202108171803 long int module

hdl4secell/include/hdl4secell.h

@@ -106,6 +106,10 @@ enum varupdatefunc {
 	VUF_UPDATING,
 };
 
+#ifndef VARBITS
+#define VARBITS   256
+#endif
+
 typedef struct sModuleVariable {
     struct sModuleVariable * pNext, *pLast;
     int index;
@@ -121,7 +125,7 @@ typedef struct sModuleVariable {
 	PointerArray functions; /* 依赖它的函数 */
 	PointerArray variables; /* 依赖它的变量 */
 	PointerArray dependvar;
-	long long updatedisset;
+	int updatedisset;
 	int genfuncindex;
 #if USEBIGINT	
     IBigNumber** data;
@@ -134,7 +138,9 @@ typedef struct sModuleVariable {
 		unsigned int		uint32;
 		long long			int64;
 		unsigned long long	uint64;
-		unsigned long long  data;
+		//unsigned long long  data;
+		unsigned int data[VARBITS / 32];
+		unsigned long long data64[VARBITS / 64];
 	};
 	union {
 		short				int16_reg;
@@ -143,7 +149,9 @@ typedef struct sModuleVariable {
 		unsigned int		uint32_reg;
 		long long			int64_reg;
 		unsigned long long	uint64_reg;
-		unsigned long long  data_reg;
+		//unsigned long long  data_reg;
+		unsigned int data_reg[VARBITS / 32];
+		unsigned long long data64_reg[VARBITS / 64];
 	};
 
 #endif
@@ -151,6 +159,10 @@ typedef struct sModuleVariable {
 	long updatefunc;
 }ModuleVariable;
 
+#define VAR_COPYDATA(dst, src)  memcpy(dst, src, VARBITS/8)
+#define VAR_NEQ(dst, src) memcmp(dst, src, VARBITS/8)
+
+
 typedef struct sModuleFunction {
 	ModuleVariable* var; /* getvalue对应的变量 */
 	void* pobj;
@@ -221,8 +233,8 @@ ModuleVariable * varUOr(ModuleVariable * a);
 ModuleVariable * varUNOr(ModuleVariable * a);
 ModuleVariable * varUXor(ModuleVariable * a);
 ModuleVariable * varUNXor(ModuleVariable * a);
-ModuleVariable * varBit(ModuleVariable * a, int bit);
-ModuleVariable * varBits(ModuleVariable * a, int lsb, int width);
+ModuleVariable * varBit(ModuleVariable * a, unsigned int bit);
+ModuleVariable * varBits(ModuleVariable * a, unsigned int lsb, unsigned int width);
 ModuleVariable * varSelect(ModuleVariable * a, ModuleVariable * b, ModuleVariable * c);
 
 int varAssign(ModuleVariable * a, ModuleVariable * b);
@@ -231,8 +243,8 @@ int varCopy(ModuleVariable * a, ModuleVariable * b);
 int varIsZero(ModuleVariable * a);
 int varIsNotZero(ModuleVariable * a);
 
-int varSetBits32(ModuleVariable* a, int index, unsigned int v);
-unsigned int varGetBits32(ModuleVariable* a, int index);
+int varSetBits32(ModuleVariable* a, unsigned int index, unsigned int v);
+unsigned int varGetBits32(ModuleVariable* a, unsigned int index);
 
 int varAssign_S32(ModuleVariable * a, int v);
 int varAssign_U32(ModuleVariable * a, unsigned int v);
@@ -324,7 +336,6 @@ IHDL4SEModuleVar* hdl4seCreateModule3(HOBJECT parent, const char* clsid, const c
 #define IDLIST enum M_ID(id_list) {
 #define END_IDLIST };
 
