FPGA: Revert "FPGA: Remove db query 9 in 2023.0 (#1257)" (#1281)

This reverts commit 86c4aa8 and
reenables the db9 code sample in 2023.1

Author: justin-rosner

DirectProgramming/C++SYCL_FPGA/ReferenceDesigns/db/README.md

@@ -76,12 +76,18 @@ This design leverages concepts discussed in the [FPGA tutorials](/DirectProgramm
 
 ### Query Implementations
 
-The following sections describe at a high level how queries 1, 11 and 12 are implemented on the FPGA using a set of generalized database operators (found in `db_utils/`). In the block diagrams below, the blocks are oneAPI kernels, and the arrows represent `pipes` that shows the flow of data from one kernel to another.
+The following sections describe at a high level how queries 1, 9, 11 and 12 are implemented on the FPGA using a set of generalized database operators (found in `db_utils/`). In the block diagrams below, the blocks are oneAPI kernels, and the arrows represent `pipes` that shows the flow of data from one kernel to another.
 
 #### Query 1
 
 Query 1 is the simplest of the four queries and only uses the `Accumulator` database operator. The query streams in each row of the LINEITEM table and performs computation on each row.
 
+#### Query 9
+
+Query 9 is the most complicated of the four queries and utilizes all database operators (`LikeRegex`, `Accumulator`, `MapJoin`, `MergeJoin`, `DuplicateMergeJoin`, and `FifoSort`). The block diagram of the design is shown below.
+
+![](assets/q9.png)
+
 #### Query 11
 
 Query 11 showcases the `MapJoin` and `FifoSort` database operators. The block diagram of the design is shown below.
@@ -101,6 +107,8 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
 |`dbdata.cpp`                           | Contains code to parse the database input files and validate the query output
 |`dbdata.hpp`                           | Definitions of database related data structures and parsing functions
 |`query1/query1_kernel.cpp`             | Contains the kernel for Query 1
+|`query9/query9_kernel.cpp`             | Contains the kernel for Query 9
+|`query9/pipe_types.cpp`                | All data types and instantiations for pipes used in query 9
 |`query11/query11_kernel.cpp`           | Contains the kernel for Query 11
 |`query11/pipe_types.cpp`               | All data types and instantiations for pipes used in query 11
 |`query12/query12_kernel.cpp`           | Contains the kernel for Query 12
@@ -142,7 +150,7 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
    cd build
    cmake .. -DQUERY=1
    ```
-   `-DQUERY=<QUERY_NUMBER>` can be any of the following query numbers: `1`, `11` or `12`.
+   `-DQUERY=<QUERY_NUMBER>` can be any of the following query numbers: `1`, `9`, `11` or `12`.
 
 3. Compile the design. (The provided targets match the recommended development flow.)
 
@@ -156,12 +164,14 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
        ```
        The report resides at `db_report.prj/reports/report.html`.
 
+       >**Note**: If you are compiling Query 9 (`-DQUERY=9`), expect a long report generation time. You can download pre-generated reports from [https://iotdk.intel.com/fpga-precompiled-binaries/latest/db.fpga.tar.gz](https://iotdk.intel.com/fpga-precompiled-binaries/latest/db.fpga.tar.gz).
+
     3. Compile for FPGA hardware (longer compile time, targets FPGA device).
 
        ```
        make fpga
        ```
-       When building for hardware, the default scale factor is **1**. To use the smaller scale factor of 0.01, add the flag `-DSF_SMALL=1` to the original `cmake` command. For example: `cmake .. -DQUERY=11 -DSF_SMALL=1`. See the [Database files](#database-files) for more information.
+       When building for hardware, the default scale factor is **1**. To use the smaller scale factor of 0.01, add the flag `-DSF_SMALL=1` to the original `cmake` command. For example: `cmake .. -DQUERY=9 -DSF_SMALL=1`. See the [Database files](#database-files) for more information.
 
