lhs.isFourState()) return setAllBitsX();
setZero();
m_value[0].m_value = lhs.countOnes();
m_data.num()[0].m_value = lhs.countOnes();
opCleanThis();
return *this;
}
@@ -1292,7 +1301,7 @@ V3Number& V3Number::opNot(const V3Number& lhs) {
NUM_ASSERT_LOGIC_ARGS1(lhs);
// op i, L(lhs) bit return
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs0(bit)) {
setBit(bit, 1);
} else if (lhs.bitIsXZ(bit)) {
@@ -1307,7 +1316,7 @@ V3Number& V3Number::opAnd(const V3Number& lhs, const V3Number& rhs) {
NUM_ASSERT_LOGIC_ARGS2(lhs, rhs);
// i op j, max(L(lhs),L(rhs)) bit return, careful need to X/Z extend.
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs1(bit) && rhs.bitIs1(bit)) {
setBit(bit, 1);
} else if (lhs.bitIs0(bit) || rhs.bitIs0(bit)) { // 0
@@ -1323,7 +1332,7 @@ V3Number& V3Number::opOr(const V3Number& lhs, const V3Number& rhs) {
NUM_ASSERT_LOGIC_ARGS2(lhs, rhs);
// i op j, max(L(lhs),L(rhs)) bit return, careful need to X/Z extend.
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs1(bit) || rhs.bitIs1(bit)) {
setBit(bit, 1);
} else if (lhs.bitIs0(bit) && rhs.bitIs0(bit)) {
@@ -1340,7 +1349,7 @@ V3Number& V3Number::opXor(const V3Number& lhs, const V3Number& rhs) {
NUM_ASSERT_OP_ARGS2(lhs, rhs);
NUM_ASSERT_LOGIC_ARGS2(lhs, rhs);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (lhs.bitIs1(bit) && rhs.bitIs0(bit)) {
setBit(bit, 1);
} else if (lhs.bitIs0(bit) && rhs.bitIs1(bit)) {
@@ -1619,9 +1628,9 @@ bool V3Number::isCaseEq(const V3Number& rhs) const {
// i op j, 1 bit return, max(L(lhs),L(rhs)) calculation, careful need to X/Z extend.
if (isString()) return toString() == rhs.toString();
if (isDouble()) return toDouble() == rhs.toDouble();
if (this->width() != rhs.width()) return false;
if (width() != rhs.width()) return false;
for (int i = 0; i < words(); ++i) {
if (!(m_value[i] == rhs.m_value[i])) return false;
if (!(m_data.num()[i] == rhs.m_data.num()[i])) return false;
}
return true;
}
@@ -1754,7 +1763,7 @@ V3Number& V3Number::opShiftR(const V3Number& lhs, const V3Number& rhs) {
}
const uint32_t rhsval = rhs.toUInt();
if (rhsval < static_cast<uint32_t>(lhs.width())) {
for (int bit = 0; bit < this->width(); bit++) setBit(bit, lhs.bitIs(bit + rhsval));
for (int bit = 0; bit < width(); bit++) setBit(bit, lhs.bitIs(bit + rhsval));
}
return *this;
}
@@ -1768,7 +1777,7 @@ V3Number& V3Number::opShiftRS(const V3Number& lhs, const V3Number& rhs, uint32_t
if (rhs.isFourState()) return setAllBitsX();
setZero();
for (int bit = 32; bit < rhs.width(); bit++) {
for (int sbit = 0; sbit < this->width(); sbit++) {
for (int sbit = 0; sbit < width(); sbit++) {
setBit(sbit, lhs.bitIs(lbits - 1)); // 0/1/X/Z
}
if (rhs.bitIs1(lbits - 1)) setAllBits1(); // -1 else 0
@@ -1776,11 +1785,11 @@ V3Number& V3Number::opShiftRS(const V3Number& lhs, const V3Number& rhs, uint32_t
}
const uint32_t rhsval = rhs.toUInt();
if (rhsval < static_cast<uint32_t>(lhs.width())) {
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
setBit(bit, lhs.bitIsExtend(bit + rhsval, lbits));
}
} else {
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
setBit(bit, lhs.bitIs(lbits - 1)); // 0/1/X/Z
}
}
@@ -1797,7 +1806,7 @@ V3Number& V3Number::opShiftL(const V3Number& lhs, const V3Number& rhs) {
if (rhs.bitIs1(bit)) return *this; // shift of over 2^32 must be zero
}
