dengzhihao
/
CMSIS-DSP
0
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merged PR #166

Browse Source
pull/19/head
GuentherMartin 7 years ago
parent 06eff37277
commit da38c27dd0
10 changed files with 124 additions and 236 deletions
Show Stats Download Patch File Download Diff File

18
Source/FilteringFunctions/arm_conv_f32.c
Unescape Escape View File

@ -294,9 +294,9 @@ void arm_conv_f32(
acc3 = 0.0f; acc3 = 0.0f;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -306,7 +306,7 @@ void arm_conv_f32(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px); x3 = *(px);
@ -325,7 +325,7 @@ void arm_conv_f32(
acc3 += x3 * c0; acc3 += x3 * c0;
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c0 = *(py--); c0 = *py--;
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px + 1U); x0 = *(px + 1U);
@ -341,7 +341,7 @@ void arm_conv_f32(
acc3 += x0 * c0; acc3 += x0 * c0;
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px + 2U); x1 = *(px + 2U);
@ -357,7 +357,7 @@ void arm_conv_f32(
acc3 += x1 * c0; acc3 += x1 * c0;
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c0 = *(py--); c0 = *py--;
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px + 3U); x2 = *(px + 3U);
@ -383,10 +383,10 @@ void arm_conv_f32(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */

26
Source/FilteringFunctions/arm_conv_fast_q31.c
Unescape Escape View File

@ -248,9 +248,9 @@ void arm_conv_fast_q31(
acc3 = 0; acc3 = 0;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -260,10 +260,10 @@ void arm_conv_fast_q31(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[0] * y[srcBLen - 1] */ /* acc0 += x[0] * y[srcBLen - 1] */
@ -279,10 +279,10 @@ void arm_conv_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32);
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c0 = *(py--); c0 = *py--;
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px++); x0 = *px++;
/* Perform the multiply-accumulate */ /* Perform the multiply-accumulate */
/* acc0 += x[1] * y[srcBLen - 2] */ /* acc0 += x[1] * y[srcBLen - 2] */
@ -295,10 +295,10 @@ void arm_conv_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32);
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px++); x1 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[2] * y[srcBLen - 3] */ /* acc0 += x[2] * y[srcBLen - 3] */
@ -311,10 +311,10 @@ void arm_conv_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32);
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c0 = *(py--); c0 = *py--;
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px++); x2 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[3] * y[srcBLen - 4] */ /* acc0 += x[3] * y[srcBLen - 4] */
@ -336,10 +336,10 @@ void arm_conv_fast_q31(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */

58
Source/FilteringFunctions/arm_conv_partial_f32.c
Unescape Escape View File

@ -263,12 +263,13 @@ arm_status arm_conv_partial_f32(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -298,9 +299,9 @@ arm_status arm_conv_partial_f32(
acc3 = 0.0f; acc3 = 0.0f;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -310,10 +311,10 @@ arm_status arm_conv_partial_f32(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulate */ /* Perform the multiply-accumulate */
/* acc0 += x[0] * y[srcBLen - 1] */ /* acc0 += x[0] * y[srcBLen - 1] */
@ -329,10 +330,10 @@ arm_status arm_conv_partial_f32(
acc3 += x3 * c0; acc3 += x3 * c0;
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c0 = *(py--); c0 = *py--;
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px++); x0 = *px++;
/* Perform the multiply-accumulate */ /* Perform the multiply-accumulate */
/* acc0 += x[1] * y[srcBLen - 2] */ /* acc0 += x[1] * y[srcBLen - 2] */
@ -345,10 +346,10 @@ arm_status arm_conv_partial_f32(
acc3 += x0 * c0; acc3 += x0 * c0;
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px++); x1 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[2] * y[srcBLen - 3] */ /* acc0 += x[2] * y[srcBLen - 3] */
@ -361,10 +362,10 @@ arm_status arm_conv_partial_f32(
acc3 += x1 * c0; acc3 += x1 * c0;
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c0 = *(py--); c0 = *py--;
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px++); x2 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[3] * y[srcBLen - 4] */ /* acc0 += x[3] * y[srcBLen - 4] */
@ -386,10 +387,10 @@ arm_status arm_conv_partial_f32(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */
@ -420,14 +421,7 @@ arm_status arm_conv_partial_f32(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -480,14 +474,7 @@ arm_status arm_conv_partial_f32(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -524,14 +511,7 @@ arm_status arm_conv_partial_f32(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