    (Optional) The hardware compile may take several hours to complete. You can download a pre-compiled binary (compatible with Linux* Ubuntu* 18.04) for an Intel® FPGA PAC D5005 (with Intel Stratix® 10 SX) from [https://iotdk.intel.com/fpga-precompiled-binaries/latest/db.fpga.tar.gz](https://iotdk.intel.com/fpga-precompiled-binaries/latest/db.fpga.tar.gz).
 
@@ -176,7 +186,7 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
    cd build
    cmake -G "NMake Makefiles" -DQUERY=1
    ```
-   `-DQUERY=<QUERY_NUMBER>` can be any of the following query numbers: `1`, `11` or `12`.
+   `-DQUERY=<QUERY_NUMBER>` can be any of the following query numbers: `1`, `9`, `11` or `12`.
 
 3. Compile the design. (The provided targets match the recommended development flow.)
 
@@ -191,6 +201,8 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
        ```
        The report resides at `db_report.prj/reports/report.html` directory.
 
+       >**Note**: If you are compiling Query 9 (`-DQUERY=9`), expect a long report generation time.
+
     3. Compile for FPGA hardware (longer compile time, targets FPGA device):
        ```
        nmake fpga
@@ -216,7 +228,7 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
     ```
     ./db.fpga_emu --dbroot=../data/sf0.01 --test
     ```
-    (Optional) Run the design for queries `11` and `12`.
+    (Optional) Run the design for queries `9`, `11` and `12`.
 
 2. Run the design on an FPGA device.
    ```
@@ -229,7 +241,7 @@ Query 12 showcases the `MergeJoin` database operator. The block diagram of the d
      ```
      db.fpga_emu.exe --dbroot=../data/sf0.01 --test
      ```
-    (Optional) Run the design for queries `11` and `12`.
+    (Optional) Run the design for queries `9`, `11` and `12`.
 
 2. Run the sample on an FPGA device.
    ```

DirectProgramming/C++SYCL_FPGA/ReferenceDesigns/db/assets/q9.png

@@ -30,6 +30,17 @@
           "./db.fpga_emu --dbroot=../data/sf0.01 --test"
         ]
       },
+      {
+        "id": "fpga_emu_q9",
+        "steps": [
+          "icpx --version",
+          "mkdir build-q9",
+          "cd build-q9",
+          "cmake .. -DQUERY=9",
+          "make fpga_emu",
+          "./db.fpga_emu --dbroot=../data/sf0.01 --test"
+        ]
+      },
       {
         "id": "fpga_emu_q11",
         "steps": [
@@ -97,6 +108,19 @@
           "db.fpga_emu.exe --dbroot=../data/sf0.01 --test"
         ]
       },
+      {
+        "id": "fpga_emu_q9",
+        "steps": [
+          "icpx --version",
+          "cd ../..",
+          "mkdir build-q9",
+          "cd build-q9",
+          "xcopy /E ..\\ReferenceDesigns\\db\\data ..\\data\\",
+          "cmake -G \"NMake Makefiles\" ../ReferenceDesigns/db -DQUERY=9",
+          "nmake fpga_emu",
+          "db.fpga_emu.exe --dbroot=../data/sf0.01 --test"
+        ]
+      },
       {
         "id": "fpga_emu_q11",
         "steps": [

DirectProgramming/C++SYCL_FPGA/ReferenceDesigns/db/sample.json

@@ -15,6 +15,9 @@ endif()
 if(${QUERY} EQUAL 1)
     set(DEFAULT_BOARD "intel_a10gx_pac:pac_a10")
     set(DEFAULT_BOARD_STR "Intel Arria(R) 10 GX")
+elseif(${QUERY} EQUAL 9)
+    set(DEFAULT_BOARD "intel_s10sx_pac:pac_s10")
+    set(DEFAULT_BOARD_STR "Intel Stratix(R) 10 SX")
 elseif(${QUERY} EQUAL 11)
     set(DEFAULT_BOARD "intel_s10sx_pac:pac_s10")
     set(DEFAULT_BOARD_STR "Intel Stratix(R) 10 SX")
@@ -42,8 +45,8 @@ else()
 endif()
 