const uint32_t rhsval = rhs.toUInt();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (bit >= static_cast<int>(rhsval)) setBit(bit, lhs.bitIs(bit - rhsval));
}
return *this;
@@ -1825,7 +1834,7 @@ V3Number& V3Number::opAdd(const V3Number& lhs, const V3Number& rhs) {
setZero();
// Addem
int carry = 0;
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
const int sum = ((lhs.bitIs1(bit) ? 1 : 0) + (rhs.bitIs1(bit) ? 1 : 0) + carry);
if (sum & 1) setBit(bit, 1);
carry = (sum >= 2);
@@ -1852,15 +1861,15 @@ V3Number& V3Number::opMul(const V3Number& lhs, const V3Number& rhs) {
opCleanThis(); // Mult produces extra bits in result
} else {
for (int lword = 0; lword < lhs.words(); lword++) {
const uint64_t lwordval = static_cast<uint64_t>(lhs.m_value[lword].m_value);
const uint64_t lwordval = static_cast<uint64_t>(lhs.m_data.num()[lword].m_value);
if (lwordval == 0) continue;
for (int rword = 0; rword < rhs.words(); rword++) {
const uint64_t rwordval = static_cast<uint64_t>(rhs.m_value[rword].m_value);
const uint64_t rwordval = static_cast<uint64_t>(rhs.m_data.num()[rword].m_value);
if (rwordval == 0) continue;
uint64_t mul = lwordval * rwordval;
for (int qword = lword + rword; qword < this->words(); qword++) {
mul += static_cast<uint64_t>(m_value[qword].m_value);
m_value[qword].m_value = (mul & 0xffffffffULL);
for (int qword = lword + rword; qword < words(); qword++) {
mul += static_cast<uint64_t>(m_data.num()[qword].m_value);
m_data.num()[qword].m_value = (mul & 0xffffffffULL);
mul = (mul >> 32ULL) & 0xffffffffULL;
if (mul == 0) break;
}
@@ -1977,17 +1986,18 @@ V3Number& V3Number::opModDivGuts(const V3Number& lhs, const V3Number& rhs, bool
if (vw == 1) { // Single divisor word breaks rest of algorithm
uint64_t k = 0;
for (int j = uw - 1; j >= 0; j--) {
const uint64_t unw64 = ((k << 32ULL) + static_cast<uint64_t>(lhs.m_value[j].m_value));
m_value[j].m_value = unw64 / static_cast<uint64_t>(rhs.m_value[0].m_value);
const uint64_t unw64
= ((k << 32ULL) + static_cast<uint64_t>(lhs.m_data.num()[j].m_value));
m_data.num()[j].m_value = unw64 / static_cast<uint64_t>(rhs.m_data.num()[0].m_value);
k = unw64
- (static_cast<uint64_t>(m_value[j].m_value)
* static_cast<uint64_t>(rhs.m_value[0].m_value));
- (static_cast<uint64_t>(m_data.num()[j].m_value)
* static_cast<uint64_t>(rhs.m_data.num()[0].m_value));
}
UINFO(9, " opmoddiv-1w " << lhs << " " << rhs << " q=" << *this << " rem=0x" << std::hex
<< k << std::dec << endl);
if (is_modulus) {
setZero();
m_value[0].m_value = k;
m_data.num()[0].m_value = k;
}
opCleanThis();
return *this;
@@ -1998,7 +2008,7 @@ V3Number& V3Number::opModDivGuts(const V3Number& lhs, const V3Number& rhs, bool
uint32_t vn[VL_MULS_MAX_WORDS + 1]; // v normalized
// Zero for ease of debugging and to save having to zero for shifts
for (int i = 0; i < words; i++) { m_value[i].m_value = 0; }
for (int i = 0; i < words; i++) { m_data.num()[i].m_value = 0; }
for (int i = 0; i < words + 1; i++) { un[i] = vn[i] = 0; } // +1 as vn may get extra word
// Algorithm requires divisor MSB to be set
@@ -2006,22 +2016,22 @@ V3Number& V3Number::opModDivGuts(const V3Number& lhs, const V3Number& rhs, bool
const int s = 31 - ((vmsbp1 - 1) & 31); // shift amount (0...31)
const uint32_t shift_mask = s ? 0xffffffff : 0; // otherwise >> 32 won't mask the value
for (int i = vw - 1; i > 0; i--) {