36
Source/FilteringFunctions/arm_conv_partial_fast_q15.c
Unescape Escape View File

@ -270,12 +270,13 @@ arm_status arm_conv_partial_fast_q15(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -473,7 +474,7 @@ arm_status arm_conv_partial_fast_q15(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
px = pIn1 + count; px = pSrc1 + count;
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -526,14 +527,7 @@ arm_status arm_conv_partial_fast_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -570,14 +564,7 @@ arm_status arm_conv_partial_fast_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -929,12 +916,13 @@ arm_status arm_conv_partial_fast_q15(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -1255,7 +1243,7 @@ arm_status arm_conv_partial_fast_q15(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
px = pIn1 + count; px = pSrc1 + count;
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -1308,7 +1296,7 @@ arm_status arm_conv_partial_fast_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
px = pIn1 + count; px = pSrc1 + count;
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -1345,7 +1333,7 @@ arm_status arm_conv_partial_fast_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
px = pIn1 + count; px = pSrc1 + count;
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

58
Source/FilteringFunctions/arm_conv_partial_fast_q31.c
Unescape Escape View File

@ -234,12 +234,13 @@ arm_status arm_conv_partial_fast_q31(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -269,9 +270,9 @@ arm_status arm_conv_partial_fast_q31(
acc3 = 0; acc3 = 0;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -281,10 +282,10 @@ arm_status arm_conv_partial_fast_q31(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulate */ /* Perform the multiply-accumulate */
/* acc0 += x[0] * y[srcBLen - 1] */ /* acc0 += x[0] * y[srcBLen - 1] */
@ -300,10 +301,10 @@ arm_status arm_conv_partial_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x3 * c0)) >> 32);
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c0 = *(py--); c0 = *py--;
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px++); x0 = *px++;
/* Perform the multiply-accumulate */ /* Perform the multiply-accumulate */
/* acc0 += x[1] * y[srcBLen - 2] */ /* acc0 += x[1] * y[srcBLen - 2] */
@ -316,10 +317,10 @@ arm_status arm_conv_partial_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x0 * c0)) >> 32);
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px++); x1 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[2] * y[srcBLen - 3] */ /* acc0 += x[2] * y[srcBLen - 3] */
@ -332,10 +333,10 @@ arm_status arm_conv_partial_fast_q31(
acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32); acc3 = (q31_t) ((((q63_t) acc3 << 32) + ((q63_t) x1 * c0)) >> 32);
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c0 = *(py--); c0 = *py--;
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px++); x2 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[3] * y[srcBLen - 4] */ /* acc0 += x[3] * y[srcBLen - 4] */
@ -357,10 +358,10 @@ arm_status arm_conv_partial_fast_q31(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */
@ -391,14 +392,7 @@ arm_status arm_conv_partial_fast_q31(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -456,14 +450,7 @@ arm_status arm_conv_partial_fast_q31(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -501,14 +488,7 @@ arm_status arm_conv_partial_fast_q31(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

32
Source/FilteringFunctions/arm_conv_partial_q15.c
Unescape Escape View File

@ -277,12 +277,13 @@ arm_status arm_conv_partial_q15(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -488,14 +489,7 @@ arm_status arm_conv_partial_q15(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -548,14 +542,7 @@ arm_status arm_conv_partial_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -592,14 +579,7 @@ arm_status arm_conv_partial_q15(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

