--- a/SHA1Stream.st Sat Mar 23 15:18:05 2019 +0100
+++ b/SHA1Stream.st Sat Mar 23 15:30:15 2019 +0100
@@ -51,11 +51,22 @@
#if defined(__GNUC__) || defined(__CLANG__)
# define STATIC_INLINE static inline
+# define __SSE4_1__
+# define __SHA__
+# include <immintrin.h>
+ static void SHA1Transform(unsigned int32 state[5], unsigned char buffer[64]);
+ // static void __attribute__ ((__target__ ("sha,sse4.1"))) SHA1Transform_x86(unsigned int32 state[5], unsigned char buffer[64]);
+ static void SHA1Transform_x86(unsigned int32 state[5], unsigned char buffer[64]);
+# define USE_SHA_INTRINSICS
+#endif
+
+#ifndef USE_SHA_INTRINSICS
+# define SHA1Transform SHA1Transform_generic
#endif
#if USE_ANSI_C
- void SHA1Transform(unsigned int32 /* long */ state[5], unsigned char buffer[64]);
+ static void SHA1Transform_generic(unsigned int32 state[5], unsigned char buffer[64]);
STATIC_INLINE void SHA1Init(SHA1_CTX* context);
void SHA1Update(SHA1_CTX* context, unsigned char* data, unsigned int len);
void SHA1Final(unsigned char digest[20], SHA1_CTX* context);
@@ -76,7 +87,6 @@
! !
!SHA1Stream primitiveFunctions!
-
%{
/*
@@ -111,9 +121,9 @@
*/
static void
#if USE_ANSI_C
-SHA1Transform (unsigned int32 state[5], unsigned char buffer[64])
+SHA1Transform_generic (unsigned int32 state[5], unsigned char buffer[64])
#else
-SHA1Transform (state, buffer)
+SHA1Transform_generic (state, buffer)
unsigned int32 state[5];
unsigned char buffer[64];
#endif
@@ -168,6 +178,221 @@
a = b = c = d = e = 0;
}
+#if defined(__CLANG__)
+
+static void
+SHA1Transform_x86 (unsigned int32 state[5], unsigned char buffer[64])
+{
+ unsigned int32 a, b, c, d, e;
+ typedef union {
+ unsigned char c[64];
+ unsigned int32 /* long */ l[16];
+ } CHAR64LONG16;
+ CHAR64LONG16* block;
+#ifdef SHA1HANDSOFF
+ static unsigned char workspace[64];
+ block = (CHAR64LONG16*)workspace;
+ memcpy(block, buffer, 64);
+#else
+ block = (CHAR64LONG16*)buffer;
+#endif
+ __m128i ABCD, ABCD_SAVE, E0, E0_SAVE, E1;
+ __m128i MSG0, MSG1, MSG2, MSG3;
+ const __m128i MASK = _mm_set_epi64x(0x0001020304050607ULL, 0x08090a0b0c0d0e0fULL);
+
+ /* Load initial values */
+ ABCD = _mm_loadu_si128((const __m128i*) state);
+ E0 = _mm_set_epi32(state[4], 0, 0, 0);
+ ABCD = _mm_shuffle_epi32(ABCD, 0x1B);
+
+ /* Save current state */
+ ABCD_SAVE = ABCD;
+ E0_SAVE = E0;
+
+ /* Rounds 0-3 */
+ MSG0 = _mm_loadu_si128((const __m128i*)(block + 0));
+ MSG0 = _mm_shuffle_epi8(MSG0, MASK);
+ E0 = _mm_add_epi32(E0, MSG0);
+ E1 = ABCD;
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 0);
+
+ /* Rounds 4-7 */
+ MSG1 = _mm_loadu_si128((const __m128i*)(block + 16));
+ MSG1 = _mm_shuffle_epi8(MSG1, MASK);
+ E1 = _mm_sha1nexte_epu32(E1, MSG1);
+ E0 = ABCD;
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 0);
+ MSG0 = _mm_sha1msg1_epu32(MSG0, MSG1);
+
+ /* Rounds 8-11 */
+ MSG2 = _mm_loadu_si128((const __m128i*)(block + 32));
+ MSG2 = _mm_shuffle_epi8(MSG2, MASK);