vn[i] = (rhs.m_value[i].m_value << s)
| (shift_mask & (rhs.m_value[i - 1].m_value >> (32 - s)));
vn[i] = (rhs.m_data.num()[i].m_value << s)
| (shift_mask & (rhs.m_data.num()[i - 1].m_value >> (32 - s)));
}
vn[0] = rhs.m_value[0].m_value << s;
vn[0] = rhs.m_data.num()[0].m_value << s;
// Copy and shift dividend by same amount; may set new upper word
if (s) {
un[uw] = lhs.m_value[uw - 1].m_value >> (32 - s);
un[uw] = lhs.m_data.num()[uw - 1].m_value >> (32 - s);
} else {
un[uw] = 0;
}
for (int i = uw - 1; i > 0; i--) {
un[i] = (lhs.m_value[i].m_value << s)
| (shift_mask & (lhs.m_value[i - 1].m_value >> (32 - s)));
un[i] = (lhs.m_data.num()[i].m_value << s)
| (shift_mask & (lhs.m_data.num()[i - 1].m_value >> (32 - s)));
}
un[0] = lhs.m_value[0].m_value << s;
un[0] = lhs.m_data.num()[0].m_value << s;
// printf(" un="); for (int i=5; i>=0; i--) printf(" %08x",un[i]); printf("\n");
// printf(" vn="); for (int i=5; i>=0; i--) printf(" %08x",vn[i]); printf("\n");
@@ -2052,11 +2062,11 @@ V3Number& V3Number::opModDivGuts(const V3Number& lhs, const V3Number& rhs, bool
}
t = un[j + vw] - k;
un[j + vw] = t;
this->m_value[j].m_value = qhat; // Save quotient digit
m_data.num()[j].m_value = qhat; // Save quotient digit
if (t < 0) {
// Over subtracted; correct by adding back
this->m_value[j].m_value--;
m_data.num()[j].m_value--;
k = 0;
for (int i = 0; i < vw; i++) {
t = static_cast<uint64_t>(un[i + j]) + static_cast<uint64_t>(vn[i]) + k;
@@ -2074,9 +2084,9 @@ V3Number& V3Number::opModDivGuts(const V3Number& lhs, const V3Number& rhs, bool
if (is_modulus) { // modulus
// Need to reverse normalization on copy to output
for (int i = 0; i < vw; i++) {
m_value[i].m_value = (un[i] >> s) | (shift_mask & (un[i + 1] << (32 - s)));
m_data.num()[i].m_value = (un[i] >> s) | (shift_mask & (un[i + 1] << (32 - s)));
}
for (int i = vw; i < words; i++) m_value[i].m_value = 0;
for (int i = vw; i < words; i++) m_data.num()[i].m_value = 0;
opCleanThis();
UINFO(9, " opmoddiv-mod " << lhs << " " << rhs << " now=" << *this << endl);
return *this;
@@ -2110,7 +2120,7 @@ V3Number& V3Number::opPow(const V3Number& lhs, const V3Number& rhs, bool lsign,
}
if (lhs.isEqZero()) return setZero();
setZero();
m_value[0].m_value = 1;
m_data.num()[0].m_value = 1;
V3Number power(&lhs, width());
power.opAssign(lhs);
for (int bit = 0; bit < rhs.width(); bit++) {
@@ -2122,7 +2132,7 @@ V3Number& V3Number::opPow(const V3Number& lhs, const V3Number& rhs, bool lsign,
if (rhs.bitIs1(bit)) { // out *= power
V3Number lastOut(&lhs, width());
lastOut.opAssign(*this);
this->opMul(lastOut, power);
opMul(lastOut, power);
// UINFO(0, "pow "<<lhs<<" "<<rhs<<" b"<<bit<<" pow="<<power<<" now="<<*this<<endl);
}
}
@@ -2142,7 +2152,7 @@ V3Number& V3Number::opBufIf1(const V3Number& ens, const V3Number& if1s) {
NUM_ASSERT_OP_ARGS2(ens, if1s);
NUM_ASSERT_LOGIC_ARGS2(ens, if1s);
setZero();
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (ens.bitIs1(bit)) {
setBit(bit, if1s.bitIs(bit));
} else {
@@ -2159,9 +2169,16 @@ V3Number& V3Number::opAssignNonXZ(const V3Number& lhs, bool ignoreXZ) {
// to itself; V3Simulate does this when hits "foo=foo;"
// So no: NUM_ASSERT_OP_ARGS1(lhs);
if (this != &lhs) {
setZero();
if (isString()) {
m_stringVal = lhs.m_stringVal;
if (VL_UNLIKELY(!lhs.isString())) {
// Non-compatible types, erase value.
