faiss 0.3.1 → 0.3.3

data/vendor/faiss/faiss/impl/index_write.cpp

@@ -1,12 +1,10 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
  */
 
-// -*- c++ -*-
-
 #include <faiss/index_io.h>
 
 #include <faiss/impl/io.h>
@@ -15,13 +13,9 @@
 #include <cstdio>
 #include <cstdlib>
 
-#include <sys/stat.h>
-#include <sys/types.h>
-
 #include <faiss/invlists/InvertedListsIOHook.h>
 
 #include <faiss/impl/FaissAssert.h>
-#include <faiss/impl/io.h>
 #include <faiss/impl/io_macros.h>
 #include <faiss/utils/hamming.h>
 
@@ -314,7 +308,11 @@ static void write_HNSW(const HNSW* hnsw, IOWriter* f) {
     WRITE1(hnsw->max_level);
     WRITE1(hnsw->efConstruction);
     WRITE1(hnsw->efSearch);
-    WRITE1(hnsw->upper_beam);
+
+    // // deprecated field
+    // WRITE1(hnsw->upper_beam);
+    constexpr int tmp_upper_beam = 1;
+    WRITE1(tmp_upper_beam);
 }
 
 static void write_NSG(const NSG* nsg, IOWriter* f) {
@@ -390,8 +388,12 @@ static void write_ivf_header(const IndexIVF* ivf, IOWriter* f) {
     write_direct_map(&ivf->direct_map, f);
 }
 
-void write_index(const Index* idx, IOWriter* f) {
-    if (const IndexFlat* idxf = dynamic_cast<const IndexFlat*>(idx)) {
+void write_index(const Index* idx, IOWriter* f, int io_flags) {
+    if (idx == nullptr) {
+        // eg. for a storage component of HNSW that is set to nullptr
+        uint32_t h = fourcc("null");
+        WRITE1(h);
+    } else if (const IndexFlat* idxf = dynamic_cast<const IndexFlat*>(idx)) {
         uint32_t h =
                 fourcc(idxf->metric_type == METRIC_INNER_PRODUCT ? "IxFI"
                                : idxf->metric_type == METRIC_L2  ? "IxF2"
@@ -760,12 +762,24 @@ void write_index(const Index* idx, IOWriter* f) {
                 : dynamic_cast<const IndexHNSWPQ*>(idx)      ? fourcc("IHNp")
                 : dynamic_cast<const IndexHNSWSQ*>(idx)      ? fourcc("IHNs")
                 : dynamic_cast<const IndexHNSW2Level*>(idx)  ? fourcc("IHN2")
+                : dynamic_cast<const IndexHNSWCagra*>(idx)   ? fourcc("IHNc")
                                                              : 0;
         FAISS_THROW_IF_NOT(h != 0);
         WRITE1(h);
         write_index_header(idxhnsw, f);
+        if (h == fourcc("IHNc")) {
+            WRITE1(idxhnsw->keep_max_size_level0);
+            auto idx_hnsw_cagra = dynamic_cast<const IndexHNSWCagra*>(idxhnsw);
+            WRITE1(idx_hnsw_cagra->base_level_only);
+            WRITE1(idx_hnsw_cagra->num_base_level_search_entrypoints);
+        }
         write_HNSW(&idxhnsw->hnsw, f);
-        write_index(idxhnsw->storage, f);
+        if (io_flags & IO_FLAG_SKIP_STORAGE) {
+            uint32_t n4 = fourcc("null");
+            WRITE1(n4);
+        } else {
+            write_index(idxhnsw->storage, f);
+        }
     } else if (const IndexNSG* idxnsg = dynamic_cast<const IndexNSG*>(idx)) {
         uint32_t h = dynamic_cast<const IndexNSGFlat*>(idx) ? fourcc("INSf")
                 : dynamic_cast<const IndexNSGPQ*>(idx)      ? fourcc("INSp")
@@ -841,14 +855,14 @@ void write_index(const Index* idx, IOWriter* f) {
     }
 }
 
-void write_index(const Index* idx, FILE* f) {
+void write_index(const Index* idx, FILE* f, int io_flags) {
     FileIOWriter writer(f);
-    write_index(idx, &writer);
+    write_index(idx, &writer, io_flags);
 }
 
-void write_index(const Index* idx, const char* fname) {
+void write_index(const Index* idx, const char* fname, int io_flags) {
     FileIOWriter writer(fname);
-    write_index(idx, &writer);
+    write_index(idx, &writer, io_flags);
 }
 
 void write_VectorTransform(const VectorTransform* vt, const char* fname) {

data/vendor/faiss/faiss/impl/io.cpp

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -20,11 +20,11 @@ namespace faiss {
  * IO functions
  ***********************************************************************/
 
-int IOReader::fileno() {
+int IOReader::filedescriptor() {
     FAISS_THROW_MSG("IOReader does not support memory mapping");
 }
 