-
 #define MODULE_DATA_TYPE M_ID(sHDL4SE_data)
 #define MODULE_DATA_STRUCT M_ID(_sHDL4SE_data)
 

hdl4secell/src/hdl4se_fifo.c

@@ -255,7 +255,7 @@ DEFINE_FUNC(hdl4se_fifo_ClkTick, "reset, nwReset, readpos, writepos, wRead, wRea
 			vput(outputcount, vget(outputcount) + 1);
 		}
 		if (vget(wWrite) && vget(wWriteEn)) {
-			int j;
+			unsigned int j;
 			var writedata = Var(bWriteData);
 			for (j = 0; j < pobj->wordsize; j++) {
 				pobj->fifo_data[writeaddr * pobj->wordsize + j] = varGetBits32(writedata, j);

hdl4secell/src/hdl4se_module.c

@@ -268,9 +268,9 @@ int hdl4se_module_Assign(sGeneralModule* pobj, int dstind, int srcind)
 	varsrc = (ModuleVariable*)pobj->variables.array[srcind];
 	UpdateVariable(varsrc);
 	if (vardst->type == VTYPE_REG)
-		vardst->data_reg = varsrc->data;
+		VAR_COPYDATA(vardst->data_reg, varsrc->data);
 	else
-		vardst->data = varsrc->data;
+		VAR_COPYDATA(vardst->data, varsrc->data);
 	return 0;
 }
 
@@ -281,9 +281,9 @@ int hdl4se_module_AssignVar(sGeneralModule* pobj, int dstind, ModuleVariable* va
 		return -1;
 	vardst = (ModuleVariable*)pobj->variables.array[dstind];
 	if (vardst->type == VTYPE_REG)
-		vardst->data_reg = varsrc->data;
+		VAR_COPYDATA(vardst->data_reg, varsrc->data);
 	else
-		vardst->data = varsrc->data;
+		VAR_COPYDATA(vardst->data, varsrc->data);
 	return 0;
 }
 
@@ -340,9 +340,9 @@ void hdl4se_module_SetVarUint32(sGeneralModule* pobj, int varindex, unsigned int
 {
 	ModuleVariable* var = (ModuleVariable*)pobj->variables.array[varindex];
 	if (var->type == VTYPE_REG)
-		var->data_reg = v;
+		var->uint64_reg = v;
 	else
-		var->data = v;
+		var->uint64 = v;
 }
 
 void hdl4se_module_SetVarInt32(sGeneralModule* pobj, int varindex, int v)
@@ -358,9 +358,9 @@ void hdl4se_module_SetVarUint64(sGeneralModule* pobj, int varindex, unsigned lon
 {
 	ModuleVariable* var = (ModuleVariable*)pobj->variables.array[varindex];
 	if (var->type == VTYPE_REG)
-		var->data_reg = v;
+		var->uint64_reg = v;
 	else
-		var->data = v;
+		var->uint64 = v;
 }
 
 void hdl4se_module_SetVarInt64(sGeneralModule* pobj, int varindex, long long v)
@@ -645,10 +645,10 @@ inline int hdl4se_module_GetValue(sGeneralModule* pobj, int index, int width, Mo
 		value->width = var->width;
 	}
 	if (value->type == VTYPE_REG) {
-		value->data_reg = var->data;
+		VAR_COPYDATA(value->data_reg, var->data);
 	}
 	else {
-		value->data = var->data;
+		VAR_COPYDATA(value->data, var->data);
 	}
 #endif
 	return 0;
@@ -687,7 +687,7 @@ int hdl4se_module_ClkTick(sGeneralModule* pobj)
 	return 0;
 }
 
-#define LOCKED_UPDATE 0
+#define LOCKED_UPDATE 1
 
 #if LOCKED_UPDATE
 static int hdl4se_var_set_update(ModuleVariable* v)
@@ -698,9 +698,6 @@ static int hdl4se_var_set_update(ModuleVariable* v)
 	if (LOCKED_SWAP(v->updatedisset, 1)) {
 		return 0;
 	}
-#if DEBUGDEPEND
-	printf("->%s.%s", v->moduledata->name, v->name);
-#endif
 	if (v->genfuncindex >= 0 || v->module != NULL)
 		LOCKED_SWAP(v->updatefunc, VUF_WAITUPDATE);
 	for (i = 0; i < v->variables.itemcount; i++) {
@@ -718,9 +715,6 @@ static int hdl4se_var_set_update(ModuleVariable* v)
 		return 0;
 	}
 	v->updatedisset = 1;
-#if DEBUGDEPEND
-	printf("->%s.%s", v->moduledata->name, v->name);
-#endif
 	if (v->genfuncindex >= 0 || v->module != NULL)
 		v->updatefunc = VUF_WAITUPDATE;
 	for (i = 0; i < v->variables.itemcount; i++) {
@@ -746,8 +740,8 @@ int hdl4se_module_Setup(sGeneralModule* pobj)
 #if USEBIGINT
 			objectCall1(var->data, Clone, var->data_reg);
 #else
-			if (var->data != var->data_reg) {
-				var->data = var->data_reg;
+			if (VAR_NEQ(var->data, var->data_reg)) {
+				VAR_COPYDATA(var->data, var->data_reg);
 #endif
 #if DEPEND_OPT
 			