 # ensure a supported query was requested
-if(NOT ${QUERY} EQUAL 1 AND NOT ${QUERY} EQUAL 11 AND NOT ${QUERY} EQUAL 12)
-  message(FATAL_ERROR "\tQUERY ${QUERY} not supported (supported queries are 1, 11 and 12)")
+if(NOT ${QUERY} EQUAL 1 AND NOT ${QUERY} EQUAL 9 AND NOT ${QUERY} EQUAL 11 AND NOT ${QUERY} EQUAL 12)
+  message(FATAL_ERROR "\tQUERY ${QUERY} not supported (supported queries are 1, 9, 11 and 12)")
 endif()
 
 # Pick the default seed if the user did not specify one to CMake.
@@ -52,6 +55,8 @@ if(NOT DEFINED SEED)
     if(${FPGA_DEVICE} MATCHES ".*a10.*")
         if(${QUERY} EQUAL 1)
             set(SEED "-Xsseed=2")
+        elseif(${QUERY} EQUAL 9)
+            set(SEED "-Xsseed=2")
         elseif(${QUERY} EQUAL 11)
             set(SEED "-Xsseed=4")
         elseif(${QUERY} EQUAL 12)
@@ -60,6 +65,8 @@ if(NOT DEFINED SEED)
     elseif(${FPGA_DEVICE} MATCHES ".*s10.*")
         if(${QUERY} EQUAL 1)
             set(SEED "-Xsseed=3")
+        elseif(${QUERY} EQUAL 9)
+            set(SEED "-Xsseed=2")
         elseif(${QUERY} EQUAL 11)
             set(SEED "-Xsseed=3")
         elseif(${QUERY} EQUAL 12)
@@ -68,6 +75,8 @@ if(NOT DEFINED SEED)
     elseif(${FPGA_DEVICE} MATCHES ".*agilex.*")
         if(${QUERY} EQUAL 1)
             set(SEED "-Xsseed=2")
+        elseif(${QUERY} EQUAL 9)
+            set(SEED "-Xsseed=2")
         elseif(${QUERY} EQUAL 11)
             set(SEED "-Xsseed=4")
         elseif(${QUERY} EQUAL 12)
@@ -82,10 +91,10 @@ if(IGNORE_DEFAULT_SEED)
     set(SEED "")
 endif()
 
-# Error out if trying to run Q11 on Arria 10
+# Error out if trying to run Q9 or Q11 on Arria 10
 if (${FPGA_DEVICE} MATCHES ".*a10.*")
-    if(${QUERY} EQUAL 11)
-      message(FATAL_ERROR "Query 11 is not supported on Arria 10 devices")
+    if(${QUERY} EQUAL 9 OR ${QUERY} EQUAL 11)
+      message(FATAL_ERROR "Queries 9 and 11 are not supported on Arria 10 devices")
     endif()
 endif()
 
@@ -106,14 +115,17 @@ endif()
 if(${QUERY} EQUAL 1)
     set(DEVICE_SOURCE query1/query1_kernel.cpp)
     set(DEVICE_HEADER query1/query1_kernel.hpp)
+elseif(${QUERY} EQUAL 9)
+    set(DEVICE_SOURCE query9/query9_kernel.cpp)
+    set(DEVICE_HEADER query9/query9_kernel.hpp)
 elseif(${QUERY} EQUAL 11)
     set(DEVICE_SOURCE query11/query11_kernel.cpp)
     set(DEVICE_HEADER query11/query11_kernel.hpp)
 elseif(${QUERY} EQUAL 12)
     set(DEVICE_SOURCE query12/query12_kernel.cpp)
     set(DEVICE_HEADER query12/query12_kernel.hpp)
 else()
-    message(FATAL_ERROR "\tQUERY ${QUERY} not supported (supported queries are 1, 11 and 12)")
+    message(FATAL_ERROR "\tQUERY ${QUERY} not supported (supported queries are 1, 9, 11 and 12)")
 endif()
 