40
Source/FilteringFunctions/arm_conv_partial_q31.c
Unescape Escape View File

@ -231,12 +231,13 @@ arm_status arm_conv_partial_q31(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -265,8 +266,8 @@ arm_status arm_conv_partial_q31(
acc2 = 0; acc2 = 0;
/* read x[0], x[1] samples */ /* read x[0], x[1] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
/* Apply loop unrolling and compute 3 MACs simultaneously. */ /* Apply loop unrolling and compute 3 MACs simultaneously. */
k = srcBLen / 3; k = srcBLen / 3;
@ -331,10 +332,10 @@ arm_status arm_conv_partial_q31(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x2 = *(px++); x2 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */
@ -361,14 +362,7 @@ arm_status arm_conv_partial_q31(
count += 3U; count += 3U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -421,14 +415,7 @@ arm_status arm_conv_partial_q31(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -465,14 +452,7 @@ arm_status arm_conv_partial_q31(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

58
Source/FilteringFunctions/arm_conv_partial_q7.c
Unescape Escape View File

@ -254,12 +254,13 @@ arm_status arm_conv_partial_q7(
/* Working pointer of inputA */ /* Working pointer of inputA */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
{ {
px = pIn1 + firstIndex - srcBLen + 1; pSrc1 = pIn1 + firstIndex - srcBLen + 1;
} }
else else
{ {
px = pIn1; pSrc1 = pIn1;
} }
px = pSrc1;
/* Working pointer of inputB */ /* Working pointer of inputB */
pSrc2 = pIn2 + (srcBLen - 1U); pSrc2 = pIn2 + (srcBLen - 1U);
@ -289,9 +290,9 @@ arm_status arm_conv_partial_q7(
acc3 = 0; acc3 = 0;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -301,12 +302,12 @@ arm_status arm_conv_partial_q7(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c1 = *(py--); c1 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px++); x3 = *px++;
/* x[0] and x[1] are packed */ /* x[0] and x[1] are packed */
in1 = (q15_t) x0; in1 = (q15_t) x0;
@ -342,7 +343,7 @@ arm_status arm_conv_partial_q7(
acc2 = __SMLAD(input1, input2, acc2); acc2 = __SMLAD(input1, input2, acc2);
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px++); x0 = *px++;
/* x[3] and x[4] are packed */ /* x[3] and x[4] are packed */
in1 = (q15_t) x3; in1 = (q15_t) x3;
@ -354,12 +355,12 @@ arm_status arm_conv_partial_q7(
acc3 = __SMLAD(input1, input2, acc3); acc3 = __SMLAD(input1, input2, acc3);
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c1 = *(py--); c1 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px++); x1 = *px++;
/* x[2] and x[3] are packed */ /* x[2] and x[3] are packed */
in1 = (q15_t) x2; in1 = (q15_t) x2;
@ -395,7 +396,7 @@ arm_status arm_conv_partial_q7(
acc2 = __SMLAD(input1, input2, acc2); acc2 = __SMLAD(input1, input2, acc2);
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px++); x2 = *px++;
/* x[5] and x[6] are packed */ /* x[5] and x[6] are packed */
in1 = (q15_t) x1; in1 = (q15_t) x1;
@ -415,10 +416,10 @@ arm_status arm_conv_partial_q7(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */
@ -449,14 +450,7 @@ arm_status arm_conv_partial_q7(
count += 4U; count += 4U;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
@ -531,14 +525,7 @@ arm_status arm_conv_partial_q7(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */
@ -575,14 +562,7 @@ arm_status arm_conv_partial_q7(
count++; count++;
/* Update the inputA and inputB pointers for next MAC calculation */ /* Update the inputA and inputB pointers for next MAC calculation */
if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0) px = pSrc1 + count;
{
px = pIn1 + firstIndex - srcBLen + 1 + count;
}
else
{
px = pIn1 + count;
}
py = pSrc2; py = pSrc2;
/* Decrement the loop counter */ /* Decrement the loop counter */