+ E0 = _mm_sha1nexte_epu32(E0, MSG2);
+ E1 = ABCD;
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 0);
+ MSG1 = _mm_sha1msg1_epu32(MSG1, MSG2);
+ MSG0 = _mm_xor_si128(MSG0, MSG2);
+
+ /* Rounds 12-15 */
+ MSG3 = _mm_loadu_si128((const __m128i*)(block + 48));
+ MSG3 = _mm_shuffle_epi8(MSG3, MASK);
+ E1 = _mm_sha1nexte_epu32(E1, MSG3);
+ E0 = ABCD;
+ MSG0 = _mm_sha1msg2_epu32(MSG0, MSG3);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 0);
+ MSG2 = _mm_sha1msg1_epu32(MSG2, MSG3);
+ MSG1 = _mm_xor_si128(MSG1, MSG3);
+
+ /* Rounds 16-19 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG0);
+ E1 = ABCD;
+ MSG1 = _mm_sha1msg2_epu32(MSG1, MSG0);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 0);
+ MSG3 = _mm_sha1msg1_epu32(MSG3, MSG0);
+ MSG2 = _mm_xor_si128(MSG2, MSG0);
+
+ /* Rounds 20-23 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG1);
+ E0 = ABCD;
+ MSG2 = _mm_sha1msg2_epu32(MSG2, MSG1);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 1);
+ MSG0 = _mm_sha1msg1_epu32(MSG0, MSG1);
+ MSG3 = _mm_xor_si128(MSG3, MSG1);
+
+ /* Rounds 24-27 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG2);
+ E1 = ABCD;
+ MSG3 = _mm_sha1msg2_epu32(MSG3, MSG2);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 1);
+ MSG1 = _mm_sha1msg1_epu32(MSG1, MSG2);
+ MSG0 = _mm_xor_si128(MSG0, MSG2);
+
+ /* Rounds 28-31 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG3);
+ E0 = ABCD;
+ MSG0 = _mm_sha1msg2_epu32(MSG0, MSG3);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 1);
+ MSG2 = _mm_sha1msg1_epu32(MSG2, MSG3);
+ MSG1 = _mm_xor_si128(MSG1, MSG3);
+
+ /* Rounds 32-35 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG0);
+ E1 = ABCD;
+ MSG1 = _mm_sha1msg2_epu32(MSG1, MSG0);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 1);
+ MSG3 = _mm_sha1msg1_epu32(MSG3, MSG0);
+ MSG2 = _mm_xor_si128(MSG2, MSG0);
+
+ /* Rounds 36-39 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG1);
+ E0 = ABCD;
+ MSG2 = _mm_sha1msg2_epu32(MSG2, MSG1);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 1);
+ MSG0 = _mm_sha1msg1_epu32(MSG0, MSG1);
+ MSG3 = _mm_xor_si128(MSG3, MSG1);
+
+ /* Rounds 40-43 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG2);
+ E1 = ABCD;
+ MSG3 = _mm_sha1msg2_epu32(MSG3, MSG2);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 2);
+ MSG1 = _mm_sha1msg1_epu32(MSG1, MSG2);
+ MSG0 = _mm_xor_si128(MSG0, MSG2);
+
+ /* Rounds 44-47 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG3);
+ E0 = ABCD;
+ MSG0 = _mm_sha1msg2_epu32(MSG0, MSG3);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 2);
+ MSG2 = _mm_sha1msg1_epu32(MSG2, MSG3);
+ MSG1 = _mm_xor_si128(MSG1, MSG3);
+
+ /* Rounds 48-51 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG0);
+ E1 = ABCD;
+ MSG1 = _mm_sha1msg2_epu32(MSG1, MSG0);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 2);
+ MSG3 = _mm_sha1msg1_epu32(MSG3, MSG0);
+ MSG2 = _mm_xor_si128(MSG2, MSG0);
+