@@ -845,7 +839,7 @@ int hdl4se_module_path_Traversal(HOBJECT object, hdl4se_module_path_TraversalFun
 		count = hdl4se_module_GetModuleCount(&detector->data);
 		for (i = 0; i < count; i++) {
 			HOBJECT unit;
-			hdl4se_module_GetModule(&detector->data, i, (void**)&unit);
+			hdl4se_module_GetModule(&detector->data, i, (IHDL4SEModuleVar**)&unit);
 			if (hdl4se_module_path_Traversal(unit, func, pname, param) != 0)
 				return -1;
 		}

hdl4secell/src/hdl4se_mux4.c

@@ -49,6 +49,8 @@
 #define hdl4se_mux4_MODULE_VERSION_STRING "0.4.0-20210714.0931 HDL4SE Mux4 cell"
 #define hdl4se_mux4_MODULE_CLSID CLSID_HDL4SE_MUX4
 
+static MODULE_CREATOR_DECLARE(hdl4se_mux4);
+
 IDLIST
 	VID(sel),
 	VID(in0),

hdl4secell/src/hdl4se_variable.c

@@ -77,11 +77,9 @@ ModuleVariable* variableCreate(int type, int width, int portdirect, int isunsign
 	var->genfuncindex = -1;
 	var->module = NULL;
 	var->moduleportindex = 0;
-	var->data = 0;
 	pointerarrayInit(&var->functions);
 	pointerarrayInit(&var->variables);
 	pointerarrayInit(&var->dependvar);
-	var->data_reg = 0;
 #if USEBIGINT
 	{
 		var->data = bigintegerCreate2(width, isunsigned);
@@ -111,7 +109,6 @@ int variableDestroy(ModuleVariable* var)
 #endif
 	if (var->depend_list != NULL)
 		free(var->depend_list);
-	var->data = 0;
 	mt_free(var);
 	return 0;
 }
@@ -329,6 +326,9 @@ long long varGet_S64(ModuleVariable* a);
 unsigned long long varGet_U64(ModuleVariable* a);
 
 #else
+
+
+
 enum TOKEN_STATE {
 	TOKEN_INITIAL,
 	TOKEN_NUM,
@@ -341,10 +341,15 @@ enum TOKEN_STATE {
 
 #define MAXNUMWIDTH (1 << 30)
 
