[impeller] makes UniformBindData 15% faster and adds unit test

[gaaclarke]

issue: flutter/flutter#159177

The improvements come from using absl::flat_hash_map.

measurements with std vs absl (host_profile_arm64):

# Std
0.041124
0.0398002
0.0396546
0.0405552
0.0397742

# Absl
0.0353708
0.0336834
0.0338616
0.033050
0.0331976

Benchmark used

TEST(BufferBindingsGLESTest, BindUniformDataMicro) {
  BufferBindingsGLES bindings;
  absl::flat_hash_map<std::string, GLint> uniform_bindings;
  uniform_bindings["SHADERMETADATA.FOOBAR"] = 1;
  bindings.SetUniformBindings(std::move(uniform_bindings));
  std::shared_ptr<MockGLES> mock_gl = MockGLES::Init();
  MockAllocator allocator;
  Bindings vertex_bindings;

  ShaderMetadata shader_metadata = {
      .name = "shader_metadata",
      .members = {ShaderStructMemberMetadata{.type = ShaderType::kFloat,
                                             .name = "foobar",
                                             .offset = 0,
                                             .size = sizeof(float),
                                             .byte_length = sizeof(float)}}};
  std::shared_ptr<ReactorGLES> reactor;
  std::shared_ptr<Allocation> backing_store = std::make_shared<Allocation>();
  ASSERT_TRUE(backing_store->Truncate(Bytes{sizeof(float)}));
  DeviceBufferGLES device_buffer(DeviceBufferDescriptor{.size = sizeof(float)},
                                 reactor, backing_store);
  BufferView buffer_view(&device_buffer, Range(0, sizeof(float)));
  vertex_bindings.buffers.push_back(BufferAndUniformSlot{
      .slot =
          ShaderUniformSlot{
              .name = "foobar", .ext_res_0 = 0, .set = 0, .binding = 0},
      .view = BufferResource(&shader_metadata, buffer_view)});
  Bindings fragment_bindings;
  int32_t count = 5'000'000;
  auto start = std::chrono::high_resolution_clock::now();
  for (int32_t i = 0; i < count; ++i) {
    bindings.BindUniformData(mock_gl->GetProcTable(), allocator,
                             vertex_bindings, fragment_bindings);
    if (i % 100 == 0) {
      mock_gl->GetCapturedCalls();
    }
  }
  auto end = std::chrono::high_resolution_clock::now();
  auto duration =
      std::chrono::duration_cast<std::chrono::microseconds>(end - start)
          .count();
  std::cout << "Execution time: " << duration / static_cast<double>(count)
            << " microseconds" << std::endl;
}

This is one of our hottest symbols on the raster thread:

Pre-launch Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I read the Tree Hygiene wiki page, which explains my responsibilities.
• I read and followed the Flutter Style Guide and the C++, Objective-C, Java style guides.
• I listed at least one issue that this PR fixes in the description above.
• I added new tests to check the change I am making or feature I am adding, or the PR is test-exempt. See testing the engine for instructions on writing and running engine tests.
• I updated/added relevant documentation (doc comments with ///).
• I signed the CLA.
• All existing and new tests are passing.

If you need help, consider asking for advice on the #hackers-new channel on Discord.

[gaaclarke]

We are planning on refactoring this at some point, in the meantime this is an easy win in case that doesn't pan out.

[jonahwilliams]

Since you're using a real allocator here I would use this in place of the mock allocator above. In fact, the transients_allocator arg is actually unused so you could also just delete it.

[gaaclarke]

done

[jonahwilliams]

LGTM