+ /* Rounds 52-55 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG1);
+ E0 = ABCD;
+ MSG2 = _mm_sha1msg2_epu32(MSG2, MSG1);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 2);
+ MSG0 = _mm_sha1msg1_epu32(MSG0, MSG1);
+ MSG3 = _mm_xor_si128(MSG3, MSG1);
+
+ /* Rounds 56-59 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG2);
+ E1 = ABCD;
+ MSG3 = _mm_sha1msg2_epu32(MSG3, MSG2);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 2);
+ MSG1 = _mm_sha1msg1_epu32(MSG1, MSG2);
+ MSG0 = _mm_xor_si128(MSG0, MSG2);
+
+ /* Rounds 60-63 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG3);
+ E0 = ABCD;
+ MSG0 = _mm_sha1msg2_epu32(MSG0, MSG3);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 3);
+ MSG2 = _mm_sha1msg1_epu32(MSG2, MSG3);
+ MSG1 = _mm_xor_si128(MSG1, MSG3);
+
+ /* Rounds 64-67 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG0);
+ E1 = ABCD;
+ MSG1 = _mm_sha1msg2_epu32(MSG1, MSG0);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 3);
+ MSG3 = _mm_sha1msg1_epu32(MSG3, MSG0);
+ MSG2 = _mm_xor_si128(MSG2, MSG0);
+
+ /* Rounds 68-71 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG1);
+ E0 = ABCD;
+ MSG2 = _mm_sha1msg2_epu32(MSG2, MSG1);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 3);
+ MSG3 = _mm_xor_si128(MSG3, MSG1);
+
+ /* Rounds 72-75 */
+ E0 = _mm_sha1nexte_epu32(E0, MSG2);
+ E1 = ABCD;
+ MSG3 = _mm_sha1msg2_epu32(MSG3, MSG2);
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E0, 3);
+
+ /* Rounds 76-79 */
+ E1 = _mm_sha1nexte_epu32(E1, MSG3);
+ E0 = ABCD;
+ ABCD = _mm_sha1rnds4_epu32(ABCD, E1, 3);
+
+ /* Combine state */
+ E0 = _mm_sha1nexte_epu32(E0, E0_SAVE);
+ ABCD = _mm_add_epi32(ABCD, ABCD_SAVE);
+
+ /* Save state */
+ ABCD = _mm_shuffle_epi32(ABCD, 0x1B);
+ _mm_storeu_si128((__m128i*) state, ABCD);
+ state[4] = _mm_extract_epi32(E0, 3);
+}
+
+#endif
+
+#ifdef USE_SHA_INTRINSICS
+
+static void
+#if USE_ANSI_C
+SHA1Transform(unsigned int32 state[5], unsigned char buffer[64])
+#else
+SHA1Transform(state, buffer)
+ unsigned int32 state[5];
+ unsigned char buffer[64];
+#endif
+{
+ SHA1Transform_generic(state, buffer);
+}
+
+#endif
+
/*
* SHA1Init - Initialize new context
*/
@@ -240,7 +465,7 @@
>> ((3-(i & 3)) * 8) ) & 255); /* Endian independent */
}
SHA1Update(context, (unsigned char *)"\200", 1);
- while ((context->count[0] & 504) != 448) {
+ while ((context->count[0] & 504 /* 16r1F8 */) != 448 /* 16r1C0 */) {
SHA1Update(context, (unsigned char *)"\0", 1);
}
SHA1Update(context, finalcount, 8); /* Should cause a SHA1Transform() */
@@ -444,27 +669,31 @@
CPU cc algo mb/sec
MAC (2010 macbook; 2.7Ghz Duo) clang -O2 slow 128.5
- 132
- MAC (2012 macbook; 2.6Ghz I7) clang -O2 203.5
+ 132
+ MAC (2012 macbook; 2.6Ghz I7) clang -O2 190
+ chunk size 10: 86.70 Mb/s
+ chunk size 50: 227.07 Mb/s
+ chunk size 1000: 405.82 Mb/s
+ chunk size 50000: 421.98 Mb/s
timing throughput:
- [exBegin]
+ [exBegin]
|hashStream n t|
hashStream := SHA1Stream new.