-int IOWriter::fileno() {
+int IOWriter::filedescriptor() {
     FAISS_THROW_MSG("IOWriter does not support memory mapping");
 }
 
@@ -85,8 +85,12 @@ size_t FileIOReader::operator()(void* ptr, size_t size, size_t nitems) {
     return fread(ptr, size, nitems, f);
 }
 
-int FileIOReader::fileno() {
+int FileIOReader::filedescriptor() {
+#ifdef _AIX
+    return fileno(f);
+#else
     return ::fileno(f);
+#endif
 }
 
 FileIOWriter::FileIOWriter(FILE* wf) : f(wf) {}
@@ -116,8 +120,12 @@ size_t FileIOWriter::operator()(const void* ptr, size_t size, size_t nitems) {
     return fwrite(ptr, size, nitems, f);
 }
 
-int FileIOWriter::fileno() {
+int FileIOWriter::filedescriptor() {
+#ifdef _AIX
+    return fileno(f);
+#else
     return ::fileno(f);
+#endif
 }
 
 /***********************************************************************
@@ -259,7 +267,7 @@ std::string fourcc_inv_printable(uint32_t x) {
             str += c;
         } else {
             char buf[10];
-            sprintf(buf, "\\x%02x", c);
+            snprintf(buf, sizeof(buf), "\\x%02x", c);
             str += buf;
         }
     }

data/vendor/faiss/faiss/impl/io.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -32,7 +32,7 @@ struct IOReader {
     virtual size_t operator()(void* ptr, size_t size, size_t nitems) = 0;
 
     // return a file number that can be memory-mapped
-    virtual int fileno();
+    virtual int filedescriptor();
 
     virtual ~IOReader() {}
 };
@@ -45,7 +45,7 @@ struct IOWriter {
     virtual size_t operator()(const void* ptr, size_t size, size_t nitems) = 0;
 
     // return a file number that can be memory-mapped
-    virtual int fileno();
+    virtual int filedescriptor();
 
     virtual ~IOWriter() noexcept(false) {}
 };
@@ -73,7 +73,7 @@ struct FileIOReader : IOReader {
 
     size_t operator()(void* ptr, size_t size, size_t nitems) override;
 
-    int fileno() override;
+    int filedescriptor() override;
 };
 
 struct FileIOWriter : IOWriter {
@@ -88,7 +88,7 @@ struct FileIOWriter : IOWriter {
 
     size_t operator()(const void* ptr, size_t size, size_t nitems) override;
 
-    int fileno() override;
+    int filedescriptor() override;
 };
 
 /*******************************************************

data/vendor/faiss/faiss/impl/io_macros.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -29,6 +29,12 @@
 
 #define READ1(x) READANDCHECK(&(x), 1)
 
+#define READ1_DUMMY(x_type) \
+    {                       \
+        x_type x = {};      \
+        READ1(x);           \
+    }
+
 // will fail if we write 256G of data at once...
 #define READVECTOR(vec)                                              \
     {                                                                \
@@ -39,13 +45,6 @@
         READANDCHECK((vec).data(), size);                            \
     }
 
-#define READSTRING(s)                     \
-    {                                     \
-        size_t size = (s).size();         \
-        WRITEANDCHECK(&size, 1);          \
-        WRITEANDCHECK((s).c_str(), size); \
-    }
-
 #define WRITEANDCHECK(ptr, n)                         \
     {                                                 \
         size_t ret = (*f)(ptr, sizeof(*(ptr)), n);    \

data/vendor/faiss/faiss/impl/kmeans1d.cpp

@@ -1,12 +1,11 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
  */
 
 #include <algorithm>
-#include <cstdint>
 #include <cstring>
 #include <functional>
 #include <numeric>

data/vendor/faiss/faiss/impl/kmeans1d.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.

data/vendor/faiss/faiss/impl/lattice_Zn.cpp

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -16,7 +16,6 @@
 
 #include <algorithm>
 #include <queue>
-#include <unordered_map>
 #include <unordered_set>
 
 #include <faiss/impl/platform_macros.h>

data/vendor/faiss/faiss/impl/lattice_Zn.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.

data/vendor/faiss/faiss/impl/platform_macros.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -39,6 +39,7 @@
 // redefine the GCC intrinsics with Windows equivalents
 
 #include <intrin.h>
+#include <limits.h>
 
 #ifndef __clang__
 inline int __builtin_ctzll(uint64_t x) {
@@ -59,7 +60,16 @@ inline int __builtin_ctz(unsigned long x) {
 
 #ifndef __clang__
 inline int __builtin_clzll(uint64_t x) {
+#if defined(_M_X64) || defined(__x86_64__)
     return (int)__lzcnt64(x);
+#elif defined(_M_ARM64)
+    unsigned long index;
+    int count = sizeof(uint64_t) * CHAR_BIT;
+    if (_BitScanReverse64(&index, x)) {
+        count = count - 1 - index;
+    }
+    return count;
+#endif
 }
 #endif
 
