SWDEV-477244: Select() Expression Dimension Support (#1091)

* add support for select function in derived counters * formatting * renaming select dims variable name from set to map * format * Update doc with select() for dimensions * use : for defining range of values in select dims * - update dimension for metric after select. - make sure to raise runtime error if user provides range for a dimension. * use map instead of unordered_map for select dim info * new line EOF * fix bug: select() operator. * Update evaluate_ast.cpp format * added a check for dim value exceeds max. * Update source/lib/rocprofiler-sdk/counters/evaluate_ast.cpp Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * Update source/lib/rocprofiler-sdk/counters/evaluate_ast.cpp Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * updated doc with data example for select operation. * changelog.md * Update CHANGELOG.md --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> [ROCm/rocprofiler-sdk commit: cc4811d27d]
2024-11-11 03:44:22 -06:00
@@ -136,3 +136,14 @@ Full documentation for ROCprofiler-SDK is available at [rocm.docs.amd.com/projec

 - Removed gfx8 metric definitions
 - Removed `rocprofv3` installation to sbin directory
+
+## ROCprofiler-SDK 0.5.0 for ROCm release 6.4
+
+### Added
+
+- Added support for select() operation in counter expression.
+### Changed
+
+### Resolved issues
+
+### Removed
@@ -290,6 +290,51 @@ Expression: 100*reduce(GL2C_HIT,sum)/(reduce(GL2C_HIT,sum)+reduce(GL2C_MISS,sum)
 The reduce function reduces counter values across all dimensions such as shader engine, SIMD, and so on, to produce a single output value. This helps to collect and compare values across the entire device.
 Here are the common reduction operations:

+### Select Function
+
+```yaml
+expression: select(Y, [DIMENSION_XCC=[0],DIMENSION_SHADER_ENGINE=[2]])
+```
+
+select() only returns counter values which match the dimension indexes provided by the user in expression. This operation is to allow a user to state they only want to select specific dimensions index. Supported dimensions include ```DIMENSION_XCC, DIMENSION_AID, DIMENSION_SHADER_ENGINE, DIMENSION_AGENT, DIMENSION_SHADER_ARRAY, DIMENSION_WGP, DIMENSION_INSTANCE```. For example ``select(Y, [DIMENSION_XCC=[0],DIMENSION_SHADER_ENGINE=[2]])`` gives counter values which are from DIMENSION_XCC= 0 and DIMENSION_SHADER_ENGINE= 2 for Y Metric.
+
+Let's say Y has XCC, SHADER_ENGINE(SE), WGP dimensions with sizes 2, 4, 4 respectively.
+
+Raw Counter Data in 3D space:
+
+#### XCC[0]:
+|       |WGP[0]|WGP[1]|WGP[2]|WGP[3]|
+|-------|------|------|------|------|
+| SE[0] |   1  |   2  |   3  |   4  |
+| SE[1] |   5  |   6  |   7  |   8  |
+| SE[2] |   9  |   10 |   11 |   12 |
+| SE[3] |   13 |   14 |   15 |   16 |
+
+#### XCC[1]:
+|       |WGP[0]|WGP[1]|WGP[2]|WGP[3]|
+|-------|------|------|------|------|
+| SE[0] |   17 |   18 |   19 |   20 |
+| SE[1] |   21 |   22 |   23 |   24 |
+| SE[2] |   25 |   26 |   27 |   28 |
+| SE[3] |   29 |   30 |   31 |   32 |
+
+Selecting at XCC=0 results to 2D space with WGP and SH dimensions, as shown below.
+
+|       |WGP[0]|WGP[1]|WGP[2]|WGP[3]|
+|-------|------|------|------|------|
+| SE[0] |   1  |   2  |   3  |   4  |
+| SE[1] |   5  |   6  |   7  |   8  |
+| SE[2] |   9  |   10 |   11 |   12 |
+| SE[3] |   13 |   14 |   15 |   16 |
+
+similarly, for `select(Y, [DIMENSION_XCC=[0],DIMENSION_SHADER_ENGINE=[2]])` results in only WGP dimension with XCC=0 and SE=2.
+
+|       |WGP[0]|WGP[1]|WGP[2]|WGP[3]|
+|-------|------|------|------|------|
+|       |  9   |  10  |  11  |  12  |
+
+### Accumulate Function
+
 - `sum`: Sums to create a single output. For example, `reduce(GL2C_HIT,sum)` sums all `GL2C_HIT` hardware register values.
 - `avr`: Calculates the average across all dimensions.
 - `min`: Selects minimum value across all dimensions.
@@ -22,10 +22,14 @@

 #include "lib/rocprofiler-sdk/counters/evaluate_ast.hpp"

+#include <algorithm>
+#include <cstdint>
 #include <exception>
 #include <numeric>
 #include <optional>
 #include <stdexcept>
+#include <unordered_map>
+#include <vector>