n := 1000000.
t := Time millisecondsToRun:[
- n timesRepeat:[
- hashStream nextPutAll:'12345678901234567890123456789012345678901234567890'.
- ].
- ].
+ n timesRepeat:[
+ hashStream nextPutAll:'12345678901234567890123456789012345678901234567890'.
+ ].
+ ].
t := (t / 1000) asFloat.
Transcript show:t; show:' seconds for '; show:(50*n/1024) asFloat; showCR:' Kb'.
Transcript show:(n*50/1024 / t); showCR:' Kb/s'
- [exEnd]
+ [exEnd]
"
! !
@@ -609,20 +838,26 @@
"update the hash value with anInteger <= 255."
%{
- if (__isSmallInteger(anInteger) && __intVal(anInteger) <= 255
- && __isByteArray(__INST(hashContext))
- && __byteArraySize(__INST(hashContext)) == sizeof(SHA1_CTX)
+ unsigned char value;
+ OBJ _hashContext = __INST(hashContext);
+
+ // fetch first; check below
+ value = __intVal(anInteger);
+ if (__isSmallInteger(anInteger) && value <= 255
+ && __isByteArray(_hashContext)
+ && __byteArraySize(_hashContext) == sizeof(SHA1_CTX)
) {
- SHA1_CTX *ctx = (SHA1_CTX *)__ByteArrayInstPtr(__INST(hashContext))->ba_element;
- unsigned char value = __intVal(anInteger);
+ SHA1_CTX *ctx = (SHA1_CTX *)__byteArrayVal(_hashContext);
- SHA1Update(ctx, &value, 1);
- RETURN(self);
+ SHA1Update(ctx, &value, 1);
+ RETURN(self);
}
bad: ;
%}.
^ self primitiveFailed
+
+ "Modified: / 23-03-2019 / 15:19:40 / Claus Gittinger"
!
nextPutBytes:count from:anObject startingAt:start
@@ -636,69 +871,78 @@
INT objSize;
int nInstBytes;
char *extPtr;
+ OBJ _hashContext = __INST(hashContext);
- if (__isByteArray(__INST(hashContext))
- && __byteArraySize(__INST(hashContext)) == sizeof(SHA1_CTX)
+ // convert here; check later
+ len = __intVal(count);
+ offs = __intVal(start) - 1;
+
+ if (__isByteArray(_hashContext)
+ && __byteArraySize(_hashContext) == sizeof(SHA1_CTX)
&& __bothSmallInteger(count, start)
- ) {
- SHA1_CTX *ctx = (SHA1_CTX *)__ByteArrayInstPtr(__INST(hashContext))->ba_element;
-
- len = __intVal(count);
- offs = __intVal(start) - 1;
-
- if (__isExternalBytesLike(anObject)) {
- OBJ sz;
+ ) {
+ SHA1_CTX *ctx = (SHA1_CTX *)__byteArrayVal(_hashContext);
- nInstBytes = 0;
- extPtr = (char *)__externalBytesAddress(anObject);
- if (extPtr == NULL) goto bad;
- sz = __externalBytesSize(anObject);
- if (__isSmallInteger(sz)) {
- objSize = __intVal(sz);
- } else {
- objSize = 0; /* unknown */
- }
- } else {
- OBJ oClass = __Class(anObject);
- int nInstVars = __intVal(__ClassInstPtr(oClass)->c_ninstvars);
+ if (__isByteArrayLike(anObject)) {
+ extPtr = (char *)__byteArrayVal(anObject);