 # A SYCL ahead-of-time (AoT) compile processes the device code in two stages.

DirectProgramming/C++SYCL_FPGA/ReferenceDesigns/db/src/CMakeLists.txt

@@ -42,6 +42,11 @@ using namespace sycl;
 bool DoQuery1(queue& q, Database& dbinfo, std::string& db_root_dir,
               std::string& args, bool test, bool print, double& kernel_latency,
               double& total_latency);
+#elif (QUERY == 9)
+#include "query9/query9_kernel.hpp"
+bool DoQuery9(queue& q, Database& dbinfo, std::string& db_root_dir,
+              std::string& args, bool test, bool print, double& kernel_latency,
+              double& total_latency);
 #elif (QUERY == 11)
 #include "query11/query11_kernel.hpp"
 bool DoQuery11(queue& q, Database& dbinfo, std::string& db_root_dir,
@@ -168,9 +173,9 @@ int main(int argc, char* argv[]) {
   }
 
   // make sure the query is supported
-  if (!(query == 1 || query == 11 || query == 12)) {
+  if (!(query == 1 || query == 9 || query == 11 || query == 12)) {
     std::cerr << "ERROR: unsupported query (" << query << "). "
-              << "Only queries 1, 11 and 12 are supported\n";
+              << "Only queries 1, 9, 11 and 12 are supported\n";
     return 1;
   }
 
@@ -224,6 +229,13 @@ int main(int argc, char* argv[]) {
         success = DoQuery1(q, dbinfo, db_root_dir, args,
                            test_query, print_result,
                            kernel_latency[run], total_latency[run]);
+#endif
+      } else if (query == 9) {
+        // query9
+#if (QUERY == 9)
+        success = DoQuery9(q, dbinfo, db_root_dir, args,
+                           test_query, print_result,
+                           kernel_latency[run], total_latency[run]);
 #endif
       } else if (query == 11) {
         // query11
@@ -351,6 +363,51 @@ bool DoQuery1(queue& q, Database& dbinfo, std::string& db_root_dir,
 }
 #endif
 
+#if (QUERY == 9)
+bool DoQuery9(queue& q, Database& dbinfo, std::string& db_root_dir,
+              std::string& args, bool test, bool print, double& kernel_latency,
+              double& total_latency) {
+  // the default colour regex based on the TPCH documents
+  std::string colour = "GREEN";
+
+  // parse the query arguments
+  if (!test && !args.empty()) {
+    std::stringstream ss(args);
+    std::getline(ss, colour, ',');
+  } else {
+    if (!args.empty()) {
+      std::cout << "Testing query 9, therefore ignoring the '--args' flag\n";
+    }
+  }
+
+  // convert the colour regex to uppercase characters (convention)
+  transform(colour.begin(), colour.end(), colour.begin(), ::toupper);
+
+  std::cout << "Running Q9 with colour regex: " << colour << std::endl;
+
+  // the output of the query
+  std::array<DBDecimal, 25 * 2020> sum_profit;
+
+  // perform the query
+  bool success = SubmitQuery9(q, dbinfo, colour, sum_profit, kernel_latency,
+                              total_latency);
+
+  if (success) {
+    // validate the results of the query, if requested
+    if (test) {
+      success = dbinfo.ValidateQ9(db_root_dir, sum_profit);
+    }
+
+    // print the results of the query, if requested
+    if (print) {
+      dbinfo.PrintQ9(sum_profit);
+    }
+  }
+
+  return success;
+}
+#endif
+
 #if (QUERY == 11)
 bool DoQuery11(queue& q, Database& dbinfo, std::string& db_root_dir,
                std::string& args, bool test, bool print, double& kernel_latency,