+void varSetWidth(ModuleVariable* v, int width)
+{
+
+}
+
 ModuleVariable* varConstStr(const char* str)
 {
 	ModuleVariable* temp = varTemp(64, 1);
-
+	/* support 64bits only, fixed me*/
 	int width;
 	enum TOKEN_STATE state;
 	const char* strt = str;
@@ -383,19 +388,19 @@ ModuleVariable* varConstStr(const char* str)
 				temp->isunsigned = 0;
 			}
 			else if (ch == 'b' || ch == 'B') {
-				temp->data = 0;
+				temp->uint64 = 0;
 				state = TOKEN_BIN;
 			}
 			else if (ch == 'h' || ch == 'H') {
-				temp->data = 0;
+				temp->uint64 = 0;
 				state = TOKEN_HEX;
 			}
 			else if (ch == 'o' || ch == 'O') {
-				temp->data = 0;
+				temp->uint64 = 0;
 				state = TOKEN_OCT;
 			}
 			else if (ch == 'd' || ch == 'D') {
-				temp->data = 0;
+				temp->uint64 = 0;
 				state = TOKEN_DEC;
 			}
 			else if (ch == ' ' || ch == '\t' || ch == '\r' || ch == '\n') {
@@ -407,7 +412,7 @@ ModuleVariable* varConstStr(const char* str)
 		}break;
 		case TOKEN_BIN: {
 			if (ch == '0' || ch == '1') {
-				temp->data = temp->data * 2 + ch - '0';
+				temp->uint64 = temp->uint64 * 2 + ch - '0';
 				numvalid = 1;
 			}
 			else if (ch == '_') {
@@ -418,7 +423,7 @@ ModuleVariable* varConstStr(const char* str)
 		}break;
 		case TOKEN_OCT: {
 			if (ch >= '0' && ch <= '7') {
-				temp->data = temp->data * 8 + ch - '0';
+				temp->uint64 = temp->uint64 * 8 + ch - '0';
 				numvalid = 1;
 			}
 			else if (ch == '_') {
@@ -429,7 +434,7 @@ ModuleVariable* varConstStr(const char* str)
 		}break;
 		case TOKEN_DEC: {
 			if (ch >= '0' && ch <= '9') {
-				temp->data = temp->data * 10 + ch - '0';
+				temp->uint64 = temp->uint64 * 10 + ch - '0';
 				numvalid = 1;
 			}
 			else if (ch == '_') {
@@ -440,13 +445,13 @@ ModuleVariable* varConstStr(const char* str)
 		}break;
 		case TOKEN_HEX: {
 			if ((ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f') || (ch >= 'A' && ch <= 'F')) {
-				temp->data = temp->data * 16;
+				temp->uint64 = temp->uint64 * 16;
 				if (ch > '0' && ch <= '9')
-					temp->data += ch - '0';
+					temp->uint64 += ch - '0';
 				else if (ch >= 'a' && ch <= 'f')
-					temp->data += ch - 'a' + 10;
+					temp->uint64 += ch - 'a' + 10;
 				else if (ch >= 'A' && ch <= 'F')
-					temp->data += ch - 'A' + 10;
+					temp->uint64 += ch - 'A' + 10;
 			}
 			else if (ch == '_') {
 			}
@@ -464,8 +469,8 @@ lastnum:
 		width = 1;
 	temp->width = width;
 	if (width < 64 && temp->isunsigned == 0) {
-		if (temp->data & (1llu << (width - 1))) {
-			temp->data |= 0xffffffffffffffffllu << width;
+		if (temp->uint64 & (1llu << (width - 1))) {
+			temp->uint64 |= 0xffffffffffffffffllu << width;
 		}
 	}
 	return temp;
@@ -486,22 +491,22 @@ int varGetStr(ModuleVariable* pobj, int base, char* str, int buflen)
 			buflen = pobj->width;
 		}
 		for (i = 0; i < buflen; i++) {
-			str[buflen - 1 - i] = (pobj->data & (1llu << i)) ? '1' : '0';
+			str[buflen - 1 - i] = (pobj->uint64 & (1llu << i)) ? '1' : '0';
 		}
 		str[buflen] = '\0';
 
 	}
 	else if (base == 10) {
 		if (pobj->isunsigned)
-			sprintf(str, "%llu", pobj->data);
+			sprintf(str, "%llu", pobj->uint64);
 		else
 			sprintf(str, "%lld", pobj->int64);
 	}
 	else {
 		if (pobj->isunsigned)
-			sprintf(str, "%d'h%llx", bc, pobj->data);
+			sprintf(str, "%d'h%llx", bc, pobj->uint64);
 		else
-			sprintf(str, "%d'sh%llx", bc, pobj->data);
+			sprintf(str, "%d'sh%llx", bc, pobj->uint64);
 	}
 	return 0;
 }
@@ -523,6 +528,7 @@ ModuleVariable* varConstSigned(int width, long long v)
 