m_data.str() = "";
} else {
m_data.str() = lhs.m_data.str();
}
} else if (VL_UNLIKELY(lhs.isString())) {
// Non-compatible types, erase value.
setZero();
} else {
// Also handles double as is just bits
for (int bit = 0; bit < this->width(); bit++) {
@@ -2198,14 +2215,14 @@ V3Number& V3Number::opClean(const V3Number& lhs, uint32_t bits) { return opSel(l
void V3Number::opCleanThis(bool warnOnTruncation) {
// Clean MSB of number
NUM_ASSERT_LOGIC_ARGS1(*this);
const ValueAndX v = m_value[words() - 1];
const ValueAndX v = m_data.num()[words() - 1];
const uint32_t newValueMsb = v.m_value & hiWordMask();
const uint32_t newValueXMsb = v.m_valueX & hiWordMask();
if (warnOnTruncation && (newValueMsb != v.m_value || newValueXMsb != v.m_valueX)) {
// Displaying in decimal avoids hiWordMask truncation
v3warn(WIDTH, "Value too large for " << width() << " bit number: " << displayed("%d"));
}
m_value[words() - 1] = {newValueMsb, newValueXMsb};
m_data.num()[words() - 1] = {newValueMsb, newValueXMsb};
}
V3Number& V3Number::opSel(const V3Number& lhs, const V3Number& msb, const V3Number& lsb) {
@@ -2221,7 +2238,7 @@ V3Number& V3Number::opSel(const V3Number& lhs, uint32_t msbval, uint32_t lsbval)
NUM_ASSERT_LOGIC_ARGS1(lhs);
setZero();
int ibit = lsbval;
for (int bit = 0; bit < this->width(); bit++) {
for (int bit = 0; bit < width(); bit++) {
if (ibit >= 0 && ibit < lhs.width() && ibit <= static_cast<int>(msbval)) {
setBit(bit, lhs.bitIs(ibit));
} else {
@@ -2317,21 +2334,17 @@ V3Number& V3Number::opRealToBits(const V3Number& lhs) {
NUM_ASSERT_OP_ARGS1(lhs);
NUM_ASSERT_DOUBLE_ARGS1(lhs);
// Conveniently our internal format is identical so we can copy bits...
if (lhs.width() != 64 || this->width() != 64) {
v3fatalSrc("Real operation on wrong sized number");
}
if (lhs.width() != 64 || width() != 64) v3fatalSrc("Real operation on wrong sized number");
m_data.setLogic();
opAssign(lhs);
m_double = false;
return *this;
}
V3Number& V3Number::opBitsToRealD(const V3Number& lhs) {
NUM_ASSERT_OP_ARGS1(lhs);
// Conveniently our internal format is identical so we can copy bits...
if (lhs.width() != 64 || this->width() != 64) {
v3fatalSrc("Real operation on wrong sized number");
}
if (lhs.width() != 64 || width() != 64) v3fatalSrc("Real operation on wrong sized number");
m_data.setDouble();
opAssign(lhs);
m_double = true;
return *this;
}
V3Number& V3Number::opNegateD(const V3Number& lhs) {