 #include <fmt/core.h>
 #include <fmt/ranges.h>
@@ -33,7 +37,9 @@

 #include "lib/common/utility.hpp"
 #include "lib/rocprofiler-sdk/counters/dimensions.hpp"
+#include "lib/rocprofiler-sdk/counters/id_decode.hpp"
 #include "lib/rocprofiler-sdk/counters/parser/reader.hpp"
+#include "rocprofiler-sdk/fwd.h"

 namespace rocprofiler
 {
@@ -128,6 +134,75 @@ perform_reduction(ReduceOperation reduce_op, std::vector<rocprofiler_record_coun
    return input_array;
 }

+int64_t
+get_int_encoded_dimensions_from_string(const std::string& rangeStr)
+{
+    int64_t            result = 0;
+    std::istringstream iss(rangeStr);
+    std::string        token;
+    size_t             bit_length = DIM_BIT_LENGTH / ROCPROFILER_DIMENSION_LAST;
+
+    while(std::getline(iss, token, ','))
+    {
+        token.erase(std::remove_if(token.begin(), token.end(), ::isspace), token.end());
+
+        size_t dash_pos = token.find(':');
+        if(dash_pos != std::string::npos)
+        {
+            throw std::runtime_error(
+                fmt::format("Range based selection not supported by Dimension API. only select "
+                            "single value for each dimension."));
+            int start = std::stoi(token.substr(0, dash_pos));
+            int end   = std::stoi(token.substr(dash_pos + 1));
+            result |= (1LL << std::min(64, end + 1)) - (1LL << std::max(start, 0));
+        }
+        else
+        {
+            int num = std::stoi(token);
+            if(num < (1 << bit_length))
+            {
+                result |= (1LL << num);
+            }
+            else
+            {
+                throw std::runtime_error(fmt::format("Dimension value exceeds max allowed."));
+            }
+        }
+    }
+    return result;
+}
+
+std::vector<rocprofiler_record_counter_t>*
+perform_selection(std::map<rocprofiler_profile_counter_instance_types, std::string>& dimension_map,
+                  std::vector<rocprofiler_record_counter_t>*                         input_array)
+{
+    if(input_array->empty()) return input_array;
+    for(auto& dim_pair : dimension_map)
+    {
+        int64_t encoded_dim_values = get_int_encoded_dimensions_from_string(dim_pair.second);
+        size_t  bit_length         = DIM_BIT_LENGTH / ROCPROFILER_DIMENSION_LAST;
+        int64_t mask = (MAX_64 >> (64 - bit_length)) << ((dim_pair.first - 1) * bit_length);
+
+        input_array->erase(std::remove_if(input_array->begin(),
+                                          input_array->end(),
+                                          [&](rocprofiler_record_counter_t& rec) {
+                                              bool should_remove =
+                                                  (encoded_dim_values &
+                                                   (1 << rocprofiler::counters::rec_to_dim_pos(
+                                                        rec.id, dim_pair.first))) == 0;
+                                              if(!should_remove)
+                                              {
+                                                  rec.id = rec.id | mask;
+                                                  rec.id = rec.id ^ mask;
+                                              }
+                                              return should_remove;
+                                          }),
+                           input_array->end());
+    }
+
+    return input_array;
+}
+
 }  // namespace

 const std::unordered_map<std::string, EvaluateASTMap>&
@@ -214,6 +289,7 @@ EvaluateAST::EvaluateAST(rocprofiler_counter_id_t                       out_id,
 , _reduce_op(get_reduce_op_type_from_string(ast.reduce_op))
 , _agent(std::move(agent))
 , _reduce_dimension_set(ast.reduce_dimension_set)
+, _select_dimension_map(ast.select_dimension_map)
 , _out_id(out_id)
 {
    if(_type == NodeType::REFERENCE_NODE || _type == NodeType::ACCUMULATE_NODE)
@@ -308,7 +384,25 @@ EvaluateAST::set_dimensions()
        break;
        case SELECT_NODE:
        {
-            // TODO: future scope
+            auto first = _children[0].set_dimensions();
+            first.erase(std::remove_if(first.begin(),
+                                       first.end(),
+                                       [&](const MetricDimension& dim) {
+                                           return _select_dimension_map.find(dim.type()) !=
+                                                  _select_dimension_map.end();
+                                       }),
+                        first.end());
+            if(first.empty())
+            {
+                _dimension_types = std::vector<MetricDimension>{
+                    {dimension_map().at(ROCPROFILER_DIMENSION_INSTANCE),
+                     1,
+                     ROCPROFILER_DIMENSION_INSTANCE}};
+            }
+            else
+            {
+                _dimension_types = first;
+            }
        }
        break;
    }
@@ -675,8 +769,11 @@ EvaluateAST::evaluate(
                                                     static_cast<int>(_reduce_op)));
            return perform_reduction(_reduce_op, result);
        }
-        // Currently unsupported
-        case SELECT_NODE: break;
+        case SELECT_NODE:
+        {
+            auto* result = _children.at(0).evaluate(results_map, cache);
+            return perform_selection(_select_dimension_map, result);
+        }
    }