 ModuleVariable* varConcat(int multicount, ...)
 {
+	/* support 64 bits only, fixed me */
 	va_list argptr;
 	int i;
 	int width = 0;
@@ -534,8 +540,8 @@ ModuleVariable* varConcat(int multicount, ...)
 		if (item == NULL)
 			break;
 		width += itemwidth = item->width;
-		temp->data <<= itemwidth;
-		temp->data |= item->data & (0xffffffffffffffffllu >> (64 - itemwidth));
+		temp->uint64 <<= itemwidth;
+		temp->uint64 |= item->uint64 & (0xffffffffffffffffllu >> (64 - itemwidth));
 	} while (1);
 	va_end(argptr);
 	temp->width = width;
@@ -544,8 +550,8 @@ ModuleVariable* varConcat(int multicount, ...)
 		ModuleVariable* temp1 = temp;
 		temp = varTemp(multicount * width, 1);
 		for (i = 0; i < multicount; i++) {
-			temp->data <<= width;
-			temp->data |= temp1->data;
+			temp->uint64 <<= width;
+			temp->uint64 |= temp1->uint64;
 		}
 	}
 	return temp;
@@ -555,7 +561,7 @@ ModuleVariable* varAdd(ModuleVariable* a, ModuleVariable* b)
 {
 	ModuleVariable* temp;
 	temp = varTemp(max(a->width, b->width), a->isunsigned || b->isunsigned);
-	temp->data = a->data + b->data;
+	temp->uint64 = a->uint64 + b->uint64;
 	return temp;
 }
 
@@ -580,7 +586,7 @@ ModuleVariable* varEQ(ModuleVariable* a, ModuleVariable* b)
 {
 	ModuleVariable* temp;
 	temp = varTemp(1, 1);
-	temp->data = (a->data == b->data)?1:0;
+	temp->uint64 = VAR_NEQ(a->data, b->data)?0:1;
 	return temp;
 }
 
@@ -597,58 +603,80 @@ ModuleVariable* varUNOr(ModuleVariable* a);
 ModuleVariable* varUXor(ModuleVariable* a);
 ModuleVariable* varUNXor(ModuleVariable* a);
 
-ModuleVariable* varBit(ModuleVariable* a, int bit)
+ModuleVariable* varBit(ModuleVariable* a, unsigned int bit)
 {
 	ModuleVariable* temp;
+	int ind;
+	int bitind;
+	if (bit >= VARBITS)
+		return 0;
+	ind = bit / 64;
+	bitind = bit % 64;
 	temp = varTemp(1, 1);
-	temp->data = (a->data & (1llu << bit)) ? 1 : 0;
+	temp->uint64 = (a->data64[ind] & (1llu << bitind)) ? 1 : 0;
 	return temp;
 }
 
-ModuleVariable* varBits(ModuleVariable* a, int from, int width)
+ModuleVariable* varBits(ModuleVariable* a, unsigned int from, unsigned int width)
 {
 	ModuleVariable* temp;
 	temp = varTemp(width, 1);
-	temp->data = (a->data >> from) & (0xffffffffffffffffllu >> (64 - width));
+	if (from >= a->width)
+		return temp;
+	if (width + from > a->width)
+		width = a->width - from;
+	if (width <= 0)
+		return temp;
+#if VARBITS <= 64
+	temp->uint64 = (a->uint64 >> from) & (0xffffffffffffffffllu >> (64 - width));
+#else
+	{
+		int i;
+		temp = varSHR(a, from);
+		temp->isunsigned = 1;
+		varSetWidth(temp, width);
+	}
+#endif
 	return temp;
 }
 
 ModuleVariable* varSelect(ModuleVariable* a, ModuleVariable* b, ModuleVariable* c)
 {
-	if (a->data)
+	if (a->uint64)
 		return b;
 	else
 		return c;
 }
 
-int varAssign(ModuleVariable* a, ModuleVariable* b);
+int varAssign(ModuleVariable* a, ModuleVariable* b)
+{
+	if (a->type = VTYPE_REG)
+		VAR_COPYDATA(a->data_reg, b->data);
+	else
+		VAR_COPYDATA(a->data, b->data);
+	return 0;
+}
+
 int varIsZero(ModuleVariable* a);
 int varIsNotZero(ModuleVariable* a);
 
