comparison projects/heavy/envelopeTrigger/SignalTabread.h @ 162:c3e8226a5651 heavy-updated

- added additional flags to C rules (-DNDEBUG, -mfpu=neon)
- sample-accurate envelope triggering pd/heavy example

+/**
+ * Copyright (c) 2014, 2015, Enzien Audio Ltd.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH
+ * REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
+ * AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
+ * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+ * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+ * OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+ * PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _HEAVY_SIGNAL_TABREAD_H_
+#define _HEAVY_SIGNAL_TABREAD_H_
+
+#include "HvBase.h"
+#include "HvTable.h"
+
+typedef struct SignalTabread {
+  HvTable *table; // the table to read
+  hv_uint32_t head;
+  bool forceAlignedLoads; // false by default, true if using __hv_tabread_f
+} SignalTabread;
+
+// random access to a table
+hv_size_t sTabread_init(SignalTabread *o, HvTable *table, bool forceAlignedLoads);
+
+
+
+#if HV_APPLE
+#pragma mark - Tabread - Random Access
+#endif
+
+static inline void __hv_tabread_if(SignalTabread *o, hv_bIni_t bIn, hv_bOutf_t bOut) {
+  const float *const b = hTable_getBuffer(o->table);
+#if HV_SIMD_AVX
+  hv_assert((int) (bIn[0] & 0xFFFFFFFFL) >= 0 && (int) (bIn[0] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[0] >> 32) >= 0 && (int) ((bIn[0] & ~0xFFFFFFFFL) >> 32) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[1] & 0xFFFFFFFFL) >= 0 && (int) (bIn[1] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[1] >> 32) >= 0 && (int) ((bIn[1] & ~0xFFFFFFFFL) >> 32) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[2] & 0xFFFFFFFFL) >= 0 && (int) (bIn[2] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[2] >> 32) >= 0 && (int) ((bIn[2] & ~0xFFFFFFFFL) >> 32) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[3] & 0xFFFFFFFFL) >= 0 && (int) (bIn[3] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[3] >> 32) >= 0 && (int) ((bIn[3] & ~0xFFFFFFFFL) >> 32) < hTable_getAllocated(o->table));
+
+  *bOut = _mm256_set_ps(
+      b[(int) (bIn[3] >> 32)],
+      b[(int) (bIn[3] & 0xFFFFFFFFL)],
+      b[(int) (bIn[2] >> 32)],
+      b[(int) (bIn[2] & 0xFFFFFFFFL)],
+      b[(int) (bIn[1] >> 32)],
+      b[(int) (bIn[1] & 0xFFFFFFFFL)],
+      b[(int) (bIn[0] >> 32)],
+      b[(int) (bIn[0] & 0xFFFFFFFFL)]);
+#elif HV_SIMD_SSE
+  hv_assert((int) (bIn[0] & 0xFFFFFFFFL) >= 0 && (int) (bIn[0] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[0] >> 32) >= 0 && (int) (bIn[0] >> 32) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[1] & 0xFFFFFFFFL) >= 0 && (int) (bIn[1] & 0xFFFFFFFFL) < hTable_getAllocated(o->table));
+  hv_assert((int) (bIn[1] >> 32) >= 0 && (int) (bIn[1] >> 32) < hTable_getAllocated(o->table));
+
+  *bOut = _mm_set_ps(
+      b[(int) (bIn[1] >> 32)],
+      b[(int) (bIn[1] & 0xFFFFFFFFL)],
+      b[(int) (bIn[0] >> 32)],
+      b[(int) (bIn[0] & 0xFFFFFFFFL)]);
+#elif HV_SIMD_NEON
+  hv_assert((bIn[0] >= 0) && (bIn[0] < hTable_getAllocated(o->table)));
+  hv_assert((bIn[1] >= 0) && (bIn[1] < hTable_getAllocated(o->table)));
+  hv_assert((bIn[2] >= 0) && (bIn[2] < hTable_getAllocated(o->table)));
+  hv_assert((bIn[3] >= 0) && (bIn[3] < hTable_getAllocated(o->table)));
+
+  *bOut = (float32x4_t) {b[bIn[0]], b[bIn[1]], b[bIn[2]], b[bIn[3]]};
+#else // HV_SIMD_NONE
+  hv_assert(bIn >= 0 && ((hv_uint32_t) bIn < hTable_getAllocated(o->table)));
+
+  *bOut = b[bIn];
+#endif
+}
+
+
+
+#if HV_APPLE
+#pragma mark - Tabread - Linear Access
+#endif
+
+// this tabread never stops reading. It is mainly intended for linear reads that loop around a table.