    return nullptr;
@@ -27,6 +27,7 @@

 #include "lib/rocprofiler-sdk/aql/packet_construct.hpp"
 #include "lib/rocprofiler-sdk/counters/dimensions.hpp"
+#include "lib/rocprofiler-sdk/counters/id_decode.hpp"
 #include "lib/rocprofiler-sdk/counters/metrics.hpp"
 #include "lib/rocprofiler-sdk/counters/parser/raw_ast.hpp"

@@ -166,17 +167,18 @@ public:
    const rocprofiler_counter_id_t& out_id() const { return _out_id; }

 private:
-    NodeType                                                       _type{NONE};
-    ReduceOperation                                                _reduce_op{REDUCE_NONE};
-    Metric                                                         _metric;
-    double                                                         _raw_value{0};
-    std::vector<EvaluateAST>                                       _children;
-    std::string                                                    _agent;
-    std::vector<MetricDimension>                                   _dimension_types{};
-    std::vector<rocprofiler_record_counter_t>                      _static_value;
-    std::unordered_set<rocprofiler_profile_counter_instance_types> _reduce_dimension_set;
-    bool                                                           _expanded{false};
-    rocprofiler_counter_id_t                                       _out_id{.handle = 0};
+    NodeType                                                          _type{NONE};
+    ReduceOperation                                                   _reduce_op{REDUCE_NONE};
+    Metric                                                            _metric;
+    double                                                            _raw_value{0};
+    std::vector<EvaluateAST>                                          _children;
+    std::string                                                       _agent;
+    std::vector<MetricDimension>                                      _dimension_types{};
+    std::vector<rocprofiler_record_counter_t>                         _static_value;
+    std::unordered_set<rocprofiler_profile_counter_instance_types>    _reduce_dimension_set;
+    std::map<rocprofiler_profile_counter_instance_types, std::string> _select_dimension_map;
+    bool                                                              _expanded{false};
+    rocprofiler_counter_id_t                                          _out_id{.handle = 0};
 };

 using EvaluateASTMap = std::unordered_map<std::string, EvaluateAST>;
@@ -23,13 +23,16 @@
 #include <algorithm>
 #include <cstdint>
 #include <tuple>
+#include <unordered_map>
 #include <utility>
+#include <vector>

 #include <fmt/core.h>
 #include <gtest/gtest.h>

 #include "lib/rocprofiler-sdk/agent.hpp"
 #include "lib/rocprofiler-sdk/counters/evaluate_ast.hpp"
+#include "lib/rocprofiler-sdk/counters/id_decode.hpp"
 #include "lib/rocprofiler-sdk/counters/parser/reader.hpp"