-int varSetBits32(ModuleVariable* a, int index, unsigned int v)
+int varSetBits32(ModuleVariable* a, unsigned int index, unsigned int v)
 {
+	if (index >= VARBITS / 32)
+		return -1;
 	if (a->type == VTYPE_REG) {
-		if (index == 0)
-			a->uint32_reg = v;
-		else
-			a->data_reg = (a->data_reg & 0xffffffff) | (((unsigned long long)v) << 32);
+		a->data_reg[index] = v;
 	}
 	else {
-		if (index == 0)
-			a->uint32 = v;
-		else
-			a->data = (a->data & 0xffffffff) | (((unsigned long long)v) << 32);
+		a->data[index] = v;
 	}
 	return 0;
 }
-unsigned int varGetBits32(ModuleVariable* a, int index)
+unsigned int varGetBits32(ModuleVariable* a, unsigned int index)
 {
-	unsigned int v;
-	if (index == 0)
-		v = a->uint32;
-	else
-		v = (unsigned int)(a->data >> 32);
-	return v;
+	if (index >= VARBITS / 32)
+		return 0;
+	return a->data[index];
 }
 
 int varAssign_S32(ModuleVariable* a, int v)

hdl4sesim/src/hdl4sesim.c

@@ -241,7 +241,7 @@ static int hdl4sesim_hdl4se_simulator_SetReset(HOBJECT object, int reset)
 	return 0;
 }
 
-#define THREADCOUNT 9
+#define THREADCOUNT 3
 
 static int hdl4sesim_hdl4se_simulator_ClkTick(HOBJECT object)
 {
@@ -281,9 +281,6 @@ static int hdl4sesim_hdl4se_simulator_Setup(HOBJECT object)
 	pobj->clk = 1;
 	pobj->clocks++;
 	vartempClean();
-#if DEBUGDEPEND
-	printf("%lld:\n", pobj->clocks);
-#endif
 	hdl4se_module_Setup(&pobj->data);
 	for (i = 0; i < pobj->singlethreadmodules.itemcount; i++) {
 		hdl4se_module_Setup(&((IHDL4SEModuleVar*)(pobj->singlethreadmodules.array[i]))->data);
@@ -292,11 +289,6 @@ static int hdl4sesim_hdl4se_simulator_Setup(HOBJECT object)
 	for (i = 0; i < pobj->multithreadmodules.itemcount; i++) {
 		hdl4se_module_Setup(&((IHDL4SEModuleVar*)(pobj->multithreadmodules.array[i]))->data);
 	}
-#if 0
-	for (i = 0; i < pobj->multithreadmodules.itemcount; i++) {
-		checkmodule(&((IHDL4SEModuleVar*)(pobj->multithreadmodules.array[i]))->data);
-	}
-#endif
 	return 0;
 }
 
@@ -318,89 +310,82 @@ IHDL4SESimulator** hdl4sesimCreateSimulator()
 
 /*
 i5 6200U 2.3G
-threadcount=4
-DEPEND_OPT=1
-clocks: 1769472, TSPD=4874.578512cps, LSPD=4681.142857cps
+threadcount=5
+LOCKED_UPDATE=0 DEPEND_OPT=1
+clocks: 1769472, TSPD=7134.967742cps, LSPD=7281.777778cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=4959.481081cps, LSPD=9362.285714cps
 
-DEPEND_OPT=0
-clocks: 1769472, TSPD=3701.824268cps, LSPD=3449.263158cps
+threadcount=4
+LOCKED_UPDATE=0 DEPEND_OPT=1
+clocks: 1769472, TSPD=8891.819095cps, LSPD=10922.666667cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=3707.086869cps, LSPD=3855.058824cps
 
-threadcount=3
-DEPEND_OPT=1
-clocks: 1769472, TSPD=6505.411765cps, LSPD=9362.285714cps
+LOCKED_UPDATE=1 DEPEND_OPT=1
+clocks: 1769472, TSPD=9120.989691cps, LSPD=10922.666667cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=6461.295775cps, LSPD=5461.333333cps
 
-DEPEND_OPT=0
-clocks: 1769472, TSPD=4021.527273cps, LSPD=5957.818182cps
+threadcount=3
+LOCKED_UPDATE=0 DEPEND_OPT=1
+clocks: 1769472, TSPD=9830.400000cps, LSPD=10922.666667cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=4015.334792cps, LSPD=3855.058824cps
 