+static inline void __hv_tabread_f(SignalTabread *o, hv_bOutf_t bOut) {
+  hv_assert((o->head + HV_N_SIMD) <= hTable_getAllocated(o->table)); // assert that we always read within the table bounds
+  hv_uint32_t head = o->head;
+#if HV_SIMD_AVX
+  *bOut = _mm256_load_ps(hTable_getBuffer(o->table) + head);
+#elif HV_SIMD_SSE
+  *bOut = _mm_load_ps(hTable_getBuffer(o->table) + head);
+#elif HV_SIMD_NEON
+  *bOut = vld1q_f32(hTable_getBuffer(o->table) + head);
+#else // HV_SIMD_NONE
+  *bOut = *(hTable_getBuffer(o->table) + head);
+#endif
+  o->head = head + HV_N_SIMD;
+}
+
+// unaligned linear tabread, as above
+static inline void __hv_tabreadu_f(SignalTabread *o, hv_bOutf_t bOut) {
+  hv_assert((o->head + HV_N_SIMD) <= hTable_getAllocated(o->table)); // assert that we always read within the table bounds
+  hv_uint32_t head = o->head;
+#if HV_SIMD_AVX
+  *bOut = _mm256_loadu_ps(hTable_getBuffer(o->table) + head);
+#elif HV_SIMD_SSE
+  *bOut = _mm_loadu_ps(hTable_getBuffer(o->table) + head);
+#elif HV_SIMD_NEON
+  *bOut = vld1q_f32(hTable_getBuffer(o->table) + head);
+#else // HV_SIMD_NONE
+  *bOut = *(hTable_getBuffer(o->table) + head);
+#endif
+  o->head = head + HV_N_SIMD;
+}
+
+// this tabread can be instructed to stop. It is mainly intended for linear reads that only process a portion of a buffer.
+static inline void __hv_tabread_stoppable_f(SignalTabread *o, hv_bOutf_t bOut) {
+#if HV_SIMD_AVX
+  if (o->head == ~0x0) {
+    *bOut = _mm256_setzero_ps();
+  } else {
+    *bOut = _mm256_load_ps(hTable_getBuffer(o->table) + o->head);
+    o->head += HV_N_SIMD;
+  }
+#elif HV_SIMD_SSE
+  if (o->head == ~0x0) {
+    *bOut = _mm_setzero_ps();
+  } else {
+    *bOut = _mm_load_ps(hTable_getBuffer(o->table) + o->head);
+    o->head += HV_N_SIMD;
+  }
+#elif HV_SIMD_NEON
+  if (o->head == ~0x0) {
+    *bOut = vdupq_n_f32(0.0f);
+  } else {
+    *bOut = vld1q_f32(hTable_getBuffer(o->table) + o->head);
+    o->head += HV_N_SIMD;
+  }
+#else // HV_SIMD_NONE
+  if (o->head == ~0x0) {
+    *bOut = 0.0f;
+  } else {
+    *bOut = *(hTable_getBuffer(o->table) + o->head);
+    o->head += HV_N_SIMD;
+  }
+#endif
+}
+
+void sTabread_onMessage(HvBase *_c, SignalTabread *o, int letIn, const HvMessage *const m);
+
+
+
+#if HV_APPLE
+#pragma mark - Tabhead
+#endif
+
+typedef struct SignalTabhead {
+  HvTable *table;
+} SignalTabhead;
+
+hv_size_t sTabhead_init(SignalTabhead *o, HvTable *table);
+
+static inline void __hv_tabhead_f(SignalTabhead *o, hv_bOutf_t bOut) {
+#if HV_SIMD_AVX
+  *bOut = _mm256_set1_ps((float) hTable_getHead(o->table));
+#elif HV_SIMD_SSE
+  *bOut = _mm_set1_ps((float) hTable_getHead(o->table));
+#elif HV_SIMD_NEON
+  *bOut = vdupq_n_f32((float32_t) hTable_getHead(o->table));
+#else // HV_SIMD_NONE
+  *bOut = (float) hTable_getHead(o->table);
+#endif
+}
+
+void sTabhead_onMessage(HvBase *_c, SignalTabhead *o, const HvMessage *const m);
+
+#endif // _HEAVY_SIGNAL_TABREAD_H_