8
Source/FilteringFunctions/arm_conv_q31.c
Unescape Escape View File

@ -245,8 +245,8 @@ void arm_conv_q31(
acc2 = 0; acc2 = 0;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
/* Apply loop unrolling and compute 3 MACs simultaneously. */ /* Apply loop unrolling and compute 3 MACs simultaneously. */
k = srcBLen / 3; k = srcBLen / 3;
@ -310,10 +310,10 @@ void arm_conv_q31(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x2 = *(px++); x2 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */

26
Source/FilteringFunctions/arm_conv_q7.c
Unescape Escape View File

@ -264,9 +264,9 @@ void arm_conv_q7(
acc3 = 0; acc3 = 0;
/* read x[0], x[1], x[2] samples */ /* read x[0], x[1], x[2] samples */
x0 = *(px++); x0 = *px++;
x1 = *(px++); x1 = *px++;
x2 = *(px++); x2 = *px++;
/* Apply loop unrolling and compute 4 MACs simultaneously. */ /* Apply loop unrolling and compute 4 MACs simultaneously. */
k = srcBLen >> 2U; k = srcBLen >> 2U;
@ -276,12 +276,12 @@ void arm_conv_q7(
do do
{ {
/* Read y[srcBLen - 1] sample */ /* Read y[srcBLen - 1] sample */
c0 = *(py--); c0 = *py--;
/* Read y[srcBLen - 2] sample */ /* Read y[srcBLen - 2] sample */
c1 = *(py--); c1 = *py--;
/* Read x[3] sample */ /* Read x[3] sample */
x3 = *(px++); x3 = *px++;
/* x[0] and x[1] are packed */ /* x[0] and x[1] are packed */
in1 = (q15_t) x0; in1 = (q15_t) x0;
@ -317,7 +317,7 @@ void arm_conv_q7(
acc2 = __SMLAD(input1, input2, acc2); acc2 = __SMLAD(input1, input2, acc2);
/* Read x[4] sample */ /* Read x[4] sample */
x0 = *(px++); x0 = *px++;
/* x[3] and x[4] are packed */ /* x[3] and x[4] are packed */
in1 = (q15_t) x3; in1 = (q15_t) x3;
@ -329,12 +329,12 @@ void arm_conv_q7(
acc3 = __SMLAD(input1, input2, acc3); acc3 = __SMLAD(input1, input2, acc3);
/* Read y[srcBLen - 3] sample */ /* Read y[srcBLen - 3] sample */
c0 = *(py--); c0 = *py--;
/* Read y[srcBLen - 4] sample */ /* Read y[srcBLen - 4] sample */
c1 = *(py--); c1 = *py--;
/* Read x[5] sample */ /* Read x[5] sample */
x1 = *(px++); x1 = *px++;
/* x[2] and x[3] are packed */ /* x[2] and x[3] are packed */
in1 = (q15_t) x2; in1 = (q15_t) x2;
@ -370,7 +370,7 @@ void arm_conv_q7(
acc2 = __SMLAD(input1, input2, acc2); acc2 = __SMLAD(input1, input2, acc2);
/* Read x[6] sample */ /* Read x[6] sample */
x2 = *(px++); x2 = *px++;
/* x[5] and x[6] are packed */ /* x[5] and x[6] are packed */
in1 = (q15_t) x1; in1 = (q15_t) x1;
@ -390,10 +390,10 @@ void arm_conv_q7(
while (k > 0U) while (k > 0U)
{ {
/* Read y[srcBLen - 5] sample */ /* Read y[srcBLen - 5] sample */
c0 = *(py--); c0 = *py--;
/* Read x[7] sample */ /* Read x[7] sample */
x3 = *(px++); x3 = *px++;
/* Perform the multiply-accumulates */ /* Perform the multiply-accumulates */
/* acc0 += x[4] * y[srcBLen - 5] */ /* acc0 += x[4] * y[srcBLen - 5] */

Write Preview
Loading…
Cancel
Save