-threadcount=2
-DEPEND_OPT=1
-clocks: 1769472, TSPD=6652.150376cps, LSPD=5461.333333cps
+LOCKED_UPDATE=1 DEPEND_OPT=1
+clocks: 1769472, TSPD=10053.818182cps, LSPD=10922.666667cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=6746.352941cps, LSPD=10922.666667cps
 
-DEPEND_OPT=0
-clocks: 1769472, TSPD=4086.540416cps, LSPD=4369.066667cps
+threadcount=2
+LOCKED_UPDATE=0 DEPEND_OPT=1
+clocks: 1769472, TSPD=8631.570732cps, LSPD=9362.285714cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=4086.877506cps, LSPD=4096.000000cps
 
-threadcount=1
-DEPEND_OPT=1
-clocks: 1769472, TSPD=4915.200000cps, LSPD=5957.818182cps
+LOCKED_UPDATE=1 DEPEND_OPT=1
+clocks: 1769472, TSPD=8759.762376cps, LSPD=9362.285714cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=4932.817204cps, LSPD=5461.333333cps
 
-DEPEND_OPT=0
-clocks: 1769472, TSPD=2886.577488cps, LSPD=2978.909091cps
+threadcount=1
+LOCKED_UPDATE=1  DEPEND_OPT=1
+clocks: 1769472, TSPD=6365.007194cps, LSPD=6553.600000cps
 softmax_168 complete
 0.4901  n03085013 computer keyboard, keypad
 0.1591  n04264628 space bar
 0.0543  n04074963 remote control, remote
 0.0439  n03832673 notebook, notebook computer
 0.0356  n04505470 typewriter keyboard
-clocks: 1835008, TSPD=2894.334385cps, LSPD=3120.761905cps
 
 i9 10900 32G
 threadcount=32

hdl4sesim/src/hdl4sevcdfile.c

@@ -225,8 +225,8 @@ static int hdl4sevcdfile_output_item_data(IDListVarPtr item, sHDL4SEVcdFile* pob
 #else
 			pobj->tempvar->width = pitem->width;
 			hdl4se_module_GetValue(&pitem->detector->data, pitem->index, pitem->width, pobj->tempvar);
-			if (pobj->tempvar->data != pitem->lastvalue->data) {
-				pitem->lastvalue->data = pobj->tempvar->data;
+			if (VAR_NEQ(pobj->tempvar->data, pitem->lastvalue->data)) {
+				VAR_COPYDATA(pitem->lastvalue->data, pobj->tempvar->data);
 				varGetStr(pitem->lastvalue, 2, pobj->tempbuf, MAXWIDTH);
 				fprintf(pobj->pFile, "b%s %s\n", pobj->tempbuf, pitem->vcdname);
 			}

hdl4seutils/src/threadlock.c

@@ -51,7 +51,7 @@ typedef struct _ThreadLock {
 THREADLOCK threadlockCreate()
 {
 	ThreadLock* pSection = NULL;
-	pSection = (ThreadLock *)mt_malloc(sizeof(ThreadLock));
+	pSection = (ThreadLock *)malloc(sizeof(ThreadLock));
 	if (pSection != NULL) {
 		InitializeCriticalSection(&pSection->section);
 		pSection->refcount = 1;
@@ -78,7 +78,7 @@ int threadlockUnlock(THREADLOCK lock)
 int threadlockDestroy(THREADLOCK lock)
 {
 	if (lock != NULL)
-		mt_free(((ThreadLock*)lock));
+		free(((ThreadLock*)lock));
 	return 0;
 }
 
@@ -180,14 +180,14 @@ int crtinit()
 
 void* mt_malloc(size_t size)
 {
-	threadlockLock(crtlock);
+	//threadlockLock(crtlock);
 	void* ret = malloc(size);
-	threadlockUnlock(crtlock);
+	//threadlockUnlock(crtlock);
 	return ret;
 }
 void  mt_free(void* data)
 {
-	threadlockLock(crtlock);
+	//threadlockLock(crtlock);
 	free(data);
-	threadlockUnlock(crtlock);
+	//threadlockUnlock(crtlock);
 }