@@ -127,6 +137,13 @@ inline int __builtin_clzll(uint64_t x) {
     __pragma(float_control(precise, off, push))
 #define FAISS_PRAGMA_IMPRECISE_FUNCTION_END __pragma(float_control(pop))
 #elif defined(__clang__)
+#if defined(__PPC__)
+#define FAISS_PRAGMA_IMPRECISE_LOOP \
+    _Pragma("clang loop vectorize_width(4) interleave_count(8)")
+#define FAISS_PRAGMA_IMPRECISE_FUNCTION_BEGIN \
+    _Pragma("float_control(precise, off, push)")
+#define FAISS_PRAGMA_IMPRECISE_FUNCTION_END _Pragma("float_control(pop)")
+#else
 #define FAISS_PRAGMA_IMPRECISE_LOOP \
     _Pragma("clang loop vectorize(enable) interleave(enable)")
 
@@ -144,6 +161,7 @@ inline int __builtin_clzll(uint64_t x) {
 #define FAISS_PRAGMA_IMPRECISE_FUNCTION_BEGIN
 #define FAISS_PRAGMA_IMPRECISE_FUNCTION_END
 #endif
+#endif
 #elif defined(__GNUC__)
 // Unfortunately, GCC does not provide a pragma for detecting it.
 // So, we have to stick to GNUC, which is defined by MANY compilers.
@@ -165,3 +183,17 @@ inline int __builtin_clzll(uint64_t x) {
 #endif
 
 // clang-format on
+
+/*******************************************************
+ * BIGENDIAN specific macros
+ *******************************************************/
+#if !defined(_MSC_VER) && \
+        (defined(__BYTE_ORDER__) && (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__))
+#define FAISS_BIG_ENDIAN
+#endif
+
+#define Swap2Bytes(val) ((((val) >> 8) & 0x00FF) | (((val) << 8) & 0xFF00))
+
+#define Swap4Bytes(val)                                           \
+    ((((val) >> 24) & 0x000000FF) | (((val) >> 8) & 0x0000FF00) | \
+     (((val) << 8) & 0x00FF0000) | (((val) << 24) & 0xFF000000))

data/vendor/faiss/faiss/impl/pq4_fast_scan.cpp

@@ -1,11 +1,12 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
  */
 
 #include <faiss/impl/FaissAssert.h>
+#include <faiss/impl/platform_macros.h>
 #include <faiss/impl/pq4_fast_scan.h>
 #include <faiss/impl/simd_result_handlers.h>
 
@@ -58,8 +59,13 @@ void pq4_pack_codes(
         return;
     }
     memset(blocks, 0, nb * nsq / 2);
+#ifdef FAISS_BIG_ENDIAN
+    const uint8_t perm0[16] = {
+            8, 0, 9, 1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7};
+#else
     const uint8_t perm0[16] = {
             0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15};
+#endif
 
     uint8_t* codes2 = blocks;
     for (size_t i0 = 0; i0 < nb; i0 += bbs) {
@@ -93,8 +99,13 @@ void pq4_pack_codes_range(
         size_t bbs,
         size_t nsq,
         uint8_t* blocks) {
+#ifdef FAISS_BIG_ENDIAN
+    const uint8_t perm0[16] = {
+            8, 0, 9, 1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7};
+#else
     const uint8_t perm0[16] = {
             0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15};
+#endif
 
     // range of affected blocks
     size_t block0 = i0 / bbs;

data/vendor/faiss/faiss/impl/pq4_fast_scan.h

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -189,4 +189,22 @@ void pq4_accumulate_loop_qbs(
         SIMDResultHandler& res,
         const NormTableScaler* scaler = nullptr);
 
+/** Wrapper of pq4_accumulate_loop_qbs using simple StoreResultHandler
+ *  and DummyScaler
+ *
+ * @param nq      number of queries
+ * @param ntotal2 number of database elements (multiple of 32)
+ * @param nsq     number of sub-quantizers (muliple of 2)
+ * @param codes   packed codes array
+ * @param LUT     packed look-up table
+ * @param accu    array to store the results
+ */
+void accumulate_to_mem(
+        int nq,
+        size_t ntotal2,
+        int nsq,
+        const uint8_t* codes,
+        const uint8_t* LUT,
+        uint16_t* accu);
+
 } // namespace faiss

data/vendor/faiss/faiss/impl/pq4_fast_scan_search_1.cpp

@@ -1,5 +1,5 @@
-/**
- * Copyright (c) Facebook, Inc. and its affiliates.
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  *
  * This source code is licensed under the MIT license found in the
  * LICENSE file in the root directory of this source tree.
@@ -217,7 +217,7 @@ void pq4_accumulate_loop(
         SIMDResultHandler& res,
         const NormTableScaler* scaler) {
     Run_pq4_accumulate_loop consumer;
-    dispatch_SIMDResultHanlder(
+    dispatch_SIMDResultHandler(
             res, consumer, nq, nb, bbs, nsq, codes, LUT, scaler);
 }