+ objSize = __byteArraySize(anObject);
+ } else if (__isStringLike(anObject)) {
+ extPtr = (char *)__stringVal(anObject);
+ objSize = __stringSize(anObject);
+ } else if (__isExternalBytesLike(anObject)) {
+ OBJ sz;
+
+ nInstBytes = 0;
+ extPtr = (char *)__externalBytesAddress(anObject);
+ if (extPtr == NULL) goto bad;
+ sz = __externalBytesSize(anObject);
+ objSize = __intVal(sz);
+ if (!__isSmallInteger(sz)) {
+ objSize = 0; /* unknown */
+ }
+ } else {
+ OBJ oClass = __Class(anObject);
+ int nInstVars = __intVal(__ClassInstPtr(oClass)->c_ninstvars);
- nInstBytes = OHDR_SIZE + __OBJS2BYTES__(nInstVars);
- switch (__intVal(__ClassInstPtr(oClass)->c_flags) & ARRAYMASK) {
- case BYTEARRAY:
- case WORDARRAY:
- case LONGARRAY:
- case SWORDARRAY:
- case SLONGARRAY:
- case FLOATARRAY:
- break;
- case DOUBLEARRAY:
+ nInstBytes = OHDR_SIZE + __OBJS2BYTES__(nInstVars);
+ switch (__intVal(__ClassInstPtr(oClass)->c_flags) & ARRAYMASK) {
+ case BYTEARRAY:
+ case WORDARRAY:
+ case LONGARRAY:
+ case SWORDARRAY:
+ case SLONGARRAY:
+ case FLOATARRAY:
+ break;
+ case DOUBLEARRAY:
#ifdef __NEED_DOUBLE_ALIGN
- nInstBytes = (nInstBytes-1+__DOUBLE_ALIGN) &~ (__DOUBLE_ALIGN-1);
+ nInstBytes = (nInstBytes-1+__DOUBLE_ALIGN) &~ (__DOUBLE_ALIGN-1);
#endif
- break;
- case LONGLONGARRAY:
- case SLONGLONGARRAY:
+ break;
+ case LONGLONGARRAY:
+ case SLONGLONGARRAY:
#ifdef __NEED_LONGLONG_ALIGN
- nInstBytes = (nInstBytes-1+__LONGLONG_ALIGN) &~ (__LONGLONG_ALIGN-1);
+ nInstBytes = (nInstBytes-1+__LONGLONG_ALIGN) &~ (__LONGLONG_ALIGN-1);
#endif
- break;
- default:
- goto bad;
- }
- // nInstBytes is the number of bytes occupied by pointer instance variables
- // subtract from size and add to byte-pointer
- objSize = __Size(anObject) - nInstBytes;
- extPtr = (char *)anObject + nInstBytes;
- }
- if ((offs >= 0) && (len >= 0) && (objSize >= (len + offs))) {
- SHA1Update(ctx, extPtr+offs, (unsigned int)len);
- RETURN (count);
- }
+ break;
+ default:
+ goto bad;
+ }
+ // nInstBytes is the number of bytes occupied by pointer instance variables
+ // subtract from size and add to byte-pointer
+ objSize = __Size(anObject) - nInstBytes;
+ extPtr = (char *)anObject + nInstBytes;
+ }
+ if ((offs >= 0) && (len >= 0) && (objSize >= (len + offs))) {
+ SHA1Update(ctx, extPtr+offs, (unsigned int)len);
+ RETURN (count);
+ }
}
bad: ;
%}.
^ self primitiveFailed
+
+ "Modified: / 23-03-2019 / 15:28:31 / Claus Gittinger"
! !
!SHA1Stream class methodsFor:'documentation'!