 namespace
@@ -364,6 +367,7 @@ construct_test_data_dim(
        record.counter_value =
            static_cast<double>(rand()) / (static_cast<double>(RAND_MAX / 50000)) + 1.0;
        auto recursive_dim = construct_test_data_dim(record.id, dims, dim_size);
+        if(!dims.empty()) ret.pop_back();
        ret.insert(ret.end(), recursive_dim.begin(), recursive_dim.end());
    }
    return ret;
@@ -1233,3 +1237,132 @@ TEST(evaluate_ast, derived_counter_reduction)
        }
    }
 }
+
+TEST(evatuate_ast, evaluate_select)
+{
+    using namespace rocprofiler::counters;
+
+    auto get_base_rec_id = [](uint64_t counter_id) {
+        rocprofiler_counter_instance_id_t base_id = 0;
+        set_counter_in_rec(base_id, {.handle = counter_id});
+        return base_id;
+    };
+
+    auto select_dim =
+        [](auto                                 a,
+           std::unordered_map<rocprofiler::counters::rocprofiler_profile_counter_instance_types,
+                              std::vector<int>> dims) -> auto
+    {
+        if(a.empty()) return a;
+        for(auto& dim_pair : dims)
+        {
+            a.erase(std::remove_if(a.begin(),
+                                   a.end(),
+                                   [&](rocprofiler_record_counter_t rec) {
+                                       return std::find(dim_pair.second.begin(),
+                                                        dim_pair.second.end(),
+                                                        rec_to_dim_pos(rec.id, dim_pair.first)) ==
+                                              dim_pair.second.end();
+                                   }),
+                    a.end());
+        }
+        for(auto& dim_pair : dims)
+        {
+            size_t  bit_length = DIM_BIT_LENGTH / ROCPROFILER_DIMENSION_LAST;
+            int64_t mask = (MAX_64 >> (64 - bit_length)) << ((dim_pair.first - 1) * bit_length);
+            for(auto& rec : a)
+            {
+                rec.id = rec.id | mask;
+                rec.id = rec.id ^ mask;
+            }
+        }
+        return a;
+    };
+
+    std::unordered_map<std::string, Metric> metrics = {
+        {"VOORHEES", Metric("gfx9", "VOORHEES", "a", "a", "a", "", "", 0)},
+        {"KRUEGER", Metric("gfx9", "KRUEGER", "a", "a", "a", "", "", 1)},
+        {"MYERS", Metric("gfx9", "MYERS", "a", "a", "a", "", "", 2)},
+        {"BATES",
+         Metric("gfx9",
+                "BATES",
+                "C",
+                "C",
+                "C",
+                "select(VOORHEES+KRUEGER,[DIMENSION_XCC=[0]])",
+                "",
+                3)},
+        {"KRAMER",
+         Metric("gfx9",
+                "KRAMER",
+                "C",
+                "C",
+                "C",
+                "select(MYERS,[DIMENSION_XCC=[1],DIMENSION_SHADER_ARRAY=[0]])",
+                "",
+                4)},
+    };
+
+    std::unordered_map<std::string, std::vector<rocprofiler_record_counter_t>> base_counter_data = {
+        {"VOORHEES", construct_test_data_dim(get_base_rec_id(0), {ROCPROFILER_DIMENSION_XCC}, 8)},
+        {"KRUEGER", construct_test_data_dim(get_base_rec_id(1), {ROCPROFILER_DIMENSION_XCC}, 8)},
+        {"MYERS",
+         construct_test_data_dim(get_base_rec_id(2),
+                                 {ROCPROFILER_DIMENSION_XCC, ROCPROFILER_DIMENSION_SHADER_ARRAY},
+                                 8)}};
+
+    std::unordered_map<std::string, std::unordered_map<std::string, EvaluateAST>> asts;
+    for(const auto& [val, metric] : metrics)
+    {
+        RawAST* ast = nullptr;
+        auto    buf = yy_scan_string(metric.expression().empty() ? metric.name().c_str()
+                                                                 : metric.expression().c_str());
+        yyparse(&ast);
+        ASSERT_TRUE(ast) << metric.expression() << " " << metric.name();
+        asts.emplace("gfx9", std::unordered_map<std::string, EvaluateAST>{})
+            .first->second.emplace(val,
+                                   EvaluateAST({.handle = metric.id()}, metrics, *ast, "gfx9"));
+        yy_delete_buffer(buf);
+        delete ast;
+    }
+
+    std::vector<std::tuple<std::string, std::vector<rocprofiler_record_counter_t>, int64_t>>
+        derived_counters = {
+            {"BATES",
+             select_dim(plus_vec(base_counter_data["VOORHEES"], base_counter_data["KRUEGER"]),
+                        {{ROCPROFILER_DIMENSION_XCC, {0}}}),
+             2},
+            {"KRAMER",
+             select_dim(
+                 base_counter_data["MYERS"],
+                 {{ROCPROFILER_DIMENSION_XCC, {1}}, {ROCPROFILER_DIMENSION_SHADER_ARRAY, {0}}}),
+             1},
+        };
+    std::unordered_map<uint64_t, std::vector<rocprofiler_record_counter_t>> base_counter_decode;
+    for(const auto& [name, base_counter_v] : base_counter_data)
+    {
+        base_counter_decode[metrics[name].id()] = base_counter_v;
+    }
+
+    for(auto& [name, expected, eval_count] : derived_counters)
+    {
+        ROCP_INFO << name;
+        auto eval_counters =
+            rocprofiler::counters::get_required_hardware_counters(asts, "gfx9", metrics[name]);
+        ASSERT_TRUE(eval_counters);
+        ASSERT_EQ(eval_counters->size(), eval_count);
+        std::vector<std::unique_ptr<std::vector<rocprofiler_record_counter_t>>> cache;
+        asts.at("gfx9").at(name).expand_derived(asts.at("gfx9"));
+        auto ret = asts.at("gfx9").at(name).evaluate(base_counter_decode, cache);
+        EXPECT_EQ(ret->size(), expected.size());
+        int pos = 0;
+        asts.at("gfx9").at(name).set_out_id(*ret);
+        for(const auto& v : *ret)
+        {
+            set_counter_in_rec(expected[pos].id, {.handle = metrics[name].id()});
+            EXPECT_EQ(v.id, expected[pos].id);
+            EXPECT_FLOAT_EQ(v.counter_value, expected[pos].counter_value);
+            pos++;
+        }
+    }
+}